RepoMirrors
/
mpv
mirror of https://github.com/mpv-player/mpv
1
0
Fork 0
Code Issues Releases Wiki Activity
mpv/player/video.c

1574 lines
52 KiB
C
Raw Normal View History

Split mplayer.c mplayer.c was a bit too big. Split it into multiple files. I hope the way it's split makes sense. Maybe some things don't make too much sense, or go against intuition. These will fixed as soon as I notice them. Some files are a bit questionable (misc.c, osd.c, configfiles.c), and suggestions how to organize this better are welcome. Regressions are possible due to reorganized include statements. Obviously I didn't just copy mplayer.c's orgy of include statements, but recreated them for each file. It's easily possible that there are oversights and mistakes, which will show up on other platforms. There is one actual change: the public avutil.h include is removed from encode.h, and I tried to replace most FFMIN/FFMAX/av_clip uses. I consider using libavutil too much as dangerous, because the set of include files they recursively pull in is rather arbitrary and is different between FFmpeg and Libav.
2013-10-29 21:38:29 +00:00
/*
Update license headers Signed-off-by: wm4 <wm4@nowhere>
2015-04-13 07:36:54 +00:00
* This file is part of mpv.
Split mplayer.c mplayer.c was a bit too big. Split it into multiple files. I hope the way it's split makes sense. Maybe some things don't make too much sense, or go against intuition. These will fixed as soon as I notice them. Some files are a bit questionable (misc.c, osd.c, configfiles.c), and suggestions how to organize this better are welcome. Regressions are possible due to reorganized include statements. Obviously I didn't just copy mplayer.c's orgy of include statements, but recreated them for each file. It's easily possible that there are oversights and mistakes, which will show up on other platforms. There is one actual change: the public avutil.h include is removed from encode.h, and I tried to replace most FFMIN/FFMAX/av_clip uses. I consider using libavutil too much as dangerous, because the set of include files they recursively pull in is rather arbitrary and is different between FFmpeg and Libav.
2013-10-29 21:38:29 +00:00
*
player: change license of most core files to LGPL These files have all in common that they were fully or mostly taken from mplayer.c. (mplayer.c was a huge file that contains almost all of the playback core, until it was split into multiple parts.) This was probably the hardest part to relicense, because so much code was moved around all the time. player/audio.c still does not compile. We'll have to redo audio filtering. Once that is done, we can probably actually provide an actual LGPL configure switch. Here is a relatively detailed list of potential issues: 8d190244: author did not reply, parts were made GPL-only in a previous commit. 7882ea9b: author could not be reached, but the code is gone. wscript still has --datadir switch, but I don't think this is relevant to copyright. f197efd5: unclear origin, but I consider the code gone anyway (replaced with generic OSD mechanisms). 8337d9c2: author did not reply, but only the option still exists (under a different name), other code was removed. d8fd7131: did not reply. Disabled in a previous commit. 05258251: same author as above. Both fields actually seem to have vanished (even when tracking renames), so no action taken. d459e644, 268b2c1a: author did not reply, but we reuse only the options (with different names and slightly or fully different semantics, and completely different implementations), so I don't think this is relevant for copyright. 09e742fe, 17c39c4e: same as above. e8a173de, bff4b3ee: author could not be reached. The commands were reworked to properties, and the code outside of the TV code were moved back to the TV code. So I don't think copyright applies to the current command.c parts (mp_property_tv_color, mp_property_tv_freq, mp_property_tv_scan). The TV parts remain GPL. 0810e427: could not be reached. Disabled in a previous commit. 43744a2d: unknown author, but this was replaced by dynamic alloc (if the change is even copyrightable). 116ca0c7: unknown author; reasoning see input.c relicensing commit. e7e4d1d8: these semantics still exist, but as generic code, and this code was fully removed. f1175cd9: the author of the cited patch is unknown, and upon inspection it turns out that I was only using the idea to pause the player on EOF, so I claim it's not copyright relevant. 25affdcc: author could not be reached (yet) - but it's only a function rename, not copyrightable. 5728504c was committed by Arpi (who agreed), but hints that it might be by a different author. In fact it seems to be mostly this patch: http://lists.mplayerhq.hu/pipermail/mplayer-dev-eng/2001-November/002041.html The author did not respond, but it all seems to have been removed later. It's a terrible mess though. Arpi reverted the A-V sync code at first, but left the RTC code for a while. The following commits remove these changes 100%: 14b35442, 7181a091, 31482783, 614f8475, df58e822. cehoyos did explicitly not agree to LGPL, but was involved in the following changes: c99d8fc8: applied a patch and didn't modify it, the original author agreed. 40ac0d31: author could not be reached, but all code is gone anyway. The "af" command has a similar function, but works completely different and actually reuses a mechanism older than this patch. 54350436: applied a patch, but didn't modify it, except for adding a German translation, which was removed later. a2dda036: same situation as above 240b743e: this was made GPL-only in a previous commit 7b25afd7: same as above (for now) kirijua could not be reached, but was a regular patch contributor: c2c997fd: video equalizer code move; probably not copyrightable. Is GPL due to Nick anyway. be54f481: technically, this became the audio track property later. But all what is left is the fact that you pass a track ID to it, so consider the original coypright non-relevant. 2f376d1b: this was rewritten in b7052b43, but for now we can afford to be careful, so this was marked as GPL only in a previous commit. 43844d09: remaining parts in main.c were reverted in a previous commit. anders has mostly disagreed with the LGPL relicensing. Does not want libaf to become LGPL, but made some concessions. In particular, he granted us permission to relicense 4943e9c52c and 242aa6ebd4. We also consider some of his changes remaining in mpv not relevant for copyright (such as 735de602 - we won't remove the this option completely). We will completely remove his other contributions, including the entire audio filter chain. For now, this stuff is marked as GPL only. The remaining question is how much code in player/audio.c (based on the former mplayer.c and dec_audio.c) is under his copyright. I made claims about this in a previous commit. Nick(ols) Kurshev, svn username "nick" and "nickols_k", could not be reached. He had a lot of changes in early MPlayer. It seems all of that was removed, at least in mpv. His main work, like VIDIX or libswscale work, does not exist in mpv anymore, but the changes to mplayer.c and other core parts still deserve attention: a4119f6b, fb927549, ad3529b8, e11b23dc, 5f2178be, 93c371d5: removed in b43d67e0, d1628d12, 24ed01fe, df58e822. 0a83c6ec, 104c125e, 4e067f62, aec5dcc8, b587a3d6, f3de6e6b: DR, VAA, and "tune" stuff was fully removed later on or replaced with other mechanisms. 340183b0: screenshots were redone later (the VOCTRL was even removed, with an independent implementation using the same VOCTRL a few years later), so not relevant anymore. Basically only the 's' shortcut remains (but not its implementation). 92c5c274, bffd4007, 555c6766: for now marked as GPL only in a previous commit. Might contain some trace amounts of "michael"'s copyright, who agreed to LGPL only once the core is relicensed. This will still be respected, but I don't think it matters at this in this case. (Some code touched by him was merged into mplayer.c, and then disappeared after heavy refactoring.) I tried to be as careful and as complete as possible. It can't be excluded that amends to this will be made later. This does not make the player LGPL yet.
2017-06-23 13:53:41 +00:00
* mpv is free software; you can redistribute it and/or
* modify it under the terms of the GNU Lesser General Public
* License as published by the Free Software Foundation; either
* version 2.1 of the License, or (at your option) any later version.
Split mplayer.c mplayer.c was a bit too big. Split it into multiple files. I hope the way it's split makes sense. Maybe some things don't make too much sense, or go against intuition. These will fixed as soon as I notice them. Some files are a bit questionable (misc.c, osd.c, configfiles.c), and suggestions how to organize this better are welcome. Regressions are possible due to reorganized include statements. Obviously I didn't just copy mplayer.c's orgy of include statements, but recreated them for each file. It's easily possible that there are oversights and mistakes, which will show up on other platforms. There is one actual change: the public avutil.h include is removed from encode.h, and I tried to replace most FFMIN/FFMAX/av_clip uses. I consider using libavutil too much as dangerous, because the set of include files they recursively pull in is rather arbitrary and is different between FFmpeg and Libav.
2013-10-29 21:38:29 +00:00
*
Update license headers Signed-off-by: wm4 <wm4@nowhere>
2015-04-13 07:36:54 +00:00
* mpv is distributed in the hope that it will be useful,
Split mplayer.c mplayer.c was a bit too big. Split it into multiple files. I hope the way it's split makes sense. Maybe some things don't make too much sense, or go against intuition. These will fixed as soon as I notice them. Some files are a bit questionable (misc.c, osd.c, configfiles.c), and suggestions how to organize this better are welcome. Regressions are possible due to reorganized include statements. Obviously I didn't just copy mplayer.c's orgy of include statements, but recreated them for each file. It's easily possible that there are oversights and mistakes, which will show up on other platforms. There is one actual change: the public avutil.h include is removed from encode.h, and I tried to replace most FFMIN/FFMAX/av_clip uses. I consider using libavutil too much as dangerous, because the set of include files they recursively pull in is rather arbitrary and is different between FFmpeg and Libav.
2013-10-29 21:38:29 +00:00
* but WITHOUT ANY WARRANTY; without even the implied warranty of
* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
player: change license of most core files to LGPL These files have all in common that they were fully or mostly taken from mplayer.c. (mplayer.c was a huge file that contains almost all of the playback core, until it was split into multiple parts.) This was probably the hardest part to relicense, because so much code was moved around all the time. player/audio.c still does not compile. We'll have to redo audio filtering. Once that is done, we can probably actually provide an actual LGPL configure switch. Here is a relatively detailed list of potential issues: 8d190244: author did not reply, parts were made GPL-only in a previous commit. 7882ea9b: author could not be reached, but the code is gone. wscript still has --datadir switch, but I don't think this is relevant to copyright. f197efd5: unclear origin, but I consider the code gone anyway (replaced with generic OSD mechanisms). 8337d9c2: author did not reply, but only the option still exists (under a different name), other code was removed. d8fd7131: did not reply. Disabled in a previous commit. 05258251: same author as above. Both fields actually seem to have vanished (even when tracking renames), so no action taken. d459e644, 268b2c1a: author did not reply, but we reuse only the options (with different names and slightly or fully different semantics, and completely different implementations), so I don't think this is relevant for copyright. 09e742fe, 17c39c4e: same as above. e8a173de, bff4b3ee: author could not be reached. The commands were reworked to properties, and the code outside of the TV code were moved back to the TV code. So I don't think copyright applies to the current command.c parts (mp_property_tv_color, mp_property_tv_freq, mp_property_tv_scan). The TV parts remain GPL. 0810e427: could not be reached. Disabled in a previous commit. 43744a2d: unknown author, but this was replaced by dynamic alloc (if the change is even copyrightable). 116ca0c7: unknown author; reasoning see input.c relicensing commit. e7e4d1d8: these semantics still exist, but as generic code, and this code was fully removed. f1175cd9: the author of the cited patch is unknown, and upon inspection it turns out that I was only using the idea to pause the player on EOF, so I claim it's not copyright relevant. 25affdcc: author could not be reached (yet) - but it's only a function rename, not copyrightable. 5728504c was committed by Arpi (who agreed), but hints that it might be by a different author. In fact it seems to be mostly this patch: http://lists.mplayerhq.hu/pipermail/mplayer-dev-eng/2001-November/002041.html The author did not respond, but it all seems to have been removed later. It's a terrible mess though. Arpi reverted the A-V sync code at first, but left the RTC code for a while. The following commits remove these changes 100%: 14b35442, 7181a091, 31482783, 614f8475, df58e822. cehoyos did explicitly not agree to LGPL, but was involved in the following changes: c99d8fc8: applied a patch and didn't modify it, the original author agreed. 40ac0d31: author could not be reached, but all code is gone anyway. The "af" command has a similar function, but works completely different and actually reuses a mechanism older than this patch. 54350436: applied a patch, but didn't modify it, except for adding a German translation, which was removed later. a2dda036: same situation as above 240b743e: this was made GPL-only in a previous commit 7b25afd7: same as above (for now) kirijua could not be reached, but was a regular patch contributor: c2c997fd: video equalizer code move; probably not copyrightable. Is GPL due to Nick anyway. be54f481: technically, this became the audio track property later. But all what is left is the fact that you pass a track ID to it, so consider the original coypright non-relevant. 2f376d1b: this was rewritten in b7052b43, but for now we can afford to be careful, so this was marked as GPL only in a previous commit. 43844d09: remaining parts in main.c were reverted in a previous commit. anders has mostly disagreed with the LGPL relicensing. Does not want libaf to become LGPL, but made some concessions. In particular, he granted us permission to relicense 4943e9c52c and 242aa6ebd4. We also consider some of his changes remaining in mpv not relevant for copyright (such as 735de602 - we won't remove the this option completely). We will completely remove his other contributions, including the entire audio filter chain. For now, this stuff is marked as GPL only. The remaining question is how much code in player/audio.c (based on the former mplayer.c and dec_audio.c) is under his copyright. I made claims about this in a previous commit. Nick(ols) Kurshev, svn username "nick" and "nickols_k", could not be reached. He had a lot of changes in early MPlayer. It seems all of that was removed, at least in mpv. His main work, like VIDIX or libswscale work, does not exist in mpv anymore, but the changes to mplayer.c and other core parts still deserve attention: a4119f6b, fb927549, ad3529b8, e11b23dc, 5f2178be, 93c371d5: removed in b43d67e0, d1628d12, 24ed01fe, df58e822. 0a83c6ec, 104c125e, 4e067f62, aec5dcc8, b587a3d6, f3de6e6b: DR, VAA, and "tune" stuff was fully removed later on or replaced with other mechanisms. 340183b0: screenshots were redone later (the VOCTRL was even removed, with an independent implementation using the same VOCTRL a few years later), so not relevant anymore. Basically only the 's' shortcut remains (but not its implementation). 92c5c274, bffd4007, 555c6766: for now marked as GPL only in a previous commit. Might contain some trace amounts of "michael"'s copyright, who agreed to LGPL only once the core is relicensed. This will still be respected, but I don't think it matters at this in this case. (Some code touched by him was merged into mplayer.c, and then disappeared after heavy refactoring.) I tried to be as careful and as complete as possible. It can't be excluded that amends to this will be made later. This does not make the player LGPL yet.
2017-06-23 13:53:41 +00:00
* GNU Lesser General Public License for more details.
Split mplayer.c mplayer.c was a bit too big. Split it into multiple files. I hope the way it's split makes sense. Maybe some things don't make too much sense, or go against intuition. These will fixed as soon as I notice them. Some files are a bit questionable (misc.c, osd.c, configfiles.c), and suggestions how to organize this better are welcome. Regressions are possible due to reorganized include statements. Obviously I didn't just copy mplayer.c's orgy of include statements, but recreated them for each file. It's easily possible that there are oversights and mistakes, which will show up on other platforms. There is one actual change: the public avutil.h include is removed from encode.h, and I tried to replace most FFMIN/FFMAX/av_clip uses. I consider using libavutil too much as dangerous, because the set of include files they recursively pull in is rather arbitrary and is different between FFmpeg and Libav.
2013-10-29 21:38:29 +00:00
*
player: change license of most core files to LGPL These files have all in common that they were fully or mostly taken from mplayer.c. (mplayer.c was a huge file that contains almost all of the playback core, until it was split into multiple parts.) This was probably the hardest part to relicense, because so much code was moved around all the time. player/audio.c still does not compile. We'll have to redo audio filtering. Once that is done, we can probably actually provide an actual LGPL configure switch. Here is a relatively detailed list of potential issues: 8d190244: author did not reply, parts were made GPL-only in a previous commit. 7882ea9b: author could not be reached, but the code is gone. wscript still has --datadir switch, but I don't think this is relevant to copyright. f197efd5: unclear origin, but I consider the code gone anyway (replaced with generic OSD mechanisms). 8337d9c2: author did not reply, but only the option still exists (under a different name), other code was removed. d8fd7131: did not reply. Disabled in a previous commit. 05258251: same author as above. Both fields actually seem to have vanished (even when tracking renames), so no action taken. d459e644, 268b2c1a: author did not reply, but we reuse only the options (with different names and slightly or fully different semantics, and completely different implementations), so I don't think this is relevant for copyright. 09e742fe, 17c39c4e: same as above. e8a173de, bff4b3ee: author could not be reached. The commands were reworked to properties, and the code outside of the TV code were moved back to the TV code. So I don't think copyright applies to the current command.c parts (mp_property_tv_color, mp_property_tv_freq, mp_property_tv_scan). The TV parts remain GPL. 0810e427: could not be reached. Disabled in a previous commit. 43744a2d: unknown author, but this was replaced by dynamic alloc (if the change is even copyrightable). 116ca0c7: unknown author; reasoning see input.c relicensing commit. e7e4d1d8: these semantics still exist, but as generic code, and this code was fully removed. f1175cd9: the author of the cited patch is unknown, and upon inspection it turns out that I was only using the idea to pause the player on EOF, so I claim it's not copyright relevant. 25affdcc: author could not be reached (yet) - but it's only a function rename, not copyrightable. 5728504c was committed by Arpi (who agreed), but hints that it might be by a different author. In fact it seems to be mostly this patch: http://lists.mplayerhq.hu/pipermail/mplayer-dev-eng/2001-November/002041.html The author did not respond, but it all seems to have been removed later. It's a terrible mess though. Arpi reverted the A-V sync code at first, but left the RTC code for a while. The following commits remove these changes 100%: 14b35442, 7181a091, 31482783, 614f8475, df58e822. cehoyos did explicitly not agree to LGPL, but was involved in the following changes: c99d8fc8: applied a patch and didn't modify it, the original author agreed. 40ac0d31: author could not be reached, but all code is gone anyway. The "af" command has a similar function, but works completely different and actually reuses a mechanism older than this patch. 54350436: applied a patch, but didn't modify it, except for adding a German translation, which was removed later. a2dda036: same situation as above 240b743e: this was made GPL-only in a previous commit 7b25afd7: same as above (for now) kirijua could not be reached, but was a regular patch contributor: c2c997fd: video equalizer code move; probably not copyrightable. Is GPL due to Nick anyway. be54f481: technically, this became the audio track property later. But all what is left is the fact that you pass a track ID to it, so consider the original coypright non-relevant. 2f376d1b: this was rewritten in b7052b43, but for now we can afford to be careful, so this was marked as GPL only in a previous commit. 43844d09: remaining parts in main.c were reverted in a previous commit. anders has mostly disagreed with the LGPL relicensing. Does not want libaf to become LGPL, but made some concessions. In particular, he granted us permission to relicense 4943e9c52c and 242aa6ebd4. We also consider some of his changes remaining in mpv not relevant for copyright (such as 735de602 - we won't remove the this option completely). We will completely remove his other contributions, including the entire audio filter chain. For now, this stuff is marked as GPL only. The remaining question is how much code in player/audio.c (based on the former mplayer.c and dec_audio.c) is under his copyright. I made claims about this in a previous commit. Nick(ols) Kurshev, svn username "nick" and "nickols_k", could not be reached. He had a lot of changes in early MPlayer. It seems all of that was removed, at least in mpv. His main work, like VIDIX or libswscale work, does not exist in mpv anymore, but the changes to mplayer.c and other core parts still deserve attention: a4119f6b, fb927549, ad3529b8, e11b23dc, 5f2178be, 93c371d5: removed in b43d67e0, d1628d12, 24ed01fe, df58e822. 0a83c6ec, 104c125e, 4e067f62, aec5dcc8, b587a3d6, f3de6e6b: DR, VAA, and "tune" stuff was fully removed later on or replaced with other mechanisms. 340183b0: screenshots were redone later (the VOCTRL was even removed, with an independent implementation using the same VOCTRL a few years later), so not relevant anymore. Basically only the 's' shortcut remains (but not its implementation). 92c5c274, bffd4007, 555c6766: for now marked as GPL only in a previous commit. Might contain some trace amounts of "michael"'s copyright, who agreed to LGPL only once the core is relicensed. This will still be respected, but I don't think it matters at this in this case. (Some code touched by him was merged into mplayer.c, and then disappeared after heavy refactoring.) I tried to be as careful and as complete as possible. It can't be excluded that amends to this will be made later. This does not make the player LGPL yet.
2017-06-23 13:53:41 +00:00
* You should have received a copy of the GNU Lesser General Public
* License along with mpv. If not, see <http://www.gnu.org/licenses/>.
*
* Parts under HAVE_GPL are licensed under GNU General Public License.
Split mplayer.c mplayer.c was a bit too big. Split it into multiple files. I hope the way it's split makes sense. Maybe some things don't make too much sense, or go against intuition. These will fixed as soon as I notice them. Some files are a bit questionable (misc.c, osd.c, configfiles.c), and suggestions how to organize this better are welcome. Regressions are possible due to reorganized include statements. Obviously I didn't just copy mplayer.c's orgy of include statements, but recreated them for each file. It's easily possible that there are oversights and mistakes, which will show up on other platforms. There is one actual change: the public avutil.h include is removed from encode.h, and I tried to replace most FFMIN/FFMAX/av_clip uses. I consider using libavutil too much as dangerous, because the set of include files they recursively pull in is rather arbitrary and is different between FFmpeg and Libav.
2013-10-29 21:38:29 +00:00
*/
#include <stddef.h>
#include <stdbool.h>
#include <inttypes.h>
#include <math.h>
#include <assert.h>
#include "config.h"
mpv_talloc.h: rename from talloc.h This change helps avoiding conflict with talloc.h from libtalloc.
2016-01-11 18:03:40 +00:00
#include "mpv_talloc.h"
Split mplayer.c mplayer.c was a bit too big. Split it into multiple files. I hope the way it's split makes sense. Maybe some things don't make too much sense, or go against intuition. These will fixed as soon as I notice them. Some files are a bit questionable (misc.c, osd.c, configfiles.c), and suggestions how to organize this better are welcome. Regressions are possible due to reorganized include statements. Obviously I didn't just copy mplayer.c's orgy of include statements, but recreated them for each file. It's easily possible that there are oversights and mistakes, which will show up on other platforms. There is one actual change: the public avutil.h include is removed from encode.h, and I tried to replace most FFMIN/FFMAX/av_clip uses. I consider using libavutil too much as dangerous, because the set of include files they recursively pull in is rather arbitrary and is different between FFmpeg and Libav.
2013-10-29 21:38:29 +00:00
Split mpvcore/ into common/, misc/, bstr/
2013-12-17 01:39:45 +00:00
#include "common/msg.h"
Move options/config related files from mpvcore/ to options/ Since m_option.h and options.h are extremely often included, a lot of files have to be changed. Moving path.c/h to options/ is a bit questionable, but since this is mainly about access to config files (which are also handled in options/), it's probably ok.
2013-12-17 01:02:25 +00:00
#include "options/options.h"
video: auto-insert software rotation filter If the VO can't do rotation, insert a filter to do this. Note that this doesn't reuse the filter insertion code from command.c (used by "vf" input command), because that would end up more complicated: we don't even want to change the user filter option.
2014-04-20 21:54:13 +00:00
#include "options/m_config.h"
#include "options/m_option.h"
Split mpvcore/ into common/, misc/, bstr/
2013-12-17 01:39:45 +00:00
#include "common/common.h"
#include "common/encode.h"
Move options/config related files from mpvcore/ to options/ Since m_option.h and options.h are extremely often included, a lot of files have to be changed. Moving path.c/h to options/ is a bit questionable, but since this is mainly about access to config files (which are also handled in options/), it's probably ok.
2013-12-17 01:02:25 +00:00
#include "options/m_property.h"
player: move video display code out of the playloop Basically move the code from playloop.c to video.c. The new function write_video() now contains the code that was part of run_playloop(). There are no functional changes, except handling "new_frame_shown" slightly differently. This is done so that we don't need new a new MPContext field or a return value for write_video() to signal this condition. Instead, it's handled indirectly.
2014-07-30 21:24:08 +00:00
#include "osdep/timer.h"
Split mplayer.c mplayer.c was a bit too big. Split it into multiple files. I hope the way it's split makes sense. Maybe some things don't make too much sense, or go against intuition. These will fixed as soon as I notice them. Some files are a bit questionable (misc.c, osd.c, configfiles.c), and suggestions how to organize this better are welcome. Regressions are possible due to reorganized include statements. Obviously I didn't just copy mplayer.c's orgy of include statements, but recreated them for each file. It's easily possible that there are oversights and mistakes, which will show up on other platforms. There is one actual change: the public avutil.h include is removed from encode.h, and I tried to replace most FFMIN/FFMAX/av_clip uses. I consider using libavutil too much as dangerous, because the set of include files they recursively pull in is rather arbitrary and is different between FFmpeg and Libav.
2013-10-29 21:38:29 +00:00
#include "audio/out/ao.h"
#include "demux/demux.h"
#include "stream/stream.h"
Rename sub.c/.h to osd.c/.h This was way too misleading. osd.c merely calls the subtitle renderers, instead of actually dealing with subtitles.
2013-11-24 11:58:06 +00:00
#include "sub/osd.h"
video: move struct mp_hwdec_info into its own header file This means most code accessing this struct must now include hwdec.h instead of dec_video.h. I just put it into dec_video.h at first because I thought a separate file would be a waste, but it's more proper to do it this way, as there are too many files which include dec_video.h only to get the mp_hwdec_info definition.
2013-11-23 20:26:31 +00:00
#include "video/hwdec.h"
Split mplayer.c mplayer.c was a bit too big. Split it into multiple files. I hope the way it's split makes sense. Maybe some things don't make too much sense, or go against intuition. These will fixed as soon as I notice them. Some files are a bit questionable (misc.c, osd.c, configfiles.c), and suggestions how to organize this better are welcome. Regressions are possible due to reorganized include statements. Obviously I didn't just copy mplayer.c's orgy of include statements, but recreated them for each file. It's easily possible that there are oversights and mistakes, which will show up on other platforms. There is one actual change: the public avutil.h include is removed from encode.h, and I tried to replace most FFMIN/FFMAX/av_clip uses. I consider using libavutil too much as dangerous, because the set of include files they recursively pull in is rather arbitrary and is different between FFmpeg and Libav.
2013-10-29 21:38:29 +00:00
#include "video/filter/vf.h"
#include "video/decode/dec_video.h"
video: move video filter chain initialization from decoder to player This should help fixing some issues (like not draining video frames correctly on reinit), as well as decoupling the decoder, filter chain, and VO code. I also wanted to make the hardware video decoding fallback work properly if software-only video filters are inserted. This currently has the issue that the fallback is too violent, and throws away a bunch of demuxer packets needed to restart software decoding properly. But keeping "backup" packets turned out as too hacky, so I'm not doing this, at least not yet.
2013-12-10 18:07:29 +00:00
#include "video/decode/vd.h"
Split mplayer.c mplayer.c was a bit too big. Split it into multiple files. I hope the way it's split makes sense. Maybe some things don't make too much sense, or go against intuition. These will fixed as soon as I notice them. Some files are a bit questionable (misc.c, osd.c, configfiles.c), and suggestions how to organize this better are welcome. Regressions are possible due to reorganized include statements. Obviously I didn't just copy mplayer.c's orgy of include statements, but recreated them for each file. It's easily possible that there are oversights and mistakes, which will show up on other platforms. There is one actual change: the public avutil.h include is removed from encode.h, and I tried to replace most FFMIN/FFMAX/av_clip uses. I consider using libavutil too much as dangerous, because the set of include files they recursively pull in is rather arbitrary and is different between FFmpeg and Libav.
2013-10-29 21:38:29 +00:00
#include "video/out/vo.h"
player: add display sync mode If this mode is enabled, the player tries to strictly synchronize video to display refresh. It will adjust playback speed to match the display, so if you play 23.976 fps video on a 24 Hz screen, playback speed is increased by approximately 1/1000. Audio wll be resampled to keep up with playback. This is different from the default sync mode, which will sync video to audio, with the consequence that video might skip or repeat a frame once in a while to make video keep up with audio. This is still unpolished. There are some major problems as well; in particular, mkv VFR files won't work well. The reason is that Matroska is terrible and rounds timestamps to milliseconds. This makes it rather hard to guess the framerate of a section of video that is playing. We could probably fix this by just accepting jittery timestamps (instead of explicitly disabling the sync code in this case), but I'm not ready to accept such a solution yet. Another issue is that we are extremely reliant on OS video and audio APIs working in an expected manner, which of course is not too often the case. Consequently, the new sync mode is a bit fragile.
2015-08-10 16:43:25 +00:00
#include "audio/filter/af.h"
#include "audio/decode/dec_audio.h"
Split mplayer.c mplayer.c was a bit too big. Split it into multiple files. I hope the way it's split makes sense. Maybe some things don't make too much sense, or go against intuition. These will fixed as soon as I notice them. Some files are a bit questionable (misc.c, osd.c, configfiles.c), and suggestions how to organize this better are welcome. Regressions are possible due to reorganized include statements. Obviously I didn't just copy mplayer.c's orgy of include statements, but recreated them for each file. It's easily possible that there are oversights and mistakes, which will show up on other platforms. There is one actual change: the public avutil.h include is removed from encode.h, and I tried to replace most FFMIN/FFMAX/av_clip uses. I consider using libavutil too much as dangerous, because the set of include files they recursively pull in is rather arbitrary and is different between FFmpeg and Libav.
2013-10-29 21:38:29 +00:00
Rename mp_core.h to core.h Get rid of the mp_ prefix.
2013-12-17 00:08:53 +00:00
#include "core.h"
Split mplayer.c mplayer.c was a bit too big. Split it into multiple files. I hope the way it's split makes sense. Maybe some things don't make too much sense, or go against intuition. These will fixed as soon as I notice them. Some files are a bit questionable (misc.c, osd.c, configfiles.c), and suggestions how to organize this better are welcome. Regressions are possible due to reorganized include statements. Obviously I didn't just copy mplayer.c's orgy of include statements, but recreated them for each file. It's easily possible that there are oversights and mistakes, which will show up on other platforms. There is one actual change: the public avutil.h include is removed from encode.h, and I tried to replace most FFMIN/FFMAX/av_clip uses. I consider using libavutil too much as dangerous, because the set of include files they recursively pull in is rather arbitrary and is different between FFmpeg and Libav.
2013-10-29 21:38:29 +00:00
#include "command.h"
player: move video display code out of the playloop Basically move the code from playloop.c to video.c. The new function write_video() now contains the code that was part of run_playloop(). There are no functional changes, except handling "new_frame_shown" slightly differently. This is done so that we don't need new a new MPContext field or a return value for write_video() to signal this condition. Instead, it's handled indirectly.
2014-07-30 21:24:08 +00:00
#include "screenshot.h"
player: rewrite deinterlace filter auto-insertion Instead of using the "vf" command code (which changes filters at runtime on user input), use the general filter-insertion code. The latter was added later, and is more suitable for automatically inserted filters. The old code failed in particular when using watch-later saving, which stored the filter list in the resume config file. If a user changed the hardware decoding mode via command line, the stored filter chain was out of date and could cause failure due to not working with hardware or software decoding mode. Storing the deinterlace filter in the filter list was unavoidable, because it was part of the user state. (The new code only edits the actually instantiated filters.)
2016-07-05 19:10:26 +00:00
#define VF_DEINTERLACE_LABEL "deinterlace"
player: move video display code out of the playloop Basically move the code from playloop.c to video.c. The new function write_video() now contains the code that was part of run_playloop(). There are no functional changes, except handling "new_frame_shown" slightly differently. This is done so that we don't need new a new MPContext field or a return value for write_video() to signal this condition. Instead, it's handled indirectly.
2014-07-30 21:24:08 +00:00
enum {
// update_video() - code also uses: <0 error, 0 eof, >0 progress
VD_ERROR = -1,
VD_EOF = 0, // end of file - no new output
VD_PROGRESS = 1, // progress, but no output; repeat call with no waiting
VD_NEW_FRAME = 2, // the call produced a new frame
VD_WAIT = 3, // no EOF, but no output; wait until wakeup
video: try harder to decode cover art picture only once For cover art, we pretend that the video stream is infinite, but also stop decoding once we have an image on the VO (this seems advantageous for the case when strange filters are inserted or the VO image gets lost). Since a while ago, the video chain started decoding 2 images though ("Non-monotonic video pts: 0.000000 <= 0.000000"), which is annoying and wasteful. Improve this by handling a certain corner case at initialization, which will decode a second image while the first one is still stuck in the filter chain. Also, just in case there are filters which buffer a lot, also force EOF filtering (which means we tell the filters to flush buffered frames). CC: @mpv-player/stable
2014-10-09 15:57:17 +00:00
VD_RECONFIG = 4,
player: move video display code out of the playloop Basically move the code from playloop.c to video.c. The new function write_video() now contains the code that was part of run_playloop(). There are no functional changes, except handling "new_frame_shown" slightly differently. This is done so that we don't need new a new MPContext field or a return value for write_video() to signal this condition. Instead, it's handled indirectly.
2014-07-30 21:24:08 +00:00
};
static const char av_desync_help_text[] =
player: slim down A/V desync warning I don't think most of these suggestions are overly helpful. Just get rid of them.
2015-06-30 17:06:00 +00:00
"\n"
"Audio/Video desynchronisation detected! Possible reasons include too slow\n"
"hardware, temporary CPU spikes, broken drivers, and broken files. Audio\n"
"position will not match to the video (see A-V status field).\n"
"\n";
Split mplayer.c mplayer.c was a bit too big. Split it into multiple files. I hope the way it's split makes sense. Maybe some things don't make too much sense, or go against intuition. These will fixed as soon as I notice them. Some files are a bit questionable (misc.c, osd.c, configfiles.c), and suggestions how to organize this better are welcome. Regressions are possible due to reorganized include statements. Obviously I didn't just copy mplayer.c's orgy of include statements, but recreated them for each file. It's easily possible that there are oversights and mistakes, which will show up on other platforms. There is one actual change: the public avutil.h include is removed from encode.h, and I tried to replace most FFMIN/FFMAX/av_clip uses. I consider using libavutil too much as dangerous, because the set of include files they recursively pull in is rather arbitrary and is different between FFmpeg and Libav.
2013-10-29 21:38:29 +00:00
player: disable video equalizer frontend code for WIP LGPL mode Nick and kiriuja could not be reached, and created/changed this in 92c5c274, 6441a5ad, bffd4007, 555c6766, c2c997fd. The video equalizer stuff was redone fully later, but there are still parts that look too similar and basically use the same approach. I'm more comfortable with declaring it GPL only for now. I plan to redo them later in a way that will remove copyright.
2017-06-23 13:13:47 +00:00
#if HAVE_GPL
video: decouple filtering/decoding slightly more Lots of noise to remove the vfilter/vo fields from dec_video. From now on, video filtering and output will still be done together, summarized under struct vo_chain. There is the question where exactly the vf_chain should go in such a decoupled architecture. The end goal is being able to place a "complex" filter between video decoders and output (which will culminate in natural integration of A->V filters for natural integration of libavfilter audio visualizations). The vf_chain is still useful for "final" processing, such as format conversions and deinterlacing. Also, there's only 1 VO and 1 --vf option. So having 1 vf_chain for a VO seems ideal, since otherwise there would be no natural way to handle all these existing options and mechanisms. There is still some work required to truly decouple decoding.
2016-01-13 23:18:48 +00:00
int video_set_colors(struct vo_chain *vo_c, const char *item, int value)
video: refactor: shuffle code around struct dec_video should have nothing to do with video filters or outputs, and this huge chunk of code was somehow stuck directly in dec_video.c.
2016-01-13 23:18:36 +00:00
{
vf_equalizer_t data;
data.item = item;
data.value = value;
video: decouple filtering/decoding slightly more Lots of noise to remove the vfilter/vo fields from dec_video. From now on, video filtering and output will still be done together, summarized under struct vo_chain. There is the question where exactly the vf_chain should go in such a decoupled architecture. The end goal is being able to place a "complex" filter between video decoders and output (which will culminate in natural integration of A->V filters for natural integration of libavfilter audio visualizations). The vf_chain is still useful for "final" processing, such as format conversions and deinterlacing. Also, there's only 1 VO and 1 --vf option. So having 1 vf_chain for a VO seems ideal, since otherwise there would be no natural way to handle all these existing options and mechanisms. There is still some work required to truly decouple decoding.
2016-01-13 23:18:48 +00:00
MP_VERBOSE(vo_c, "set video colors %s=%d \n", item, value);
if (video_vf_vo_control(vo_c, VFCTRL_SET_EQUALIZER, &data) == CONTROL_TRUE)
return 1;
MP_VERBOSE(vo_c, "Video attribute '%s' is not supported by selected vo.\n",
video: refactor: shuffle code around struct dec_video should have nothing to do with video filters or outputs, and this huge chunk of code was somehow stuck directly in dec_video.c.
2016-01-13 23:18:36 +00:00
item);
return 0;
}
video: decouple filtering/decoding slightly more Lots of noise to remove the vfilter/vo fields from dec_video. From now on, video filtering and output will still be done together, summarized under struct vo_chain. There is the question where exactly the vf_chain should go in such a decoupled architecture. The end goal is being able to place a "complex" filter between video decoders and output (which will culminate in natural integration of A->V filters for natural integration of libavfilter audio visualizations). The vf_chain is still useful for "final" processing, such as format conversions and deinterlacing. Also, there's only 1 VO and 1 --vf option. So having 1 vf_chain for a VO seems ideal, since otherwise there would be no natural way to handle all these existing options and mechanisms. There is still some work required to truly decouple decoding.
2016-01-13 23:18:48 +00:00
int video_get_colors(struct vo_chain *vo_c, const char *item, int *value)
video: refactor: shuffle code around struct dec_video should have nothing to do with video filters or outputs, and this huge chunk of code was somehow stuck directly in dec_video.c.
2016-01-13 23:18:36 +00:00
{
vf_equalizer_t data;
data.item = item;
video: decouple filtering/decoding slightly more Lots of noise to remove the vfilter/vo fields from dec_video. From now on, video filtering and output will still be done together, summarized under struct vo_chain. There is the question where exactly the vf_chain should go in such a decoupled architecture. The end goal is being able to place a "complex" filter between video decoders and output (which will culminate in natural integration of A->V filters for natural integration of libavfilter audio visualizations). The vf_chain is still useful for "final" processing, such as format conversions and deinterlacing. Also, there's only 1 VO and 1 --vf option. So having 1 vf_chain for a VO seems ideal, since otherwise there would be no natural way to handle all these existing options and mechanisms. There is still some work required to truly decouple decoding.
2016-01-13 23:18:48 +00:00
MP_VERBOSE(vo_c, "get video colors %s \n", item);
if (video_vf_vo_control(vo_c, VFCTRL_GET_EQUALIZER, &data) == CONTROL_TRUE) {
*value = data.value;
return 1;
video: refactor: shuffle code around struct dec_video should have nothing to do with video filters or outputs, and this huge chunk of code was somehow stuck directly in dec_video.c.
2016-01-13 23:18:36 +00:00
}
return 0;
}
player: disable video equalizer frontend code for WIP LGPL mode Nick and kiriuja could not be reached, and created/changed this in 92c5c274, 6441a5ad, bffd4007, 555c6766, c2c997fd. The video equalizer stuff was redone fully later, but there are still parts that look too similar and basically use the same approach. I'm more comfortable with declaring it GPL only for now. I plan to redo them later in a way that will remove copyright.
2017-06-23 13:13:47 +00:00
#endif
video: refactor: shuffle code around struct dec_video should have nothing to do with video filters or outputs, and this huge chunk of code was somehow stuck directly in dec_video.c.
2016-01-13 23:18:36 +00:00
// Send a VCTRL, or if it doesn't work, translate it to a VOCTRL and try the VO.
video: decouple filtering/decoding slightly more Lots of noise to remove the vfilter/vo fields from dec_video. From now on, video filtering and output will still be done together, summarized under struct vo_chain. There is the question where exactly the vf_chain should go in such a decoupled architecture. The end goal is being able to place a "complex" filter between video decoders and output (which will culminate in natural integration of A->V filters for natural integration of libavfilter audio visualizations). The vf_chain is still useful for "final" processing, such as format conversions and deinterlacing. Also, there's only 1 VO and 1 --vf option. So having 1 vf_chain for a VO seems ideal, since otherwise there would be no natural way to handle all these existing options and mechanisms. There is still some work required to truly decouple decoding.
2016-01-13 23:18:48 +00:00
int video_vf_vo_control(struct vo_chain *vo_c, int vf_cmd, void *data)
video: refactor: shuffle code around struct dec_video should have nothing to do with video filters or outputs, and this huge chunk of code was somehow stuck directly in dec_video.c.
2016-01-13 23:18:36 +00:00
{
video: decouple filtering/decoding slightly more Lots of noise to remove the vfilter/vo fields from dec_video. From now on, video filtering and output will still be done together, summarized under struct vo_chain. There is the question where exactly the vf_chain should go in such a decoupled architecture. The end goal is being able to place a "complex" filter between video decoders and output (which will culminate in natural integration of A->V filters for natural integration of libavfilter audio visualizations). The vf_chain is still useful for "final" processing, such as format conversions and deinterlacing. Also, there's only 1 VO and 1 --vf option. So having 1 vf_chain for a VO seems ideal, since otherwise there would be no natural way to handle all these existing options and mechanisms. There is still some work required to truly decouple decoding.
2016-01-13 23:18:48 +00:00
if (vo_c->vf->initialized > 0) {
int r = vf_control_any(vo_c->vf, vf_cmd, data);
video: refactor: shuffle code around struct dec_video should have nothing to do with video filters or outputs, and this huge chunk of code was somehow stuck directly in dec_video.c.
2016-01-13 23:18:36 +00:00
if (r != CONTROL_UNKNOWN)
return r;
}
switch (vf_cmd) {
case VFCTRL_SET_EQUALIZER: {
vf_equalizer_t *eq = data;
video: decouple filtering/decoding slightly more Lots of noise to remove the vfilter/vo fields from dec_video. From now on, video filtering and output will still be done together, summarized under struct vo_chain. There is the question where exactly the vf_chain should go in such a decoupled architecture. The end goal is being able to place a "complex" filter between video decoders and output (which will culminate in natural integration of A->V filters for natural integration of libavfilter audio visualizations). The vf_chain is still useful for "final" processing, such as format conversions and deinterlacing. Also, there's only 1 VO and 1 --vf option. So having 1 vf_chain for a VO seems ideal, since otherwise there would be no natural way to handle all these existing options and mechanisms. There is still some work required to truly decouple decoding.
2016-01-13 23:18:48 +00:00
if (!vo_c->vo->config_ok)
video: refactor: shuffle code around struct dec_video should have nothing to do with video filters or outputs, and this huge chunk of code was somehow stuck directly in dec_video.c.
2016-01-13 23:18:36 +00:00
return CONTROL_FALSE; // vo not configured?
struct voctrl_set_equalizer_args param = {
eq->item, eq->value
};
video: decouple filtering/decoding slightly more Lots of noise to remove the vfilter/vo fields from dec_video. From now on, video filtering and output will still be done together, summarized under struct vo_chain. There is the question where exactly the vf_chain should go in such a decoupled architecture. The end goal is being able to place a "complex" filter between video decoders and output (which will culminate in natural integration of A->V filters for natural integration of libavfilter audio visualizations). The vf_chain is still useful for "final" processing, such as format conversions and deinterlacing. Also, there's only 1 VO and 1 --vf option. So having 1 vf_chain for a VO seems ideal, since otherwise there would be no natural way to handle all these existing options and mechanisms. There is still some work required to truly decouple decoding.
2016-01-13 23:18:48 +00:00
return vo_control(vo_c->vo, VOCTRL_SET_EQUALIZER, &param) == VO_TRUE;
video: refactor: shuffle code around struct dec_video should have nothing to do with video filters or outputs, and this huge chunk of code was somehow stuck directly in dec_video.c.
2016-01-13 23:18:36 +00:00
}
case VFCTRL_GET_EQUALIZER: {
vf_equalizer_t *eq = data;
video: decouple filtering/decoding slightly more Lots of noise to remove the vfilter/vo fields from dec_video. From now on, video filtering and output will still be done together, summarized under struct vo_chain. There is the question where exactly the vf_chain should go in such a decoupled architecture. The end goal is being able to place a "complex" filter between video decoders and output (which will culminate in natural integration of A->V filters for natural integration of libavfilter audio visualizations). The vf_chain is still useful for "final" processing, such as format conversions and deinterlacing. Also, there's only 1 VO and 1 --vf option. So having 1 vf_chain for a VO seems ideal, since otherwise there would be no natural way to handle all these existing options and mechanisms. There is still some work required to truly decouple decoding.
2016-01-13 23:18:48 +00:00
if (!vo_c->vo->config_ok)
video: refactor: shuffle code around struct dec_video should have nothing to do with video filters or outputs, and this huge chunk of code was somehow stuck directly in dec_video.c.
2016-01-13 23:18:36 +00:00
return CONTROL_FALSE; // vo not configured?
struct voctrl_get_equalizer_args param = {
eq->item, &eq->value
};
video: decouple filtering/decoding slightly more Lots of noise to remove the vfilter/vo fields from dec_video. From now on, video filtering and output will still be done together, summarized under struct vo_chain. There is the question where exactly the vf_chain should go in such a decoupled architecture. The end goal is being able to place a "complex" filter between video decoders and output (which will culminate in natural integration of A->V filters for natural integration of libavfilter audio visualizations). The vf_chain is still useful for "final" processing, such as format conversions and deinterlacing. Also, there's only 1 VO and 1 --vf option. So having 1 vf_chain for a VO seems ideal, since otherwise there would be no natural way to handle all these existing options and mechanisms. There is still some work required to truly decouple decoding.
2016-01-13 23:18:48 +00:00
return vo_control(vo_c->vo, VOCTRL_GET_EQUALIZER, &param) == VO_TRUE;
video: refactor: shuffle code around struct dec_video should have nothing to do with video filters or outputs, and this huge chunk of code was somehow stuck directly in dec_video.c.
2016-01-13 23:18:36 +00:00
}
}
return CONTROL_UNKNOWN;
}
video: decouple filtering/decoding slightly more Lots of noise to remove the vfilter/vo fields from dec_video. From now on, video filtering and output will still be done together, summarized under struct vo_chain. There is the question where exactly the vf_chain should go in such a decoupled architecture. The end goal is being able to place a "complex" filter between video decoders and output (which will culminate in natural integration of A->V filters for natural integration of libavfilter audio visualizations). The vf_chain is still useful for "final" processing, such as format conversions and deinterlacing. Also, there's only 1 VO and 1 --vf option. So having 1 vf_chain for a VO seems ideal, since otherwise there would be no natural way to handle all these existing options and mechanisms. There is still some work required to truly decouple decoding.
2016-01-13 23:18:48 +00:00
static void set_allowed_vo_formats(struct vo_chain *vo_c)
video: move VO reinit from filter chain to player This gets rid of the vf_vo pseudo-filter. It ends the idea of MPlayer's architecture that the VO is just a (terminating) video filter. It didn't really work for us with respect to video timing (the "end" of the video chain isn't really made for video timing, and making it do so would be awkward), and now we're removing it entirely. We will be able to fix some things, such as properly draining video on reconfiguration.
2013-12-10 18:24:58 +00:00
{
video: decouple filtering/decoding slightly more Lots of noise to remove the vfilter/vo fields from dec_video. From now on, video filtering and output will still be done together, summarized under struct vo_chain. There is the question where exactly the vf_chain should go in such a decoupled architecture. The end goal is being able to place a "complex" filter between video decoders and output (which will culminate in natural integration of A->V filters for natural integration of libavfilter audio visualizations). The vf_chain is still useful for "final" processing, such as format conversions and deinterlacing. Also, there's only 1 VO and 1 --vf option. So having 1 vf_chain for a VO seems ideal, since otherwise there would be no natural way to handle all these existing options and mechanisms. There is still some work required to truly decouple decoding.
2016-01-13 23:18:48 +00:00
vo_query_formats(vo_c->vo, vo_c->vf->allowed_output_formats);
video: move VO reinit from filter chain to player This gets rid of the vf_vo pseudo-filter. It ends the idea of MPlayer's architecture that the VO is just a (terminating) video filter. It didn't really work for us with respect to video timing (the "end" of the video chain isn't really made for video timing, and making it do so would be awkward), and now we're removing it entirely. We will be able to fix some things, such as properly draining video on reconfiguration.
2013-12-10 18:24:58 +00:00
}
player: fix breakage when combining 3D and rotate auto-filters This would get stuck in reconfiguring the filter chain forever, because params was mutated ("params.rotate = 0;"). This was used as input for vf_reconfig(), but the filter chain input must always be equivalent to the decoder output, or filter chain reconfiguration will be triggered. The line of code to reset the rotation is from a time when this used to work differently. Also remove the unnecessary try_filter() parameter.
2016-03-28 17:39:33 +00:00
static int try_filter(struct vo_chain *vo_c, char *name, char *label, char **args)
video: auto-insert software rotation filter If the VO can't do rotation, insert a filter to do this. Note that this doesn't reuse the filter insertion code from command.c (used by "vf" input command), because that would end up more complicated: we don't even want to change the user filter option.
2014-04-20 21:54:13 +00:00
{
video: decouple filtering/decoding slightly more Lots of noise to remove the vfilter/vo fields from dec_video. From now on, video filtering and output will still be done together, summarized under struct vo_chain. There is the question where exactly the vf_chain should go in such a decoupled architecture. The end goal is being able to place a "complex" filter between video decoders and output (which will culminate in natural integration of A->V filters for natural integration of libavfilter audio visualizations). The vf_chain is still useful for "final" processing, such as format conversions and deinterlacing. Also, there's only 1 VO and 1 --vf option. So having 1 vf_chain for a VO seems ideal, since otherwise there would be no natural way to handle all these existing options and mechanisms. There is still some work required to truly decouple decoding.
2016-01-13 23:18:48 +00:00
struct vf_instance *vf = vf_append_filter(vo_c->vf, name, args);
video: auto-insert software rotation filter If the VO can't do rotation, insert a filter to do this. Note that this doesn't reuse the filter insertion code from command.c (used by "vf" input command), because that would end up more complicated: we don't even want to change the user filter option.
2014-04-20 21:54:13 +00:00
if (!vf)
return -1;
vf->label = talloc_strdup(vf, label);
player: fix breakage when combining 3D and rotate auto-filters This would get stuck in reconfiguring the filter chain forever, because params was mutated ("params.rotate = 0;"). This was used as input for vf_reconfig(), but the filter chain input must always be equivalent to the decoder output, or filter chain reconfiguration will be triggered. The line of code to reset the rotation is from a time when this used to work differently. Also remove the unnecessary try_filter() parameter.
2016-03-28 17:39:33 +00:00
if (vf_reconfig(vo_c->vf, &vo_c->input_format) < 0) {
video: decouple filtering/decoding slightly more Lots of noise to remove the vfilter/vo fields from dec_video. From now on, video filtering and output will still be done together, summarized under struct vo_chain. There is the question where exactly the vf_chain should go in such a decoupled architecture. The end goal is being able to place a "complex" filter between video decoders and output (which will culminate in natural integration of A->V filters for natural integration of libavfilter audio visualizations). The vf_chain is still useful for "final" processing, such as format conversions and deinterlacing. Also, there's only 1 VO and 1 --vf option. So having 1 vf_chain for a VO seems ideal, since otherwise there would be no natural way to handle all these existing options and mechanisms. There is still some work required to truly decouple decoding.
2016-01-13 23:18:48 +00:00
vf_remove_filter(vo_c->vf, vf);
video: auto-insert software rotation filter If the VO can't do rotation, insert a filter to do this. Note that this doesn't reuse the filter insertion code from command.c (used by "vf" input command), because that would end up more complicated: we don't even want to change the user filter option.
2014-04-20 21:54:13 +00:00
// restore
player: fix breakage when combining 3D and rotate auto-filters This would get stuck in reconfiguring the filter chain forever, because params was mutated ("params.rotate = 0;"). This was used as input for vf_reconfig(), but the filter chain input must always be equivalent to the decoder output, or filter chain reconfiguration will be triggered. The line of code to reset the rotation is from a time when this used to work differently. Also remove the unnecessary try_filter() parameter.
2016-03-28 17:39:33 +00:00
vf_reconfig(vo_c->vf, &vo_c->input_format);
video: auto-insert software rotation filter If the VO can't do rotation, insert a filter to do this. Note that this doesn't reuse the filter insertion code from command.c (used by "vf" input command), because that would end up more complicated: we don't even want to change the user filter option.
2014-04-20 21:54:13 +00:00
return -1;
}
return 0;
}
player: rewrite deinterlace filter auto-insertion Instead of using the "vf" command code (which changes filters at runtime on user input), use the general filter-insertion code. The latter was added later, and is more suitable for automatically inserted filters. The old code failed in particular when using watch-later saving, which stored the filter list in the resume config file. If a user changed the hardware decoding mode via command line, the stored filter chain was out of date and could cause failure due to not working with hardware or software decoding mode. Storing the deinterlace filter in the filter list was unavoidable, because it was part of the user state. (The new code only edits the actually instantiated filters.)
2016-07-05 19:10:26 +00:00
static bool check_output_format(struct vo_chain *vo_c, int imgfmt)
{
return vo_c->vf->output_params.imgfmt == imgfmt;
}
static int probe_deint_filters(struct vo_chain *vo_c)
{
player: disable deinterlace property for WIP LGPL mode cehoyos has not agreed to the LGPL relicensing. He added the deinterlace property in commit 7b25afd7. Make it GPL-only for now. The still working parts of the --deinterlace option are not affected by his copyright.
2017-06-23 13:09:21 +00:00
#if HAVE_GPL
video: fix deinterlace filter handling for VFCTRL_SET_DEINTERLACE filters Some filters support VFCTRL_SET_DEINTERLACE. This affects most hardware deinterlace filters. They can be inserted by the user manually, or auto- inserted by vf.c itself as conversion filter (vf_d3d11vpp). In these cases, we shouldn't insert or remove filters outselves, and instead VFCTRL_SET_DEINTERLACE should be invoked to switch the mode. This wasn't done correctly in the recently refactored code and could have broken with --deinterlace. (The refactor only considered switching via property in this case.) Fix it by making it a proper part of the filter_reconfig() function, and making set_deinterlacing() (which is called by the property handler) merely call filter_reconfig() in all cases to do the real work. We can even avoid rebuilding the filter chain - though only if no other auto-filters are inserted. It probably also provides a slightly cleaner way to implement functionality in the VO while still inserting video filter fallbacks correctly if required.
2016-07-06 12:08:47 +00:00
// Usually, we prefer inserting/removing deint filters. But If there's VO
// support, or the user inserted a filter that supports swichting deint and
// that has no VF_DEINTERLACE_LABEL, or if the filter was auto-inserted
// for other reasons and supports switching deint (like vf_d3d11vpp), then
// use the runtime switching method.
if (video_vf_vo_control(vo_c, VFCTRL_SET_DEINTERLACE, &(int){1}) == CONTROL_OK)
return 0;
player: disable deinterlace property for WIP LGPL mode cehoyos has not agreed to the LGPL relicensing. He added the deinterlace property in commit 7b25afd7. Make it GPL-only for now. The still working parts of the --deinterlace option are not affected by his copyright.
2017-06-23 13:09:21 +00:00
#endif
video: fix deinterlace filter handling for VFCTRL_SET_DEINTERLACE filters Some filters support VFCTRL_SET_DEINTERLACE. This affects most hardware deinterlace filters. They can be inserted by the user manually, or auto- inserted by vf.c itself as conversion filter (vf_d3d11vpp). In these cases, we shouldn't insert or remove filters outselves, and instead VFCTRL_SET_DEINTERLACE should be invoked to switch the mode. This wasn't done correctly in the recently refactored code and could have broken with --deinterlace. (The refactor only considered switching via property in this case.) Fix it by making it a proper part of the filter_reconfig() function, and making set_deinterlacing() (which is called by the property handler) merely call filter_reconfig() in all cases to do the real work. We can even avoid rebuilding the filter chain - though only if no other auto-filters are inserted. It probably also provides a slightly cleaner way to implement functionality in the VO while still inserting video filter fallbacks correctly if required.
2016-07-06 12:08:47 +00:00
player: rewrite deinterlace filter auto-insertion Instead of using the "vf" command code (which changes filters at runtime on user input), use the general filter-insertion code. The latter was added later, and is more suitable for automatically inserted filters. The old code failed in particular when using watch-later saving, which stored the filter list in the resume config file. If a user changed the hardware decoding mode via command line, the stored filter chain was out of date and could cause failure due to not working with hardware or software decoding mode. Storing the deinterlace filter in the filter list was unavoidable, because it was part of the user state. (The new code only edits the actually instantiated filters.)
2016-07-05 19:10:26 +00:00
if (check_output_format(vo_c, IMGFMT_VDPAU)) {
char *args[5] = {"deint", "yes"};
int pref = 0;
vo_control(vo_c->vo, VOCTRL_GET_PREF_DEINT, &pref);
pref = pref < 0 ? -pref : pref;
if (pref > 0 && pref <= 4) {
const char *types[] =
{"", "first-field", "bob", "temporal", "temporal-spatial"};
args[2] = "deint-mode";
args[3] = (char *)types[pref];
}
return try_filter(vo_c, "vdpaupp", VF_DEINTERLACE_LABEL, args);
}
if (check_output_format(vo_c, IMGFMT_VAAPI))
return try_filter(vo_c, "vavpp", VF_DEINTERLACE_LABEL, NULL);
if (check_output_format(vo_c, IMGFMT_D3D11VA) ||
check_output_format(vo_c, IMGFMT_D3D11NV12))
return try_filter(vo_c, "d3d11vpp", VF_DEINTERLACE_LABEL, NULL);
video: deprecate almost all video filters The plan is to nuke the custom filter chain completely. It's not clear what will happen to the still needed builtin filters (mostly hardware deinterlacing and vf_vapoursynth). Most likely we'll replace them with different filter chain concept (whose main purpose will be providing builtin things and bridging to libavfilter). The undocumented "warn" options are there to disable deprecation warnings when the player inserts filter automatically. The same will be done to audio filters, at a later point.
2017-04-02 15:59:45 +00:00
char *args[] = {"warn", "no", NULL};
return try_filter(vo_c, "yadif", VF_DEINTERLACE_LABEL, args);
player: rewrite deinterlace filter auto-insertion Instead of using the "vf" command code (which changes filters at runtime on user input), use the general filter-insertion code. The latter was added later, and is more suitable for automatically inserted filters. The old code failed in particular when using watch-later saving, which stored the filter list in the resume config file. If a user changed the hardware decoding mode via command line, the stored filter chain was out of date and could cause failure due to not working with hardware or software decoding mode. Storing the deinterlace filter in the filter list was unavoidable, because it was part of the user state. (The new code only edits the actually instantiated filters.)
2016-07-05 19:10:26 +00:00
}
video: refactor: disentangle decoding/filtering some more This moves some code related to decoding from video.c to dec_video.c, and also removes some accesses to dec_video.c from the filtering code. dec_video.ch is starting to make sense, and simply returns video frames from a demuxer stream. The API exposed is also somewhat intended to be easily changeable to move decoding to a separate thread, if we ever want this (due to libavcodec already being threaded, I don't see much of a reason, but it might still be helpful).
2016-01-16 20:19:52 +00:00
// Reconfigure the filter chain according to the new input format.
player: rewrite deinterlace filter auto-insertion Instead of using the "vf" command code (which changes filters at runtime on user input), use the general filter-insertion code. The latter was added later, and is more suitable for automatically inserted filters. The old code failed in particular when using watch-later saving, which stored the filter list in the resume config file. If a user changed the hardware decoding mode via command line, the stored filter chain was out of date and could cause failure due to not working with hardware or software decoding mode. Storing the deinterlace filter in the filter list was unavoidable, because it was part of the user state. (The new code only edits the actually instantiated filters.)
2016-07-05 19:10:26 +00:00
static void filter_reconfig(struct MPContext *mpctx, struct vo_chain *vo_c)
video: move VO reinit from filter chain to player This gets rid of the vf_vo pseudo-filter. It ends the idea of MPlayer's architecture that the VO is just a (terminating) video filter. It didn't really work for us with respect to video timing (the "end" of the video chain isn't really made for video timing, and making it do so would be awkward), and now we're removing it entirely. We will be able to fix some things, such as properly draining video on reconfiguration.
2013-12-10 18:24:58 +00:00
{
video: refactor: disentangle decoding/filtering some more This moves some code related to decoding from video.c to dec_video.c, and also removes some accesses to dec_video.c from the filtering code. dec_video.ch is starting to make sense, and simply returns video frames from a demuxer stream. The API exposed is also somewhat intended to be easily changeable to move decoding to a separate thread, if we ever want this (due to libavcodec already being threaded, I don't see much of a reason, but it might still be helpful).
2016-01-16 20:19:52 +00:00
struct mp_image_params params = vo_c->input_format;
if (!params.imgfmt)
return;
video: decouple filtering/decoding slightly more Lots of noise to remove the vfilter/vo fields from dec_video. From now on, video filtering and output will still be done together, summarized under struct vo_chain. There is the question where exactly the vf_chain should go in such a decoupled architecture. The end goal is being able to place a "complex" filter between video decoders and output (which will culminate in natural integration of A->V filters for natural integration of libavfilter audio visualizations). The vf_chain is still useful for "final" processing, such as format conversions and deinterlacing. Also, there's only 1 VO and 1 --vf option. So having 1 vf_chain for a VO seems ideal, since otherwise there would be no natural way to handle all these existing options and mechanisms. There is still some work required to truly decouple decoding.
2016-01-13 23:18:48 +00:00
set_allowed_vo_formats(vo_c);
video: move VO reinit from filter chain to player This gets rid of the vf_vo pseudo-filter. It ends the idea of MPlayer's architecture that the VO is just a (terminating) video filter. It didn't really work for us with respect to video timing (the "end" of the video chain isn't really made for video timing, and making it do so would be awkward), and now we're removing it entirely. We will be able to fix some things, such as properly draining video on reconfiguration.
2013-12-10 18:24:58 +00:00
player: rewrite deinterlace filter auto-insertion Instead of using the "vf" command code (which changes filters at runtime on user input), use the general filter-insertion code. The latter was added later, and is more suitable for automatically inserted filters. The old code failed in particular when using watch-later saving, which stored the filter list in the resume config file. If a user changed the hardware decoding mode via command line, the stored filter chain was out of date and could cause failure due to not working with hardware or software decoding mode. Storing the deinterlace filter in the filter list was unavoidable, because it was part of the user state. (The new code only edits the actually instantiated filters.)
2016-07-05 19:10:26 +00:00
char *filters[] = {"autorotate", "autostereo3d", "deinterlace", NULL};
player: remove auto-inserted filters before adding them again Makes certain cases of runtime changes actually work. Also change the label for the stereo3d filter and make it consistent with the rotate one.
2016-03-28 17:36:11 +00:00
for (int n = 0; filters[n]; n++) {
struct vf_instance *vf = vf_find_by_label(vo_c->vf, filters[n]);
video: fix deinterlace filter handling on pixel format changes The test scenario at hand was hardware decoding a file with d3d11 and with deinterlacing enabled. The file switches to a non-hardware dedocdeable format mid-stream. This failed, because it tried to call vf_reconfig() with the old filters inserted, with was fatal due to vf_d3d11vpp accepting only hardware input formats. Fix this by always strictly removing all auto-inserted filters (including the deinterlacing one), and reconfiguring only after that. Note that this change is good for other situations too, because we generally don't want to use a hardware deinterlacer for software decoding by default. They're not necessarily optimal, and VAAPI VPP even has incomprehensible deinterlacer bugs specifically with software frames not coming from a hardware decoder.
2016-07-06 11:38:43 +00:00
if (vf)
player: remove auto-inserted filters before adding them again Makes certain cases of runtime changes actually work. Also change the label for the stereo3d filter and make it consistent with the rotate one.
2016-03-28 17:36:11 +00:00
vf_remove_filter(vo_c->vf, vf);
}
video: fix deinterlace filter handling for VFCTRL_SET_DEINTERLACE filters Some filters support VFCTRL_SET_DEINTERLACE. This affects most hardware deinterlace filters. They can be inserted by the user manually, or auto- inserted by vf.c itself as conversion filter (vf_d3d11vpp). In these cases, we shouldn't insert or remove filters outselves, and instead VFCTRL_SET_DEINTERLACE should be invoked to switch the mode. This wasn't done correctly in the recently refactored code and could have broken with --deinterlace. (The refactor only considered switching via property in this case.) Fix it by making it a proper part of the filter_reconfig() function, and making set_deinterlacing() (which is called by the property handler) merely call filter_reconfig() in all cases to do the real work. We can even avoid rebuilding the filter chain - though only if no other auto-filters are inserted. It probably also provides a slightly cleaner way to implement functionality in the VO while still inserting video filter fallbacks correctly if required.
2016-07-06 12:08:47 +00:00
if (vo_c->vf->initialized < 1) {
if (vf_reconfig(vo_c->vf, &params) < 0)
return;
}
player: disable deinterlace property for WIP LGPL mode cehoyos has not agreed to the LGPL relicensing. He added the deinterlace property in commit 7b25afd7. Make it GPL-only for now. The still working parts of the --deinterlace option are not affected by his copyright.
2017-06-23 13:09:21 +00:00
#if HAVE_GPL
video: fix deinterlace filter handling for VFCTRL_SET_DEINTERLACE filters Some filters support VFCTRL_SET_DEINTERLACE. This affects most hardware deinterlace filters. They can be inserted by the user manually, or auto- inserted by vf.c itself as conversion filter (vf_d3d11vpp). In these cases, we shouldn't insert or remove filters outselves, and instead VFCTRL_SET_DEINTERLACE should be invoked to switch the mode. This wasn't done correctly in the recently refactored code and could have broken with --deinterlace. (The refactor only considered switching via property in this case.) Fix it by making it a proper part of the filter_reconfig() function, and making set_deinterlacing() (which is called by the property handler) merely call filter_reconfig() in all cases to do the real work. We can even avoid rebuilding the filter chain - though only if no other auto-filters are inserted. It probably also provides a slightly cleaner way to implement functionality in the VO while still inserting video filter fallbacks correctly if required.
2016-07-06 12:08:47 +00:00
// Make sure to reset this even if runtime deint switching is used.
video: respect --deinterlace=auto --deinterlace=auto is the default, and has the obscure semantics that deinterlacing is disabled, unless the user has manually inserted a deinterlacing filter. While in software decoding this doesn't matter, and we will happily insert 2 yadif filters (if the user has already added one), or not remove the yadif filter (if deinterlacing is disabled, but the user has added the filter manually), this is different with hardware deinterlacer filters. These support VFCTRL_SET_DEINTERLACE for toggling deinterlacing filtering at runtime. It exists mainly for legacy reasons, and possibly because it makes switching deinterlacing modes more efficient. It might also gives us an entry-point for VO deinterlacing, maybe. For whatever reasons this mechanism exists, we still support and use it. This commit fixes that video.c always used VFCTRL_SET_DEINTERLACE to disable deinterlacing, even if --deinterlace=auto was set. Fix this by checking the value of the option directly.
2016-07-12 17:35:47 +00:00
if (mpctx->opts->deinterlace >= 0)
video_vf_vo_control(vo_c, VFCTRL_SET_DEINTERLACE, &(int){0});
player: disable deinterlace property for WIP LGPL mode cehoyos has not agreed to the LGPL relicensing. He added the deinterlace property in commit 7b25afd7. Make it GPL-only for now. The still working parts of the --deinterlace option are not affected by his copyright.
2017-06-23 13:09:21 +00:00
#endif
video: fix deinterlace filter handling on pixel format changes The test scenario at hand was hardware decoding a file with d3d11 and with deinterlacing enabled. The file switches to a non-hardware dedocdeable format mid-stream. This failed, because it tried to call vf_reconfig() with the old filters inserted, with was fatal due to vf_d3d11vpp accepting only hardware input formats. Fix this by always strictly removing all auto-inserted filters (including the deinterlacing one), and reconfiguring only after that. Note that this change is good for other situations too, because we generally don't want to use a hardware deinterlacer for software decoding by default. They're not necessarily optimal, and VAAPI VPP even has incomprehensible deinterlacer bugs specifically with software frames not coming from a hardware decoder.
2016-07-06 11:38:43 +00:00
vf_rotate: allow arbitrary rotation vf_rotate selects the correct filter for 90° rotation, but it can be extended to use lavfi's vf_rotate as fallback. See #3434.
2016-08-19 07:37:52 +00:00
if (params.rotate) {
if (!(vo_c->vo->driver->caps & VO_CAP_ROTATE90) || params.rotate % 90) {
video: auto-insert software rotation filter If the VO can't do rotation, insert a filter to do this. Note that this doesn't reuse the filter insertion code from command.c (used by "vf" input command), because that would end up more complicated: we don't even want to change the user filter option.
2014-04-20 21:54:13 +00:00
// Try to insert a rotation filter.
video: deprecate almost all video filters The plan is to nuke the custom filter chain completely. It's not clear what will happen to the still needed builtin filters (mostly hardware deinterlacing and vf_vapoursynth). Most likely we'll replace them with different filter chain concept (whose main purpose will be providing builtin things and bridging to libavfilter). The undocumented "warn" options are there to disable deprecation warnings when the player inserts filter automatically. The same will be done to audio filters, at a later point.
2017-04-02 15:59:45 +00:00
char *args[] = {"angle", "auto", "warn", "no", NULL};
player: fix breakage when combining 3D and rotate auto-filters This would get stuck in reconfiguring the filter chain forever, because params was mutated ("params.rotate = 0;"). This was used as input for vf_reconfig(), but the filter chain input must always be equivalent to the decoder output, or filter chain reconfiguration will be triggered. The line of code to reset the rotation is from a time when this used to work differently. Also remove the unnecessary try_filter() parameter.
2016-03-28 17:39:33 +00:00
if (try_filter(vo_c, "rotate", "autorotate", args) < 0)
video: refactor: slightly disentangle video filtering
2016-01-15 21:23:01 +00:00
MP_ERR(vo_c, "Can't insert rotation filter.\n");
video: auto-insert software rotation filter If the VO can't do rotation, insert a filter to do this. Note that this doesn't reuse the filter insertion code from command.c (used by "vf" input command), because that would end up more complicated: we don't even want to change the user filter option.
2014-04-20 21:54:13 +00:00
}
}
video: initial Matroska 3D support This inserts an automatic conversion filter if a Matroska file is marked as 3D (StereoMode element). The basic idea is similar to video rotation and colorspace handling: the 3D mode is added as a property to the video params. Depending on this property, a video filter can be inserted. As of this commit, extending mp_image_params is actually completely unnecessary - but the idea is that it will make it easier to integrate with VOs supporting stereo 3D mogrification. Although vo_opengl does support some stereo rendering, it didn't support the mode my sample file used, so I'll leave that part for later. Not that most mappings from Matroska mode to vf_stereo3d mode are probably wrong, and some are missing. Assuming that Matroska modes, and vf_stereo3d in modes, and out modes are all the same might be an oversimplification - we'll see. See issue #1045.
2014-08-30 21:24:46 +00:00
if (params.stereo_in != params.stereo_out &&
params.stereo_in > 0 && params.stereo_out >= 0)
{
video: cleanup stereo mode parsing Use OPT_CHOICE_C() instead of the custom parser. The functionality is pretty much equivalent. (On a side note, it seems --video-stereo-mode can't be removed, because it controls whether to "reduce" stereo video to mono, which is also the default. In fact I'm not sure how this should be handled at all.)
2015-04-02 21:54:08 +00:00
char *to = (char *)MP_STEREO3D_NAME(params.stereo_out);
video: change automatic rotation and 3D filter insertion We inserted these filters with fixed parameters, which was ok. But this also didn't change image parameters for the filters down the filter chain and the VO. For example, if rotation by 90° was requested by the file, we would insert a filter and rotate the video, but the VO would still receive image parameters that direct rotation by 90°. This wasn't a problem, but it could become one. Fix this by letting the filters automatically pick up the image params. The image params are reset on application. (We could probably also always try to apply and reset image params in a filter, instead of having special "auto" parameters. This would probably work, and video.c would insert a "rotate=0" filter. But I'm afraid this would be confusing and the current solution is cosmetically slightly nicer.) Unfortunately, the vf_stereo3d.c change turned out a big mess, but once the "internal" filter is fully replaced with libavfilter, most of this can be radically simplified.
2014-09-27 16:13:14 +00:00
if (to) {
video: deprecate almost all video filters The plan is to nuke the custom filter chain completely. It's not clear what will happen to the still needed builtin filters (mostly hardware deinterlacing and vf_vapoursynth). Most likely we'll replace them with different filter chain concept (whose main purpose will be providing builtin things and bridging to libavfilter). The undocumented "warn" options are there to disable deprecation warnings when the player inserts filter automatically. The same will be done to audio filters, at a later point.
2017-04-02 15:59:45 +00:00
char *args[] = {"in", "auto", "out", to, "warn", "no", NULL, NULL};
player: fix breakage when combining 3D and rotate auto-filters This would get stuck in reconfiguring the filter chain forever, because params was mutated ("params.rotate = 0;"). This was used as input for vf_reconfig(), but the filter chain input must always be equivalent to the decoder output, or filter chain reconfiguration will be triggered. The line of code to reset the rotation is from a time when this used to work differently. Also remove the unnecessary try_filter() parameter.
2016-03-28 17:39:33 +00:00
if (try_filter(vo_c, "stereo3d", "autostereo3d", args) < 0)
video: refactor: slightly disentangle video filtering
2016-01-15 21:23:01 +00:00
MP_ERR(vo_c, "Can't insert 3D conversion filter.\n");
video: initial Matroska 3D support This inserts an automatic conversion filter if a Matroska file is marked as 3D (StereoMode element). The basic idea is similar to video rotation and colorspace handling: the 3D mode is added as a property to the video params. Depending on this property, a video filter can be inserted. As of this commit, extending mp_image_params is actually completely unnecessary - but the idea is that it will make it easier to integrate with VOs supporting stereo 3D mogrification. Although vo_opengl does support some stereo rendering, it didn't support the mode my sample file used, so I'll leave that part for later. Not that most mappings from Matroska mode to vf_stereo3d mode are probably wrong, and some are missing. Assuming that Matroska modes, and vf_stereo3d in modes, and out modes are all the same might be an oversimplification - we'll see. See issue #1045.
2014-08-30 21:24:46 +00:00
}
}
player: rewrite deinterlace filter auto-insertion Instead of using the "vf" command code (which changes filters at runtime on user input), use the general filter-insertion code. The latter was added later, and is more suitable for automatically inserted filters. The old code failed in particular when using watch-later saving, which stored the filter list in the resume config file. If a user changed the hardware decoding mode via command line, the stored filter chain was out of date and could cause failure due to not working with hardware or software decoding mode. Storing the deinterlace filter in the filter list was unavoidable, because it was part of the user state. (The new code only edits the actually instantiated filters.)
2016-07-05 19:10:26 +00:00
if (mpctx->opts->deinterlace == 1)
probe_deint_filters(vo_c);
}
static void recreate_auto_filters(struct MPContext *mpctx)
{
filter_reconfig(mpctx, mpctx->vo_chain);
mp_force_video_refresh(mpctx);
mp_notify(mpctx, MPV_EVENT_VIDEO_RECONFIG, NULL);
}
int get_deinterlacing(struct MPContext *mpctx)
{
struct vo_chain *vo_c = mpctx->vo_chain;
int enabled = 0;
player: disable deinterlace property for WIP LGPL mode cehoyos has not agreed to the LGPL relicensing. He added the deinterlace property in commit 7b25afd7. Make it GPL-only for now. The still working parts of the --deinterlace option are not affected by his copyright.
2017-06-23 13:09:21 +00:00
#if HAVE_GPL
player: rewrite deinterlace filter auto-insertion Instead of using the "vf" command code (which changes filters at runtime on user input), use the general filter-insertion code. The latter was added later, and is more suitable for automatically inserted filters. The old code failed in particular when using watch-later saving, which stored the filter list in the resume config file. If a user changed the hardware decoding mode via command line, the stored filter chain was out of date and could cause failure due to not working with hardware or software decoding mode. Storing the deinterlace filter in the filter list was unavoidable, because it was part of the user state. (The new code only edits the actually instantiated filters.)
2016-07-05 19:10:26 +00:00
if (video_vf_vo_control(vo_c, VFCTRL_GET_DEINTERLACE, &enabled) != CONTROL_OK)
enabled = -1;
player: disable deinterlace property for WIP LGPL mode cehoyos has not agreed to the LGPL relicensing. He added the deinterlace property in commit 7b25afd7. Make it GPL-only for now. The still working parts of the --deinterlace option are not affected by his copyright.
2017-06-23 13:09:21 +00:00
#endif
player: rewrite deinterlace filter auto-insertion Instead of using the "vf" command code (which changes filters at runtime on user input), use the general filter-insertion code. The latter was added later, and is more suitable for automatically inserted filters. The old code failed in particular when using watch-later saving, which stored the filter list in the resume config file. If a user changed the hardware decoding mode via command line, the stored filter chain was out of date and could cause failure due to not working with hardware or software decoding mode. Storing the deinterlace filter in the filter list was unavoidable, because it was part of the user state. (The new code only edits the actually instantiated filters.)
2016-07-05 19:10:26 +00:00
if (enabled < 0) {
// vf_lavfi doesn't support VFCTRL_GET_DEINTERLACE
if (vf_find_by_label(vo_c->vf, VF_DEINTERLACE_LABEL))
enabled = 1;
}
return enabled;
}
player: more option/property consistency fixes Some properties had a different type from their equivalent options (such as mute, volume, deinterlace, edition). This wasn't really sane, as raw option values should be always within their bounds. On the other hand, these properties use a different type to reflect runtime limits (such as range of available editions), or simply to improve the "UI" (you don't want to cycle throuhg the completely useless "auto" value when cycling the "mute" property). Handle this by making them always return the option type, but also allowing them to provide a "constricted" type, which is used for UI purposes. All M_PROPERTY_GET_CONSTRICTED_TYPE changes are related to this. One consequence is that you can set the volume property to arbitrary high values just like with the --volume option, but using the "add" command it still restricts it to the --volume-max range. Also deprecate --chapter, as it is grossly incompatible to the chapter property. We pondered renaming it to --chapters, or introducing a more powerful --range option, but concluded that --start --end is actually enough. These changes appear to take care of the last gross property/option incompatibilities, although there might still be a few lurking.
2016-09-18 14:06:12 +00:00
void set_deinterlacing(struct MPContext *mpctx, int opt_val)
player: rewrite deinterlace filter auto-insertion Instead of using the "vf" command code (which changes filters at runtime on user input), use the general filter-insertion code. The latter was added later, and is more suitable for automatically inserted filters. The old code failed in particular when using watch-later saving, which stored the filter list in the resume config file. If a user changed the hardware decoding mode via command line, the stored filter chain was out of date and could cause failure due to not working with hardware or software decoding mode. Storing the deinterlace filter in the filter list was unavoidable, because it was part of the user state. (The new code only edits the actually instantiated filters.)
2016-07-05 19:10:26 +00:00
{
player: more option/property consistency fixes Some properties had a different type from their equivalent options (such as mute, volume, deinterlace, edition). This wasn't really sane, as raw option values should be always within their bounds. On the other hand, these properties use a different type to reflect runtime limits (such as range of available editions), or simply to improve the "UI" (you don't want to cycle throuhg the completely useless "auto" value when cycling the "mute" property). Handle this by making them always return the option type, but also allowing them to provide a "constricted" type, which is used for UI purposes. All M_PROPERTY_GET_CONSTRICTED_TYPE changes are related to this. One consequence is that you can set the volume property to arbitrary high values just like with the --volume option, but using the "add" command it still restricts it to the --volume-max range. Also deprecate --chapter, as it is grossly incompatible to the chapter property. We pondered renaming it to --chapters, or introducing a more powerful --range option, but concluded that --start --end is actually enough. These changes appear to take care of the last gross property/option incompatibilities, although there might still be a few lurking.
2016-09-18 14:06:12 +00:00
if ((opt_val < 0 && mpctx->opts->deinterlace == opt_val) ||
(opt_val == (get_deinterlacing(mpctx) > 0)))
video: fix deinterlace filter handling for VFCTRL_SET_DEINTERLACE filters Some filters support VFCTRL_SET_DEINTERLACE. This affects most hardware deinterlace filters. They can be inserted by the user manually, or auto- inserted by vf.c itself as conversion filter (vf_d3d11vpp). In these cases, we shouldn't insert or remove filters outselves, and instead VFCTRL_SET_DEINTERLACE should be invoked to switch the mode. This wasn't done correctly in the recently refactored code and could have broken with --deinterlace. (The refactor only considered switching via property in this case.) Fix it by making it a proper part of the filter_reconfig() function, and making set_deinterlacing() (which is called by the property handler) merely call filter_reconfig() in all cases to do the real work. We can even avoid rebuilding the filter chain - though only if no other auto-filters are inserted. It probably also provides a slightly cleaner way to implement functionality in the VO while still inserting video filter fallbacks correctly if required.
2016-07-06 12:08:47 +00:00
return;
player: more option/property consistency fixes Some properties had a different type from their equivalent options (such as mute, volume, deinterlace, edition). This wasn't really sane, as raw option values should be always within their bounds. On the other hand, these properties use a different type to reflect runtime limits (such as range of available editions), or simply to improve the "UI" (you don't want to cycle throuhg the completely useless "auto" value when cycling the "mute" property). Handle this by making them always return the option type, but also allowing them to provide a "constricted" type, which is used for UI purposes. All M_PROPERTY_GET_CONSTRICTED_TYPE changes are related to this. One consequence is that you can set the volume property to arbitrary high values just like with the --volume option, but using the "add" command it still restricts it to the --volume-max range. Also deprecate --chapter, as it is grossly incompatible to the chapter property. We pondered renaming it to --chapters, or introducing a more powerful --range option, but concluded that --start --end is actually enough. These changes appear to take care of the last gross property/option incompatibilities, although there might still be a few lurking.
2016-09-18 14:06:12 +00:00
mpctx->opts->deinterlace = opt_val;
video: fix deinterlace filter handling for VFCTRL_SET_DEINTERLACE filters Some filters support VFCTRL_SET_DEINTERLACE. This affects most hardware deinterlace filters. They can be inserted by the user manually, or auto- inserted by vf.c itself as conversion filter (vf_d3d11vpp). In these cases, we shouldn't insert or remove filters outselves, and instead VFCTRL_SET_DEINTERLACE should be invoked to switch the mode. This wasn't done correctly in the recently refactored code and could have broken with --deinterlace. (The refactor only considered switching via property in this case.) Fix it by making it a proper part of the filter_reconfig() function, and making set_deinterlacing() (which is called by the property handler) merely call filter_reconfig() in all cases to do the real work. We can even avoid rebuilding the filter chain - though only if no other auto-filters are inserted. It probably also provides a slightly cleaner way to implement functionality in the VO while still inserting video filter fallbacks correctly if required.
2016-07-06 12:08:47 +00:00
recreate_auto_filters(mpctx);
player: more option/property consistency fixes Some properties had a different type from their equivalent options (such as mute, volume, deinterlace, edition). This wasn't really sane, as raw option values should be always within their bounds. On the other hand, these properties use a different type to reflect runtime limits (such as range of available editions), or simply to improve the "UI" (you don't want to cycle throuhg the completely useless "auto" value when cycling the "mute" property). Handle this by making them always return the option type, but also allowing them to provide a "constricted" type, which is used for UI purposes. All M_PROPERTY_GET_CONSTRICTED_TYPE changes are related to this. One consequence is that you can set the volume property to arbitrary high values just like with the --volume option, but using the "add" command it still restricts it to the --volume-max range. Also deprecate --chapter, as it is grossly incompatible to the chapter property. We pondered renaming it to --chapters, or introducing a more powerful --range option, but concluded that --start --end is actually enough. These changes appear to take care of the last gross property/option incompatibilities, although there might still be a few lurking.
2016-09-18 14:06:12 +00:00
if (opt_val >= 0)
mpctx->opts->deinterlace = get_deinterlacing(mpctx) > 0;
video: move VO reinit from filter chain to player This gets rid of the vf_vo pseudo-filter. It ends the idea of MPlayer's architecture that the VO is just a (terminating) video filter. It didn't really work for us with respect to video timing (the "end" of the video chain isn't really made for video timing, and making it do so would be awkward), and now we're removing it entirely. We will be able to fix some things, such as properly draining video on reconfiguration.
2013-12-10 18:24:58 +00:00
}
Split mplayer.c mplayer.c was a bit too big. Split it into multiple files. I hope the way it's split makes sense. Maybe some things don't make too much sense, or go against intuition. These will fixed as soon as I notice them. Some files are a bit questionable (misc.c, osd.c, configfiles.c), and suggestions how to organize this better are welcome. Regressions are possible due to reorganized include statements. Obviously I didn't just copy mplayer.c's orgy of include statements, but recreated them for each file. It's easily possible that there are oversights and mistakes, which will show up on other platforms. There is one actual change: the public avutil.h include is removed from encode.h, and I tried to replace most FFMIN/FFMAX/av_clip uses. I consider using libavutil too much as dangerous, because the set of include files they recursively pull in is rather arbitrary and is different between FFmpeg and Libav.
2013-10-29 21:38:29 +00:00
static void recreate_video_filters(struct MPContext *mpctx)
{
struct MPOpts *opts = mpctx->opts;
video: decouple filtering/decoding slightly more Lots of noise to remove the vfilter/vo fields from dec_video. From now on, video filtering and output will still be done together, summarized under struct vo_chain. There is the question where exactly the vf_chain should go in such a decoupled architecture. The end goal is being able to place a "complex" filter between video decoders and output (which will culminate in natural integration of A->V filters for natural integration of libavfilter audio visualizations). The vf_chain is still useful for "final" processing, such as format conversions and deinterlacing. Also, there's only 1 VO and 1 --vf option. So having 1 vf_chain for a VO seems ideal, since otherwise there would be no natural way to handle all these existing options and mechanisms. There is still some work required to truly decouple decoding.
2016-01-13 23:18:48 +00:00
struct vo_chain *vo_c = mpctx->vo_chain;
player: refactor: eliminate MPContext.d_video Eventually we want the VO be driven by a A->V filter, so a decoder doesn't even have to exist. Some features definitely require a decoder though (like reporting the decoder in use, hardware decoding, etc.), so for each thing which accessed d_video, it has to be redecided if and how it can access decoder state. At least the "framedrop" property slightly changes semantics: you can now always set this property, even if no video is active. Some untested changes in this commit, but our bio-based distributed test suite has to take care of this.
2016-01-17 17:07:50 +00:00
assert(vo_c);
Split mplayer.c mplayer.c was a bit too big. Split it into multiple files. I hope the way it's split makes sense. Maybe some things don't make too much sense, or go against intuition. These will fixed as soon as I notice them. Some files are a bit questionable (misc.c, osd.c, configfiles.c), and suggestions how to organize this better are welcome. Regressions are possible due to reorganized include statements. Obviously I didn't just copy mplayer.c's orgy of include statements, but recreated them for each file. It's easily possible that there are oversights and mistakes, which will show up on other platforms. There is one actual change: the public avutil.h include is removed from encode.h, and I tried to replace most FFMIN/FFMAX/av_clip uses. I consider using libavutil too much as dangerous, because the set of include files they recursively pull in is rather arbitrary and is different between FFmpeg and Libav.
2013-10-29 21:38:29 +00:00
video: decouple filtering/decoding slightly more Lots of noise to remove the vfilter/vo fields from dec_video. From now on, video filtering and output will still be done together, summarized under struct vo_chain. There is the question where exactly the vf_chain should go in such a decoupled architecture. The end goal is being able to place a "complex" filter between video decoders and output (which will culminate in natural integration of A->V filters for natural integration of libavfilter audio visualizations). The vf_chain is still useful for "final" processing, such as format conversions and deinterlacing. Also, there's only 1 VO and 1 --vf option. So having 1 vf_chain for a VO seems ideal, since otherwise there would be no natural way to handle all these existing options and mechanisms. There is still some work required to truly decouple decoding.
2016-01-13 23:18:48 +00:00
vf_destroy(vo_c->vf);
vo_c->vf = vf_new(mpctx->global);
video: refactor how VO exports hwdec device handles The main change is with video/hwdec.h. mp_hwdec_info is made opaque (and renamed to mp_hwdec_devices). Its accessors are mainly thread-safe (or documented where not), which makes the whole thing saner and cleaner. In particular, thread-safety rules become less subtle and more obvious. The new internal API makes it easier to support multiple OpenGL interop backends. (Although this is not done yet, and it's not clear whether it ever will.) This also removes all the API-specific fields from mp_hwdec_ctx and replaces them with a "ctx" field. For d3d in particular, we drop the mp_d3d_ctx struct completely, and pass the interfaces directly. Remove the emulation checks from vaapi.c and vdpau.c; they are pointless, and the checks that matter are done on the VO layer. The d3d hardware decoders might slightly change behavior: dxva2-copy will not use the VO device anymore if the VO supports proper interop. This pretty much assumes that any in such cases the VO will not use any form of exclusive mode, which makes using the VO device in copy mode unnecessary. This is a big refactor. Some things may be untested and could be broken.
2016-05-09 17:42:03 +00:00
vo_c->vf->hwdec_devs = vo_c->hwdec_devs;
player: minor changes in init code Move the MPV_LEAK_REPORT env query to mp_create(), where it will also be used by the client API (it might be helpful, so why not). The same applies to MPV_VERBOSE. The prepare_playlist() call doesn't need to be in mp_initialize() and can just be in mp_play_files() to reduce the size of mp_initialize(). Also, remove wakeup_playloop(), which is 100% redundant with mp_wakeup_core_cb().
2016-09-19 17:57:31 +00:00
vo_c->vf->wakeup_callback = mp_wakeup_core_cb;
video: decouple filtering/decoding slightly more Lots of noise to remove the vfilter/vo fields from dec_video. From now on, video filtering and output will still be done together, summarized under struct vo_chain. There is the question where exactly the vf_chain should go in such a decoupled architecture. The end goal is being able to place a "complex" filter between video decoders and output (which will culminate in natural integration of A->V filters for natural integration of libavfilter audio visualizations). The vf_chain is still useful for "final" processing, such as format conversions and deinterlacing. Also, there's only 1 VO and 1 --vf option. So having 1 vf_chain for a VO seems ideal, since otherwise there would be no natural way to handle all these existing options and mechanisms. There is still some work required to truly decouple decoding.
2016-01-13 23:18:48 +00:00
vo_c->vf->wakeup_callback_ctx = mpctx;
player: refactor: eliminate MPContext.d_video Eventually we want the VO be driven by a A->V filter, so a decoder doesn't even have to exist. Some features definitely require a decoder though (like reporting the decoder in use, hardware decoding, etc.), so for each thing which accessed d_video, it has to be redecided if and how it can access decoder state. At least the "framedrop" property slightly changes semantics: you can now always set this property, even if no video is active. Some untested changes in this commit, but our bio-based distributed test suite has to take care of this.
2016-01-17 17:07:50 +00:00
vo_c->vf->container_fps = vo_c->container_fps;
video: decouple filtering/decoding slightly more Lots of noise to remove the vfilter/vo fields from dec_video. From now on, video filtering and output will still be done together, summarized under struct vo_chain. There is the question where exactly the vf_chain should go in such a decoupled architecture. The end goal is being able to place a "complex" filter between video decoders and output (which will culminate in natural integration of A->V filters for natural integration of libavfilter audio visualizations). The vf_chain is still useful for "final" processing, such as format conversions and deinterlacing. Also, there's only 1 VO and 1 --vf option. So having 1 vf_chain for a VO seems ideal, since otherwise there would be no natural way to handle all these existing options and mechanisms. There is still some work required to truly decouple decoding.
2016-01-13 23:18:48 +00:00
vo_control(vo_c->vo, VOCTRL_GET_DISPLAY_FPS, &vo_c->vf->display_fps);
Split mplayer.c mplayer.c was a bit too big. Split it into multiple files. I hope the way it's split makes sense. Maybe some things don't make too much sense, or go against intuition. These will fixed as soon as I notice them. Some files are a bit questionable (misc.c, osd.c, configfiles.c), and suggestions how to organize this better are welcome. Regressions are possible due to reorganized include statements. Obviously I didn't just copy mplayer.c's orgy of include statements, but recreated them for each file. It's easily possible that there are oversights and mistakes, which will show up on other platforms. There is one actual change: the public avutil.h include is removed from encode.h, and I tried to replace most FFMIN/FFMAX/av_clip uses. I consider using libavutil too much as dangerous, because the set of include files they recursively pull in is rather arbitrary and is different between FFmpeg and Libav.
2013-10-29 21:38:29 +00:00
video: decouple filtering/decoding slightly more Lots of noise to remove the vfilter/vo fields from dec_video. From now on, video filtering and output will still be done together, summarized under struct vo_chain. There is the question where exactly the vf_chain should go in such a decoupled architecture. The end goal is being able to place a "complex" filter between video decoders and output (which will culminate in natural integration of A->V filters for natural integration of libavfilter audio visualizations). The vf_chain is still useful for "final" processing, such as format conversions and deinterlacing. Also, there's only 1 VO and 1 --vf option. So having 1 vf_chain for a VO seems ideal, since otherwise there would be no natural way to handle all these existing options and mechanisms. There is still some work required to truly decouple decoding.
2016-01-13 23:18:48 +00:00
vf_append_filter_list(vo_c->vf, opts->vf_settings);
Split mplayer.c mplayer.c was a bit too big. Split it into multiple files. I hope the way it's split makes sense. Maybe some things don't make too much sense, or go against intuition. These will fixed as soon as I notice them. Some files are a bit questionable (misc.c, osd.c, configfiles.c), and suggestions how to organize this better are welcome. Regressions are possible due to reorganized include statements. Obviously I didn't just copy mplayer.c's orgy of include statements, but recreated them for each file. It's easily possible that there are oversights and mistakes, which will show up on other platforms. There is one actual change: the public avutil.h include is removed from encode.h, and I tried to replace most FFMIN/FFMAX/av_clip uses. I consider using libavutil too much as dangerous, because the set of include files they recursively pull in is rather arbitrary and is different between FFmpeg and Libav.
2013-10-29 21:38:29 +00:00
video: create a separate context for video filter chain This adds vf_chain, which unlike vf_instance refers to the filter chain as a whole. This makes the filter API less awkward, and will allow handling format negotiation better.
2013-12-07 18:32:44 +00:00
// for vf_sub
sub: uglify OSD code path with locking Do two things: 1. add locking to struct osd_state 2. make struct osd_state opaque While 1. is somewhat simple, 2. is quite horrible. Lots of code accesses lots of osd_state (and osd_object) members. To make sure everything is accessed synchronously, I prefer making osd_state opaque, even if it means adding pretty dumb accessors. All of this is meant to allow running VO in their own threads. Eventually, VOs will request OSD on their own, which means osd_state will be accessed from foreign threads.
2014-01-18 00:19:20 +00:00
osd_set_render_subs_in_filter(mpctx->osd,
video: decouple filtering/decoding slightly more Lots of noise to remove the vfilter/vo fields from dec_video. From now on, video filtering and output will still be done together, summarized under struct vo_chain. There is the question where exactly the vf_chain should go in such a decoupled architecture. The end goal is being able to place a "complex" filter between video decoders and output (which will culminate in natural integration of A->V filters for natural integration of libavfilter audio visualizations). The vf_chain is still useful for "final" processing, such as format conversions and deinterlacing. Also, there's only 1 VO and 1 --vf option. So having 1 vf_chain for a VO seems ideal, since otherwise there would be no natural way to handle all these existing options and mechanisms. There is still some work required to truly decouple decoding.
2016-01-13 23:18:48 +00:00
vf_control_any(vo_c->vf, VFCTRL_INIT_OSD, mpctx->osd) > 0);
video: move VO reinit from filter chain to player This gets rid of the vf_vo pseudo-filter. It ends the idea of MPlayer's architecture that the VO is just a (terminating) video filter. It didn't really work for us with respect to video timing (the "end" of the video chain isn't really made for video timing, and making it do so would be awkward), and now we're removing it entirely. We will be able to fix some things, such as properly draining video on reconfiguration.
2013-12-10 18:24:58 +00:00
video: decouple filtering/decoding slightly more Lots of noise to remove the vfilter/vo fields from dec_video. From now on, video filtering and output will still be done together, summarized under struct vo_chain. There is the question where exactly the vf_chain should go in such a decoupled architecture. The end goal is being able to place a "complex" filter between video decoders and output (which will culminate in natural integration of A->V filters for natural integration of libavfilter audio visualizations). The vf_chain is still useful for "final" processing, such as format conversions and deinterlacing. Also, there's only 1 VO and 1 --vf option. So having 1 vf_chain for a VO seems ideal, since otherwise there would be no natural way to handle all these existing options and mechanisms. There is still some work required to truly decouple decoding.
2016-01-13 23:18:48 +00:00
set_allowed_vo_formats(vo_c);
Split mplayer.c mplayer.c was a bit too big. Split it into multiple files. I hope the way it's split makes sense. Maybe some things don't make too much sense, or go against intuition. These will fixed as soon as I notice them. Some files are a bit questionable (misc.c, osd.c, configfiles.c), and suggestions how to organize this better are welcome. Regressions are possible due to reorganized include statements. Obviously I didn't just copy mplayer.c's orgy of include statements, but recreated them for each file. It's easily possible that there are oversights and mistakes, which will show up on other platforms. There is one actual change: the public avutil.h include is removed from encode.h, and I tried to replace most FFMIN/FFMAX/av_clip uses. I consider using libavutil too much as dangerous, because the set of include files they recursively pull in is rather arbitrary and is different between FFmpeg and Libav.
2013-10-29 21:38:29 +00:00
}
int reinit_video_filters(struct MPContext *mpctx)
{
video: decouple filtering/decoding slightly more Lots of noise to remove the vfilter/vo fields from dec_video. From now on, video filtering and output will still be done together, summarized under struct vo_chain. There is the question where exactly the vf_chain should go in such a decoupled architecture. The end goal is being able to place a "complex" filter between video decoders and output (which will culminate in natural integration of A->V filters for natural integration of libavfilter audio visualizations). The vf_chain is still useful for "final" processing, such as format conversions and deinterlacing. Also, there's only 1 VO and 1 --vf option. So having 1 vf_chain for a VO seems ideal, since otherwise there would be no natural way to handle all these existing options and mechanisms. There is still some work required to truly decouple decoding.
2016-01-13 23:18:48 +00:00
struct vo_chain *vo_c = mpctx->vo_chain;
Split mplayer.c mplayer.c was a bit too big. Split it into multiple files. I hope the way it's split makes sense. Maybe some things don't make too much sense, or go against intuition. These will fixed as soon as I notice them. Some files are a bit questionable (misc.c, osd.c, configfiles.c), and suggestions how to organize this better are welcome. Regressions are possible due to reorganized include statements. Obviously I didn't just copy mplayer.c's orgy of include statements, but recreated them for each file. It's easily possible that there are oversights and mistakes, which will show up on other platforms. There is one actual change: the public avutil.h include is removed from encode.h, and I tried to replace most FFMIN/FFMAX/av_clip uses. I consider using libavutil too much as dangerous, because the set of include files they recursively pull in is rather arbitrary and is different between FFmpeg and Libav.
2013-10-29 21:38:29 +00:00
player: refactor: eliminate MPContext.d_video Eventually we want the VO be driven by a A->V filter, so a decoder doesn't even have to exist. Some features definitely require a decoder though (like reporting the decoder in use, hardware decoding, etc.), so for each thing which accessed d_video, it has to be redecided if and how it can access decoder state. At least the "framedrop" property slightly changes semantics: you can now always set this property, even if no video is active. Some untested changes in this commit, but our bio-based distributed test suite has to take care of this.
2016-01-17 17:07:50 +00:00
if (!vo_c)
command: allow changing filters before video chain initialization Apparently this is more intuitive. Somewhat tricky, because of the odd state after loading a file but before initializing the VO.
2014-03-30 17:21:54 +00:00
return 0;
video: decouple filtering/decoding slightly more Lots of noise to remove the vfilter/vo fields from dec_video. From now on, video filtering and output will still be done together, summarized under struct vo_chain. There is the question where exactly the vf_chain should go in such a decoupled architecture. The end goal is being able to place a "complex" filter between video decoders and output (which will culminate in natural integration of A->V filters for natural integration of libavfilter audio visualizations). The vf_chain is still useful for "final" processing, such as format conversions and deinterlacing. Also, there's only 1 VO and 1 --vf option. So having 1 vf_chain for a VO seems ideal, since otherwise there would be no natural way to handle all these existing options and mechanisms. There is still some work required to truly decouple decoding.
2016-01-13 23:18:48 +00:00
bool need_reconfig = vo_c->vf->initialized != 0;
Split mplayer.c mplayer.c was a bit too big. Split it into multiple files. I hope the way it's split makes sense. Maybe some things don't make too much sense, or go against intuition. These will fixed as soon as I notice them. Some files are a bit questionable (misc.c, osd.c, configfiles.c), and suggestions how to organize this better are welcome. Regressions are possible due to reorganized include statements. Obviously I didn't just copy mplayer.c's orgy of include statements, but recreated them for each file. It's easily possible that there are oversights and mistakes, which will show up on other platforms. There is one actual change: the public avutil.h include is removed from encode.h, and I tried to replace most FFMIN/FFMAX/av_clip uses. I consider using libavutil too much as dangerous, because the set of include files they recursively pull in is rather arbitrary and is different between FFmpeg and Libav.
2013-10-29 21:38:29 +00:00
recreate_video_filters(mpctx);
command: allow changing filters before video chain initialization Apparently this is more intuitive. Somewhat tricky, because of the odd state after loading a file but before initializing the VO.
2014-03-30 17:21:54 +00:00
if (need_reconfig)
player: rewrite deinterlace filter auto-insertion Instead of using the "vf" command code (which changes filters at runtime on user input), use the general filter-insertion code. The latter was added later, and is more suitable for automatically inserted filters. The old code failed in particular when using watch-later saving, which stored the filter list in the resume config file. If a user changed the hardware decoding mode via command line, the stored filter chain was out of date and could cause failure due to not working with hardware or software decoding mode. Storing the deinterlace filter in the filter list was unavoidable, because it was part of the user state. (The new code only edits the actually instantiated filters.)
2016-07-05 19:10:26 +00:00
filter_reconfig(mpctx, vo_c);
video: refactor: slightly disentangle video filtering
2016-01-15 21:23:01 +00:00
player: force refresh seek when changing audio filters Unfortunately I see no better solution. The refresh seek is skipped if the amount of buffered audio is not overly huge. Unfortunately softvol af_volume insertion still can cause this issue, because it's outside of the normal dynamic filter chain changing code. Move the video refresh call to reinit_video_filters() to make it more uniform along with the audio code.
2016-02-09 21:19:01 +00:00
mp_force_video_refresh(mpctx);
video: refactor: slightly disentangle video filtering
2016-01-15 21:23:01 +00:00
mp_notify(mpctx, MPV_EVENT_VIDEO_RECONFIG, NULL);
command: allow changing filters before video chain initialization Apparently this is more intuitive. Somewhat tricky, because of the odd state after loading a file but before initializing the VO.
2014-03-30 17:21:54 +00:00
video: decouple filtering/decoding slightly more Lots of noise to remove the vfilter/vo fields from dec_video. From now on, video filtering and output will still be done together, summarized under struct vo_chain. There is the question where exactly the vf_chain should go in such a decoupled architecture. The end goal is being able to place a "complex" filter between video decoders and output (which will culminate in natural integration of A->V filters for natural integration of libavfilter audio visualizations). The vf_chain is still useful for "final" processing, such as format conversions and deinterlacing. Also, there's only 1 VO and 1 --vf option. So having 1 vf_chain for a VO seems ideal, since otherwise there would be no natural way to handle all these existing options and mechanisms. There is still some work required to truly decouple decoding.
2016-01-13 23:18:48 +00:00
return vo_c->vf->initialized;
}
static void vo_chain_reset_state(struct vo_chain *vo_c)
{
video: refactor: disentangle decoding/filtering some more This moves some code related to decoding from video.c to dec_video.c, and also removes some accesses to dec_video.c from the filtering code. dec_video.ch is starting to make sense, and simply returns video frames from a demuxer stream. The API exposed is also somewhat intended to be easily changeable to move decoding to a separate thread, if we ever want this (due to libavcodec already being threaded, I don't see much of a reason, but it might still be helpful).
2016-01-16 20:19:52 +00:00
mp_image_unrefp(&vo_c->input_mpi);
video: decouple filtering/decoding slightly more Lots of noise to remove the vfilter/vo fields from dec_video. From now on, video filtering and output will still be done together, summarized under struct vo_chain. There is the question where exactly the vf_chain should go in such a decoupled architecture. The end goal is being able to place a "complex" filter between video decoders and output (which will culminate in natural integration of A->V filters for natural integration of libavfilter audio visualizations). The vf_chain is still useful for "final" processing, such as format conversions and deinterlacing. Also, there's only 1 VO and 1 --vf option. So having 1 vf_chain for a VO seems ideal, since otherwise there would be no natural way to handle all these existing options and mechanisms. There is still some work required to truly decouple decoding.
2016-01-13 23:18:48 +00:00
if (vo_c->vf->initialized == 1)
vf_seek_reset(vo_c->vf);
vo_seek_reset(vo_c->vo);
player: refactor: some more minor decoder/output decoupling These changes don't make too much sense without context, but are preparation for later. Then the audio_src/video_src fields will be actually be NULL under circumstances.
2016-01-29 21:46:28 +00:00
if (vo_c->video_src)
video_reset(vo_c->video_src);
player: change aspects of cover art handling Cover art handling is a disgusting hack that causes a mess in all components. And this will stay this way. This is the Xth time I've changed cover art handling, and that will probably also continue. But change the code such that cover art is injected into the demux packet stream, instead of having an explicit special case it in the decoder glue code. (This is somewhat more similar to the cover art hack in libavformat.) To avoid that the over art picture is decoded again on each seek, we need some additional "caching" in player/video.c. Decoding it after each seek would work as well, but since cover art pictures can be pretty huge, it's probably ok to invest some lines of code into caching it. One weird thing is that the cover art packet will remain queued after seeks, but that is probably not an issue. In exchange, we can drop the dec_video.c code, which is pretty convenient for one of the following commits. This code duplicates a bunch of lower-level decode calls and does icky messing with this weird state stuff, so I'm glad it goes away.
2017-01-10 13:03:41 +00:00
// Prepare for continued playback after a seek.
if (!vo_c->input_mpi && vo_c->cached_coverart)
vo_c->input_mpi = mp_image_new_ref(vo_c->cached_coverart);
Split mplayer.c mplayer.c was a bit too big. Split it into multiple files. I hope the way it's split makes sense. Maybe some things don't make too much sense, or go against intuition. These will fixed as soon as I notice them. Some files are a bit questionable (misc.c, osd.c, configfiles.c), and suggestions how to organize this better are welcome. Regressions are possible due to reorganized include statements. Obviously I didn't just copy mplayer.c's orgy of include statements, but recreated them for each file. It's easily possible that there are oversights and mistakes, which will show up on other platforms. There is one actual change: the public avutil.h include is removed from encode.h, and I tried to replace most FFMIN/FFMAX/av_clip uses. I consider using libavutil too much as dangerous, because the set of include files they recursively pull in is rather arbitrary and is different between FFmpeg and Libav.
2013-10-29 21:38:29 +00:00
}
player: split seek_reset() This also reduces some code duplication with other parts of the code. The changfe is mostly cosmetic, although there are also some subtle changes in behavior. At least one change is that the big desync message is now printed after every seek.
2014-07-30 21:01:55 +00:00
void reset_video_state(struct MPContext *mpctx)
{
player: refactor: some more minor decoder/output decoupling These changes don't make too much sense without context, but are preparation for later. Then the audio_src/video_src fields will be actually be NULL under circumstances.
2016-01-29 21:46:28 +00:00
if (mpctx->vo_chain)
video: decouple filtering/decoding slightly more Lots of noise to remove the vfilter/vo fields from dec_video. From now on, video filtering and output will still be done together, summarized under struct vo_chain. There is the question where exactly the vf_chain should go in such a decoupled architecture. The end goal is being able to place a "complex" filter between video decoders and output (which will culminate in natural integration of A->V filters for natural integration of libavfilter audio visualizations). The vf_chain is still useful for "final" processing, such as format conversions and deinterlacing. Also, there's only 1 VO and 1 --vf option. So having 1 vf_chain for a VO seems ideal, since otherwise there would be no natural way to handle all these existing options and mechanisms. There is still some work required to truly decouple decoding.
2016-01-13 23:18:48 +00:00
vo_chain_reset_state(mpctx->vo_chain);
player: split seek_reset() This also reduces some code duplication with other parts of the code. The changfe is mostly cosmetic, although there are also some subtle changes in behavior. At least one change is that the big desync message is now printed after every seek.
2014-07-30 21:01:55 +00:00
video: pass future frames to VO Now the VO can request a number of future frames with the last parameter of vo_set_queue_params(). This will be helpful to fix the interpolation code. Note that the first frame (after playback start or seeking) will usually not have any future frames (to make seeking fast). Near the end of the file, the number of future frames will become lower as well.
2015-07-01 17:22:40 +00:00
for (int n = 0; n < mpctx->num_next_frames; n++)
mp_image_unrefp(&mpctx->next_frames[n]);
mpctx->num_next_frames = 0;
player: when seeking past EOF with --keep-open, seek to last frame It feels strange that seeking past EOF with --keep-open actually leaves the player at a random position. You can't even unpause, because the demuxer is in the EOF state, and what you see on screen is just what was around before the seek. Improve this by attempting to seek to the last video frame if EOF happens. We explicitly don't do this if EOF was reached normally to increase robustness (if the VO got a frame since the last seek, it obviously means we had normal playback before EOF). If an error happens when trying to find the last frame (such as not actually finding a last frame because e.g. the demuxer misbehaves), this will probably turn your CPU into a heater. There is no logic to prevent reinitiating the last-frame search if the last-frame search reached EOF. (Pausing usually prevents that EOF is reached again after a successful last-frame search.) Fixes #819.
2014-12-07 01:47:09 +00:00
mp_image_unrefp(&mpctx->saved_frame);
video: fix and simplify video format changes and last frame display The previous commit broke these things, and fixing them is separate in this commit in order to reduce the volume of changes. Move the image queue from the VO to the playback core. The image queue is a remnant of the old way how vdpau was implemented, and increasingly became more and more an artifact. In the end, it did only one thing: computing the duration of the current frame. This was done by taking the PTS difference between the current and the future frame. We keep this, but by moving it out of the VO, we don't have to special-case format changes anymore. This simplifies the code a lot. Since we need the queue to compute the duration only, a queue size larger than 2 makes no sense, and we can hardcode that. Also change how the last frame is handled. The last frame is a bit of a problem, because video timing works by showing one frame after another, which makes it a special case. Make the VO provide a function to notify us when the frame is done, instead. The frame duration is used for that. This is not perfect. For example, changing playback speed during the last frame doesn't update the end time. Pausing will not stop the clock that times the last frame. But I don't think this matters for such a corner case.
2014-08-12 21:17:35 +00:00
player: split seek_reset() This also reduces some code duplication with other parts of the code. The changfe is mostly cosmetic, although there are also some subtle changes in behavior. At least one change is that the big desync message is now printed after every seek.
2014-07-30 21:01:55 +00:00
mpctx->delay = 0;
mpctx->time_frame = 0;
video: get rid of video_next_pts field Not really needed anymore. Code should be mostly equivalent. Also get rid of some other now-unused or outdated things.
2014-08-22 12:20:00 +00:00
mpctx->video_pts = MP_NOPTS_VALUE;
video: use demuxer-signaled duration for last video frame Helps with gif, probably does unwanted things with other formats. This doesn't handle --end quite correctly, but this could be added later. Fixes #3924.
2016-12-21 17:18:24 +00:00
mpctx->last_frame_duration = 0;
player: refactor display-sync frame duration calculations Get rid of get_past_frame_durations(), which was a bit too messy. Add a past_frames array, which contains the same information in a more reasonable way. This also means that we can get the exact current and past frame durations without going through awful stuff. (The main problem is that vo_pts_history contains future frames as well, which is needed for frame backstepping etc., but gets in the way here.) Also disable the automatic disabling of display-sync if the frame duration changes, and extend the frame durations allowed for display sync. To allow arbitrarily high durations, vo.c needs to be changed to pause and potentially redraw OSD while showing a single frame, so they're still limited. In an attempt to deal with VFR, calculate the overall speed using the average FPS. The frame scheduling itself does not use the average FPS, but the duration of the current frame. This does not work too well, but provides a good base for further improvements. Where this commit actually helps a lot is dealing with rounded timestamps, e.g. if the container framerate is wrong or unknown, or if the muxer wrote incorrectly rounded timestamps. While the rounding errors apparently can't be get rid of completely in the general case, this is still much better than e.g. disabling display-sync completely just because some frame durations go out of bounds.
2015-11-13 21:45:40 +00:00
mpctx->num_past_frames = 0;
player: split seek_reset() This also reduces some code duplication with other parts of the code. The changfe is mostly cosmetic, although there are also some subtle changes in behavior. At least one change is that the big desync message is now printed after every seek.
2014-07-30 21:01:55 +00:00
mpctx->total_avsync_change = 0;
video: do not use MP_NOPTS_VALUE for A/V difference There's no need for this, it just creates more corner cases. Also always reset it on seeks etc..
2015-05-24 18:57:31 +00:00
mpctx->last_av_difference = 0;
video: refactor: disentangle decoding/filtering some more This moves some code related to decoding from video.c to dec_video.c, and also removes some accesses to dec_video.c from the filtering code. dec_video.ch is starting to make sense, and simply returns video frames from a demuxer stream. The API exposed is also somewhat intended to be easily changeable to move decoding to a separate thread, if we ever want this (due to libavcodec already being threaded, I don't see much of a reason, but it might still be helpful).
2016-01-16 20:19:52 +00:00
mpctx->dropped_frames_start = 0;
command: add mistimed-frame-count property Does what the manpage says. This is a replacement incrementing the dropped frame counter (see previous commit).
2015-10-30 13:05:41 +00:00
mpctx->mistimed_frames_total = 0;
player: split seek_reset() This also reduces some code duplication with other parts of the code. The changfe is mostly cosmetic, although there are also some subtle changes in behavior. At least one change is that the big desync message is now printed after every seek.
2014-07-30 21:01:55 +00:00
mpctx->drop_message_shown = 0;
player: add display sync mode If this mode is enabled, the player tries to strictly synchronize video to display refresh. It will adjust playback speed to match the display, so if you play 23.976 fps video on a 24 Hz screen, playback speed is increased by approximately 1/1000. Audio wll be resampled to keep up with playback. This is different from the default sync mode, which will sync video to audio, with the consequence that video might skip or repeat a frame once in a while to make video keep up with audio. This is still unpolished. There are some major problems as well; in particular, mkv VFR files won't work well. The reason is that Matroska is terrible and rounds timestamps to milliseconds. This makes it rather hard to guess the framerate of a section of video that is playing. We could probably fix this by just accepting jittery timestamps (instead of explicitly disabling the sync code in this case), but I'm not ready to accept such a solution yet. Another issue is that we are extremely reliant on OS video and audio APIs working in an expected manner, which of course is not too often the case. Consequently, the new sync mode is a bit fragile.
2015-08-10 16:43:25 +00:00
mpctx->display_sync_drift_dir = 0;
player: always disable display-sync on desyncs Instead of periodically trying to enable it again. There are two cases that can happen: 1. A random discontinuity messed everything up, 2. Things are just broken and will desync all the time Until now, it tried to deal with case 1 - but maybe this is really rare, and we don't really need to care about it. On the other hand, case 2 is kind of hard to diagnose if the user doesn't use the terminal. Seeking will reenable display-sync, so you can fix playback if case 1 happens, but still get predictable behavior in case 2.
2015-11-27 13:40:52 +00:00
mpctx->display_sync_broken = false;
player: split seek_reset() This also reduces some code duplication with other parts of the code. The changfe is mostly cosmetic, although there are also some subtle changes in behavior. At least one change is that the big desync message is now printed after every seek.
2014-07-30 21:01:55 +00:00
player: refactor: eliminate MPContext.d_video Eventually we want the VO be driven by a A->V filter, so a decoder doesn't even have to exist. Some features definitely require a decoder though (like reporting the decoder in use, hardware decoding, etc.), so for each thing which accessed d_video, it has to be redecided if and how it can access decoder state. At least the "framedrop" property slightly changes semantics: you can now always set this property, even if no video is active. Some untested changes in this commit, but our bio-based distributed test suite has to take care of this.
2016-01-17 17:07:50 +00:00
mpctx->video_status = mpctx->vo_chain ? STATUS_SYNCING : STATUS_EOF;
player: split seek_reset() This also reduces some code duplication with other parts of the code. The changfe is mostly cosmetic, although there are also some subtle changes in behavior. At least one change is that the big desync message is now printed after every seek.
2014-07-30 21:01:55 +00:00
}
player: remove central uninit_player() function and flags mess Each subsystem (or similar thing) had an INITIALIZED_ flag assigned. The main use of this was that you could pass a bitmask of these flags to uninit_player(). Except in some situations where you wanted to uninitialize nearly everything, this wasn't really useful. Moreover, it was quite annoying that subsystems had most of the code in a specific file, but the uninit code in loadfile.c (because that's where uninit_player() was implemented). Simplify all this. Remove the flags; e.g. instead of testing for the INITIALIZED_AO flag, test whether mpctx->ao is set. Move uninit code to separate functions, e.g. uninit_audio_out().
2014-10-03 17:57:49 +00:00
void uninit_video_out(struct MPContext *mpctx)
{
uninit_video_chain(mpctx);
player: fix excessive CPU usage in audio-only mode Caused by one of the --force-window commits. The unconditional uninit_video_out() call (which normally should be idempotent) raised sporadic MPV_EVENT_VIDEO_RECONFIG events. This is ok, except for the fact that clients (like a Lua script or libmpv users) would cause the event loop to run again after receiving it, triggering a feedback loop. Fix it by sending the events really only on a change.
2015-09-22 09:19:39 +00:00
if (mpctx->video_out) {
player: remove central uninit_player() function and flags mess Each subsystem (or similar thing) had an INITIALIZED_ flag assigned. The main use of this was that you could pass a bitmask of these flags to uninit_player(). Except in some situations where you wanted to uninitialize nearly everything, this wasn't really useful. Moreover, it was quite annoying that subsystems had most of the code in a specific file, but the uninit code in loadfile.c (because that's where uninit_player() was implemented). Simplify all this. Remove the flags; e.g. instead of testing for the INITIALIZED_AO flag, test whether mpctx->ao is set. Move uninit code to separate functions, e.g. uninit_audio_out().
2014-10-03 17:57:49 +00:00
vo_destroy(mpctx->video_out);
player: fix excessive CPU usage in audio-only mode Caused by one of the --force-window commits. The unconditional uninit_video_out() call (which normally should be idempotent) raised sporadic MPV_EVENT_VIDEO_RECONFIG events. This is ok, except for the fact that clients (like a Lua script or libmpv users) would cause the event loop to run again after receiving it, triggering a feedback loop. Fix it by sending the events really only on a change.
2015-09-22 09:19:39 +00:00
mp_notify(mpctx, MPV_EVENT_VIDEO_RECONFIG, NULL);
}
player: remove central uninit_player() function and flags mess Each subsystem (or similar thing) had an INITIALIZED_ flag assigned. The main use of this was that you could pass a bitmask of these flags to uninit_player(). Except in some situations where you wanted to uninitialize nearly everything, this wasn't really useful. Moreover, it was quite annoying that subsystems had most of the code in a specific file, but the uninit code in loadfile.c (because that's where uninit_player() was implemented). Simplify all this. Remove the flags; e.g. instead of testing for the INITIALIZED_AO flag, test whether mpctx->ao is set. Move uninit code to separate functions, e.g. uninit_audio_out().
2014-10-03 17:57:49 +00:00
mpctx->video_out = NULL;
}
video: decouple filtering/decoding slightly more Lots of noise to remove the vfilter/vo fields from dec_video. From now on, video filtering and output will still be done together, summarized under struct vo_chain. There is the question where exactly the vf_chain should go in such a decoupled architecture. The end goal is being able to place a "complex" filter between video decoders and output (which will culminate in natural integration of A->V filters for natural integration of libavfilter audio visualizations). The vf_chain is still useful for "final" processing, such as format conversions and deinterlacing. Also, there's only 1 VO and 1 --vf option. So having 1 vf_chain for a VO seems ideal, since otherwise there would be no natural way to handle all these existing options and mechanisms. There is still some work required to truly decouple decoding.
2016-01-13 23:18:48 +00:00
static void vo_chain_uninit(struct vo_chain *vo_c)
{
player: add complex filter graph support See --lavfi-complex option. This is still quite rough. There's no support for dynamic configuration of any kind. There are probably corner cases where playback might freeze or burn 100% CPU (due to dataflow problems when interaction with libavfilter). Future possible plans might include: - freely switch tracks by providing some sort of default track graph label - automatically enabling audio visualization - automatically mix audio or stack video when multiple tracks are selected at once (similar to how multiple sub tracks can be selected)
2016-02-05 22:19:56 +00:00
struct track *track = vo_c->track;
if (track) {
assert(track->vo_c == vo_c);
track->vo_c = NULL;
assert(track->d_video == vo_c->video_src);
track->d_video = NULL;
video_uninit(vo_c->video_src);
}
if (vo_c->filter_src)
lavfi_set_connected(vo_c->filter_src, false);
video: refactor: disentangle decoding/filtering some more This moves some code related to decoding from video.c to dec_video.c, and also removes some accesses to dec_video.c from the filtering code. dec_video.ch is starting to make sense, and simply returns video frames from a demuxer stream. The API exposed is also somewhat intended to be easily changeable to move decoding to a separate thread, if we ever want this (due to libavcodec already being threaded, I don't see much of a reason, but it might still be helpful).
2016-01-16 20:19:52 +00:00
mp_image_unrefp(&vo_c->input_mpi);
player: change aspects of cover art handling Cover art handling is a disgusting hack that causes a mess in all components. And this will stay this way. This is the Xth time I've changed cover art handling, and that will probably also continue. But change the code such that cover art is injected into the demux packet stream, instead of having an explicit special case it in the decoder glue code. (This is somewhat more similar to the cover art hack in libavformat.) To avoid that the over art picture is decoded again on each seek, we need some additional "caching" in player/video.c. Decoding it after each seek would work as well, but since cover art pictures can be pretty huge, it's probably ok to invest some lines of code into caching it. One weird thing is that the cover art packet will remain queued after seeks, but that is probably not an issue. In exchange, we can drop the dec_video.c code, which is pretty convenient for one of the following commits. This code duplicates a bunch of lower-level decode calls and does icky messing with this weird state stuff, so I'm glad it goes away.
2017-01-10 13:03:41 +00:00
mp_image_unrefp(&vo_c->cached_coverart);
player: fix some oversights in video refactoring vo_chain_uninit() isn't supposed to care much about the decoder (although decoders and outputs still go strictly together, so there is not much of an actual difference now). Also unset track.d_video correctly. Remove a stale declaration from dec_video.h as well.
2016-01-21 21:33:28 +00:00
vf_destroy(vo_c->vf);
video: decouple filtering/decoding slightly more Lots of noise to remove the vfilter/vo fields from dec_video. From now on, video filtering and output will still be done together, summarized under struct vo_chain. There is the question where exactly the vf_chain should go in such a decoupled architecture. The end goal is being able to place a "complex" filter between video decoders and output (which will culminate in natural integration of A->V filters for natural integration of libavfilter audio visualizations). The vf_chain is still useful for "final" processing, such as format conversions and deinterlacing. Also, there's only 1 VO and 1 --vf option. So having 1 vf_chain for a VO seems ideal, since otherwise there would be no natural way to handle all these existing options and mechanisms. There is still some work required to truly decouple decoding.
2016-01-13 23:18:48 +00:00
talloc_free(vo_c);
video: refactor: disentangle decoding/filtering some more This moves some code related to decoding from video.c to dec_video.c, and also removes some accesses to dec_video.c from the filtering code. dec_video.ch is starting to make sense, and simply returns video frames from a demuxer stream. The API exposed is also somewhat intended to be easily changeable to move decoding to a separate thread, if we ever want this (due to libavcodec already being threaded, I don't see much of a reason, but it might still be helpful).
2016-01-16 20:19:52 +00:00
// this does not free the VO
video: decouple filtering/decoding slightly more Lots of noise to remove the vfilter/vo fields from dec_video. From now on, video filtering and output will still be done together, summarized under struct vo_chain. There is the question where exactly the vf_chain should go in such a decoupled architecture. The end goal is being able to place a "complex" filter between video decoders and output (which will culminate in natural integration of A->V filters for natural integration of libavfilter audio visualizations). The vf_chain is still useful for "final" processing, such as format conversions and deinterlacing. Also, there's only 1 VO and 1 --vf option. So having 1 vf_chain for a VO seems ideal, since otherwise there would be no natural way to handle all these existing options and mechanisms. There is still some work required to truly decouple decoding.
2016-01-13 23:18:48 +00:00
}
player: remove central uninit_player() function and flags mess Each subsystem (or similar thing) had an INITIALIZED_ flag assigned. The main use of this was that you could pass a bitmask of these flags to uninit_player(). Except in some situations where you wanted to uninitialize nearly everything, this wasn't really useful. Moreover, it was quite annoying that subsystems had most of the code in a specific file, but the uninit code in loadfile.c (because that's where uninit_player() was implemented). Simplify all this. Remove the flags; e.g. instead of testing for the INITIALIZED_AO flag, test whether mpctx->ao is set. Move uninit code to separate functions, e.g. uninit_audio_out().
2014-10-03 17:57:49 +00:00
void uninit_video_chain(struct MPContext *mpctx)
{
player: refactor: eliminate MPContext.d_video Eventually we want the VO be driven by a A->V filter, so a decoder doesn't even have to exist. Some features definitely require a decoder though (like reporting the decoder in use, hardware decoding, etc.), so for each thing which accessed d_video, it has to be redecided if and how it can access decoder state. At least the "framedrop" property slightly changes semantics: you can now always set this property, even if no video is active. Some untested changes in this commit, but our bio-based distributed test suite has to take care of this.
2016-01-17 17:07:50 +00:00
if (mpctx->vo_chain) {
player: remove central uninit_player() function and flags mess Each subsystem (or similar thing) had an INITIALIZED_ flag assigned. The main use of this was that you could pass a bitmask of these flags to uninit_player(). Except in some situations where you wanted to uninitialize nearly everything, this wasn't really useful. Moreover, it was quite annoying that subsystems had most of the code in a specific file, but the uninit code in loadfile.c (because that's where uninit_player() was implemented). Simplify all this. Remove the flags; e.g. instead of testing for the INITIALIZED_AO flag, test whether mpctx->ao is set. Move uninit code to separate functions, e.g. uninit_audio_out().
2014-10-03 17:57:49 +00:00
reset_video_state(mpctx);
video: decouple filtering/decoding slightly more Lots of noise to remove the vfilter/vo fields from dec_video. From now on, video filtering and output will still be done together, summarized under struct vo_chain. There is the question where exactly the vf_chain should go in such a decoupled architecture. The end goal is being able to place a "complex" filter between video decoders and output (which will culminate in natural integration of A->V filters for natural integration of libavfilter audio visualizations). The vf_chain is still useful for "final" processing, such as format conversions and deinterlacing. Also, there's only 1 VO and 1 --vf option. So having 1 vf_chain for a VO seems ideal, since otherwise there would be no natural way to handle all these existing options and mechanisms. There is still some work required to truly decouple decoding.
2016-01-13 23:18:48 +00:00
vo_chain_uninit(mpctx->vo_chain);
mpctx->vo_chain = NULL;
player: add complex filter graph support See --lavfi-complex option. This is still quite rough. There's no support for dynamic configuration of any kind. There are probably corner cases where playback might freeze or burn 100% CPU (due to dataflow problems when interaction with libavfilter). Future possible plans might include: - freely switch tracks by providing some sort of default track graph label - automatically enabling audio visualization - automatically mix audio or stack video when multiple tracks are selected at once (similar to how multiple sub tracks can be selected)
2016-02-05 22:19:56 +00:00
player: remove central uninit_player() function and flags mess Each subsystem (or similar thing) had an INITIALIZED_ flag assigned. The main use of this was that you could pass a bitmask of these flags to uninit_player(). Except in some situations where you wanted to uninitialize nearly everything, this wasn't really useful. Moreover, it was quite annoying that subsystems had most of the code in a specific file, but the uninit code in loadfile.c (because that's where uninit_player() was implemented). Simplify all this. Remove the flags; e.g. instead of testing for the INITIALIZED_AO flag, test whether mpctx->ao is set. Move uninit code to separate functions, e.g. uninit_audio_out().
2014-10-03 17:57:49 +00:00
mpctx->video_status = STATUS_EOF;
player: slightly simplify how demuxer streams are enabled/disabled Instead of having reselect_demux_streams() look at all streams, make it look at the current stream that is being enabled/disabled.
2016-02-25 21:44:50 +00:00
player: fix excessive CPU usage in audio-only mode Caused by one of the --force-window commits. The unconditional uninit_video_out() call (which normally should be idempotent) raised sporadic MPV_EVENT_VIDEO_RECONFIG events. This is ok, except for the fact that clients (like a Lua script or libmpv users) would cause the event loop to run again after receiving it, triggering a feedback loop. Fix it by sending the events really only on a change.
2015-09-22 09:19:39 +00:00
mp_notify(mpctx, MPV_EVENT_VIDEO_RECONFIG, NULL);
player: remove central uninit_player() function and flags mess Each subsystem (or similar thing) had an INITIALIZED_ flag assigned. The main use of this was that you could pass a bitmask of these flags to uninit_player(). Except in some situations where you wanted to uninitialize nearly everything, this wasn't really useful. Moreover, it was quite annoying that subsystems had most of the code in a specific file, but the uninit code in loadfile.c (because that's where uninit_player() was implemented). Simplify all this. Remove the flags; e.g. instead of testing for the INITIALIZED_AO flag, test whether mpctx->ao is set. Move uninit code to separate functions, e.g. uninit_audio_out().
2014-10-03 17:57:49 +00:00
}
}
player: move audio and video decoder init to separate functions Preparation.
2016-02-05 20:50:37 +00:00
int init_video_decoder(struct MPContext *mpctx, struct track *track)
{
assert(!track->d_video);
if (!track->stream)
goto err_out;
track->d_video = talloc_zero(NULL, struct dec_video);
struct dec_video *d_video = track->d_video;
d_video->global = mpctx->global;
d_video->log = mp_log_new(d_video, mpctx->log, "!vd");
d_video->opts = mpctx->opts;
d_video->header = track->stream;
audio/video: expose codec info as separate field Preparation for the timeline rewrite. The codec will be able to change, the stream header not.
2016-02-15 19:34:45 +00:00
d_video->codec = track->stream->codec;
player: move audio and video decoder init to separate functions Preparation.
2016-02-05 20:50:37 +00:00
d_video->fps = d_video->header->codec->fps;
vo_opengl: add direct rendering support Can be enabled via --vd-lavc-dr=yes. See manpage additions for what it does. This reminds of the MPlayer -dr flag, but the implementation is completely different. It's the same basic concept: letting the decoder render into a GPU buffer to avoid a copy. Unlike MPlayer, this doesn't try to go through filters (libavfilter doesn't support this anyway). Unless a filter can work in-place, DR will be silently disabled. MPlayer had very complex semantics about buffer types and management (which apparently nobody ever understood) and weird restrictions that mostly limited it to mpeg2 style codecs. The mpv code does not do any of this, and just lets the decoder allocate an arbitrary number of untyped images. (No MPlayer code was used.) Parts of the code based on work by atomnuker (starting point for the generic code) and haasn (some GL definitions, some basic PBO code, and correct fencing).
2017-07-23 07:41:51 +00:00
d_video->vo = mpctx->vo_chain->vo;
client API: make mpv_opengl_cb_uninit_gl() behavior slightly nicer Instead of deselecting the video stream plainly, use the slightly more robust error_on_track() function. Also give it an error code (although I'm not sure if this one is confusing, it's better than the one before).
2016-09-09 08:59:09 +00:00
// Note: at least mpv_opengl_cb_uninit_gl() relies on being able to get
// rid of all references to the VO by destroying the VO chain. Thus,
// decoders not linked to vo_chain must not use the hwdec context.
player: move audio and video decoder init to separate functions Preparation.
2016-02-05 20:50:37 +00:00
if (mpctx->vo_chain)
video: refactor how VO exports hwdec device handles The main change is with video/hwdec.h. mp_hwdec_info is made opaque (and renamed to mp_hwdec_devices). Its accessors are mainly thread-safe (or documented where not), which makes the whole thing saner and cleaner. In particular, thread-safety rules become less subtle and more obvious. The new internal API makes it easier to support multiple OpenGL interop backends. (Although this is not done yet, and it's not clear whether it ever will.) This also removes all the API-specific fields from mp_hwdec_ctx and replaces them with a "ctx" field. For d3d in particular, we drop the mp_d3d_ctx struct completely, and pass the interfaces directly. Remove the emulation checks from vaapi.c and vdpau.c; they are pointless, and the checks that matter are done on the VO layer. The d3d hardware decoders might slightly change behavior: dxva2-copy will not use the VO device anymore if the VO supports proper interop. This pretty much assumes that any in such cases the VO will not use any form of exclusive mode, which makes using the VO device in copy mode unnecessary. This is a big refactor. Some things may be untested and could be broken.
2016-05-09 17:42:03 +00:00
d_video->hwdec_devs = mpctx->vo_chain->hwdec_devs;
player: move audio and video decoder init to separate functions Preparation.
2016-02-05 20:50:37 +00:00
MP_VERBOSE(d_video, "Container reported FPS: %f\n", d_video->fps);
if (d_video->opts->force_fps) {
d_video->fps = d_video->opts->force_fps;
MP_INFO(mpctx, "FPS forced to %5.3f.\n", d_video->fps);
MP_INFO(mpctx, "Use --no-correct-pts to force FPS based timing.\n");
}
video: remove pointless parameter indirection This is always the same value.
2016-02-15 19:28:36 +00:00
if (!video_init_best_codec(d_video))
player: move audio and video decoder init to separate functions Preparation.
2016-02-05 20:50:37 +00:00
goto err_out;
return 1;
err_out:
player: add complex filter graph support See --lavfi-complex option. This is still quite rough. There's no support for dynamic configuration of any kind. There are probably corner cases where playback might freeze or burn 100% CPU (due to dataflow problems when interaction with libavfilter). Future possible plans might include: - freely switch tracks by providing some sort of default track graph label - automatically enabling audio visualization - automatically mix audio or stack video when multiple tracks are selected at once (similar to how multiple sub tracks can be selected)
2016-02-05 22:19:56 +00:00
if (track->sink)
lavfi_set_connected(track->sink, false);
track->sink = NULL;
player: move audio and video decoder init to separate functions Preparation.
2016-02-05 20:50:37 +00:00
video_uninit(track->d_video);
track->d_video = NULL;
error_on_track(mpctx, track);
return 0;
}
Split mplayer.c mplayer.c was a bit too big. Split it into multiple files. I hope the way it's split makes sense. Maybe some things don't make too much sense, or go against intuition. These will fixed as soon as I notice them. Some files are a bit questionable (misc.c, osd.c, configfiles.c), and suggestions how to organize this better are welcome. Regressions are possible due to reorganized include statements. Obviously I didn't just copy mplayer.c's orgy of include statements, but recreated them for each file. It's easily possible that there are oversights and mistakes, which will show up on other platforms. There is one actual change: the public avutil.h include is removed from encode.h, and I tried to replace most FFMIN/FFMAX/av_clip uses. I consider using libavutil too much as dangerous, because the set of include files they recursively pull in is rather arbitrary and is different between FFmpeg and Libav.
2013-10-29 21:38:29 +00:00
int reinit_video_chain(struct MPContext *mpctx)
player: add complex filter graph support See --lavfi-complex option. This is still quite rough. There's no support for dynamic configuration of any kind. There are probably corner cases where playback might freeze or burn 100% CPU (due to dataflow problems when interaction with libavfilter). Future possible plans might include: - freely switch tracks by providing some sort of default track graph label - automatically enabling audio visualization - automatically mix audio or stack video when multiple tracks are selected at once (similar to how multiple sub tracks can be selected)
2016-02-05 22:19:56 +00:00
{
return reinit_video_chain_src(mpctx, NULL);
}
int reinit_video_chain_src(struct MPContext *mpctx, struct lavfi_pad *src)
Split mplayer.c mplayer.c was a bit too big. Split it into multiple files. I hope the way it's split makes sense. Maybe some things don't make too much sense, or go against intuition. These will fixed as soon as I notice them. Some files are a bit questionable (misc.c, osd.c, configfiles.c), and suggestions how to organize this better are welcome. Regressions are possible due to reorganized include statements. Obviously I didn't just copy mplayer.c's orgy of include statements, but recreated them for each file. It's easily possible that there are oversights and mistakes, which will show up on other platforms. There is one actual change: the public avutil.h include is removed from encode.h, and I tried to replace most FFMIN/FFMAX/av_clip uses. I consider using libavutil too much as dangerous, because the set of include files they recursively pull in is rather arbitrary and is different between FFmpeg and Libav.
2013-10-29 21:38:29 +00:00
{
player: add complex filter graph support See --lavfi-complex option. This is still quite rough. There's no support for dynamic configuration of any kind. There are probably corner cases where playback might freeze or burn 100% CPU (due to dataflow problems when interaction with libavfilter). Future possible plans might include: - freely switch tracks by providing some sort of default track graph label - automatically enabling audio visualization - automatically mix audio or stack video when multiple tracks are selected at once (similar to how multiple sub tracks can be selected)
2016-02-05 22:19:56 +00:00
struct track *track = NULL;
struct sh_stream *sh = NULL;
if (!src) {
track = mpctx->current_track[0][STREAM_VIDEO];
if (!track)
return 0;
player: remove dead code Fixes CID 1350055 and CID 1350054.
2016-02-12 15:05:46 +00:00
sh = track->stream;
player: add complex filter graph support See --lavfi-complex option. This is still quite rough. There's no support for dynamic configuration of any kind. There are probably corner cases where playback might freeze or burn 100% CPU (due to dataflow problems when interaction with libavfilter). Future possible plans might include: - freely switch tracks by providing some sort of default track graph label - automatically enabling audio visualization - automatically mix audio or stack video when multiple tracks are selected at once (similar to how multiple sub tracks can be selected)
2016-02-05 22:19:56 +00:00
if (!sh)
goto no_video;
}
video: decouple filtering/decoding slightly more Lots of noise to remove the vfilter/vo fields from dec_video. From now on, video filtering and output will still be done together, summarized under struct vo_chain. There is the question where exactly the vf_chain should go in such a decoupled architecture. The end goal is being able to place a "complex" filter between video decoders and output (which will culminate in natural integration of A->V filters for natural integration of libavfilter audio visualizations). The vf_chain is still useful for "final" processing, such as format conversions and deinterlacing. Also, there's only 1 VO and 1 --vf option. So having 1 vf_chain for a VO seems ideal, since otherwise there would be no natural way to handle all these existing options and mechanisms. There is still some work required to truly decouple decoding.
2016-01-13 23:18:48 +00:00
assert(!mpctx->vo_chain);
Split mplayer.c mplayer.c was a bit too big. Split it into multiple files. I hope the way it's split makes sense. Maybe some things don't make too much sense, or go against intuition. These will fixed as soon as I notice them. Some files are a bit questionable (misc.c, osd.c, configfiles.c), and suggestions how to organize this better are welcome. Regressions are possible due to reorganized include statements. Obviously I didn't just copy mplayer.c's orgy of include statements, but recreated them for each file. It's easily possible that there are oversights and mistakes, which will show up on other platforms. There is one actual change: the public avutil.h include is removed from encode.h, and I tried to replace most FFMIN/FFMAX/av_clip uses. I consider using libavutil too much as dangerous, because the set of include files they recursively pull in is rather arbitrary and is different between FFmpeg and Libav.
2013-10-29 21:38:29 +00:00
player: remove --fixed-vo In ancient times, this was needed because it was not default, and many VOs had problems with it. But it was always default in mpv, and all VOs are required to deal with it. Also, running --fixed-vo=no is not useful and just creates weird corner cases. Get rid of it.
2015-02-03 22:09:02 +00:00
if (!mpctx->video_out) {
video: pass some VO params as struct Not particularly elegant, but better than adding more and more stuff to the relevant function signatures.
2014-12-31 18:01:28 +00:00
struct vo_extra ex = {
.input_ctx = mpctx->input,
.osd = mpctx->osd,
.encode_lavc_ctx = mpctx->encode_lavc_ctx,
vo_opengl_cb: pass context directly This is simpler than setting the context after VO creation, which requires the code to check for the context on every entrypoint.
2014-12-31 18:12:44 +00:00
.opengl_cb_context = mpctx->gl_cb_ctx,
player, ao, vo: don't call mp_input_wakeup() directly Currently, calling mp_input_wakeup() will wake up the core thread (also called the playloop). This seems odd, but currently the core indeed calls mp_input_wait() when it has nothing more to do. It's done this way because MPlayer used input_ctx as central "mainloop". This is probably going to change. Remove direct calls to this function, and replace it with mp_wakeup_core() calls. ao and vo are changed to use opaque callbacks and not use input_ctx for this purpose. Other code already uses opaque callbacks, or has legitimate reasons to use input_ctx directly (such as sending actual user input).
2016-09-16 12:23:54 +00:00
.wakeup_cb = mp_wakeup_core_cb,
.wakeup_ctx = mpctx,
video: pass some VO params as struct Not particularly elegant, but better than adding more and more stuff to the relevant function signatures.
2014-12-31 18:01:28 +00:00
};
mpctx->video_out = init_best_video_out(mpctx->global, &ex);
Split mplayer.c mplayer.c was a bit too big. Split it into multiple files. I hope the way it's split makes sense. Maybe some things don't make too much sense, or go against intuition. These will fixed as soon as I notice them. Some files are a bit questionable (misc.c, osd.c, configfiles.c), and suggestions how to organize this better are welcome. Regressions are possible due to reorganized include statements. Obviously I didn't just copy mplayer.c's orgy of include statements, but recreated them for each file. It's easily possible that there are oversights and mistakes, which will show up on other platforms. There is one actual change: the public avutil.h include is removed from encode.h, and I tried to replace most FFMIN/FFMAX/av_clip uses. I consider using libavutil too much as dangerous, because the set of include files they recursively pull in is rather arbitrary and is different between FFmpeg and Libav.
2013-10-29 21:38:29 +00:00
if (!mpctx->video_out) {
MP_FATAL(mpctx, "Error opening/initializing "
manpage: replace `-vo` with `--vo`
2016-12-08 17:08:04 +00:00
"the selected video_out (--vo) device.\n");
client API: better error reporting Give somewhat more information on playback failure.
2014-10-28 15:19:07 +00:00
mpctx->error_playing = MPV_ERROR_VO_INIT_FAILED;
Split mplayer.c mplayer.c was a bit too big. Split it into multiple files. I hope the way it's split makes sense. Maybe some things don't make too much sense, or go against intuition. These will fixed as soon as I notice them. Some files are a bit questionable (misc.c, osd.c, configfiles.c), and suggestions how to organize this better are welcome. Regressions are possible due to reorganized include statements. Obviously I didn't just copy mplayer.c's orgy of include statements, but recreated them for each file. It's easily possible that there are oversights and mistakes, which will show up on other platforms. There is one actual change: the public avutil.h include is removed from encode.h, and I tried to replace most FFMIN/FFMAX/av_clip uses. I consider using libavutil too much as dangerous, because the set of include files they recursively pull in is rather arbitrary and is different between FFmpeg and Libav.
2013-10-29 21:38:29 +00:00
goto err_out;
}
mpctx->mouse_cursor_visible = true;
}
player: set PulseAudio stream title to window title Set the PulseAudio stream title, just like the VO window title is set. Refactor update_vo_window_title() so that we can use it for AOs too. The ao_pulse.c bit is stolen from MPlayer.
2013-11-09 23:49:13 +00:00
update_window_title(mpctx, true);
Split mplayer.c mplayer.c was a bit too big. Split it into multiple files. I hope the way it's split makes sense. Maybe some things don't make too much sense, or go against intuition. These will fixed as soon as I notice them. Some files are a bit questionable (misc.c, osd.c, configfiles.c), and suggestions how to organize this better are welcome. Regressions are possible due to reorganized include statements. Obviously I didn't just copy mplayer.c's orgy of include statements, but recreated them for each file. It's easily possible that there are oversights and mistakes, which will show up on other platforms. There is one actual change: the public avutil.h include is removed from encode.h, and I tried to replace most FFMIN/FFMAX/av_clip uses. I consider using libavutil too much as dangerous, because the set of include files they recursively pull in is rather arbitrary and is different between FFmpeg and Libav.
2013-10-29 21:38:29 +00:00
player: refactor: eliminate MPContext.d_video Eventually we want the VO be driven by a A->V filter, so a decoder doesn't even have to exist. Some features definitely require a decoder though (like reporting the decoder in use, hardware decoding, etc.), so for each thing which accessed d_video, it has to be redecided if and how it can access decoder state. At least the "framedrop" property slightly changes semantics: you can now always set this property, even if no video is active. Some untested changes in this commit, but our bio-based distributed test suite has to take care of this.
2016-01-17 17:07:50 +00:00
struct vo_chain *vo_c = talloc_zero(NULL, struct vo_chain);
mpctx->vo_chain = vo_c;
vo_c->log = mpctx->log;
vo_c->vo = mpctx->video_out;
player: fix crash if no video decoder can be initialized Caused by the recent refactoring for complex filters.
2016-02-09 23:07:01 +00:00
vo_c->vf = vf_new(mpctx->global);
player: refactor: eliminate MPContext.d_video Eventually we want the VO be driven by a A->V filter, so a decoder doesn't even have to exist. Some features definitely require a decoder though (like reporting the decoder in use, hardware decoding, etc.), so for each thing which accessed d_video, it has to be redecided if and how it can access decoder state. At least the "framedrop" property slightly changes semantics: you can now always set this property, even if no video is active. Some untested changes in this commit, but our bio-based distributed test suite has to take care of this.
2016-01-17 17:07:50 +00:00
video: refactor how VO exports hwdec device handles The main change is with video/hwdec.h. mp_hwdec_info is made opaque (and renamed to mp_hwdec_devices). Its accessors are mainly thread-safe (or documented where not), which makes the whole thing saner and cleaner. In particular, thread-safety rules become less subtle and more obvious. The new internal API makes it easier to support multiple OpenGL interop backends. (Although this is not done yet, and it's not clear whether it ever will.) This also removes all the API-specific fields from mp_hwdec_ctx and replaces them with a "ctx" field. For d3d in particular, we drop the mp_d3d_ctx struct completely, and pass the interfaces directly. Remove the emulation checks from vaapi.c and vdpau.c; they are pointless, and the checks that matter are done on the VO layer. The d3d hardware decoders might slightly change behavior: dxva2-copy will not use the VO device anymore if the VO supports proper interop. This pretty much assumes that any in such cases the VO will not use any form of exclusive mode, which makes using the VO device in copy mode unnecessary. This is a big refactor. Some things may be untested and could be broken.
2016-05-09 17:42:03 +00:00
vo_c->hwdec_devs = vo_c->vo->hwdec_devs;
player: refactor: eliminate MPContext.d_video Eventually we want the VO be driven by a A->V filter, so a decoder doesn't even have to exist. Some features definitely require a decoder though (like reporting the decoder in use, hardware decoding, etc.), so for each thing which accessed d_video, it has to be redecided if and how it can access decoder state. At least the "framedrop" property slightly changes semantics: you can now always set this property, even if no video is active. Some untested changes in this commit, but our bio-based distributed test suite has to take care of this.
2016-01-17 17:07:50 +00:00
lavfi: support hwdec filters for --lavfi-complex Not so important by itself, but important for when we replace the vf libavfilter wrapper with the common implementation. (Which will hopefully happen, but not too soon.)
2017-02-20 12:42:08 +00:00
if (mpctx->lavfi)
lavfi_set_hwdec_devs(mpctx->lavfi, vo_c->hwdec_devs);
player: add complex filter graph support See --lavfi-complex option. This is still quite rough. There's no support for dynamic configuration of any kind. There are probably corner cases where playback might freeze or burn 100% CPU (due to dataflow problems when interaction with libavfilter). Future possible plans might include: - freely switch tracks by providing some sort of default track graph label - automatically enabling audio visualization - automatically mix audio or stack video when multiple tracks are selected at once (similar to how multiple sub tracks can be selected)
2016-02-05 22:19:56 +00:00
vo_c->filter_src = src;
if (!vo_c->filter_src) {
vo_c->track = track;
track->vo_c = vo_c;
if (!init_video_decoder(mpctx, track))
goto err_out;
vo_c->video_src = track->d_video;
vo_c->container_fps = vo_c->video_src->fps;
vo_c->is_coverart = !!sh->attached_picture;
player: cleanup update_fps() function It was called only in 2 places, one of them redundant (the container FPS can not change).
2015-04-20 21:21:38 +00:00
player: add complex filter graph support See --lavfi-complex option. This is still quite rough. There's no support for dynamic configuration of any kind. There are probably corner cases where playback might freeze or burn 100% CPU (due to dataflow problems when interaction with libavfilter). Future possible plans might include: - freely switch tracks by providing some sort of default track graph label - automatically enabling audio visualization - automatically mix audio or stack video when multiple tracks are selected at once (similar to how multiple sub tracks can be selected)
2016-02-05 22:19:56 +00:00
track->vo_c = vo_c;
vo_c->track = track;
}
player: refactor: eliminate MPContext.d_video Eventually we want the VO be driven by a A->V filter, so a decoder doesn't even have to exist. Some features definitely require a decoder though (like reporting the decoder in use, hardware decoding, etc.), so for each thing which accessed d_video, it has to be redecided if and how it can access decoder state. At least the "framedrop" property slightly changes semantics: you can now always set this property, even if no video is active. Some untested changes in this commit, but our bio-based distributed test suite has to take care of this.
2016-01-17 17:07:50 +00:00
player: cleanup update_fps() function It was called only in 2 places, one of them redundant (the container FPS can not change).
2015-04-20 21:21:38 +00:00
#if HAVE_ENCODING
player: remove redundant check Found by Coverity.
2015-12-05 22:53:48 +00:00
if (mpctx->encode_lavc_ctx)
player: move audio and video decoder init to separate functions Preparation.
2016-02-05 20:50:37 +00:00
encode_lavc_set_video_fps(mpctx->encode_lavc_ctx, vo_c->container_fps);
player: cleanup update_fps() function It was called only in 2 places, one of them redundant (the container FPS can not change).
2015-04-20 21:21:38 +00:00
#endif
vf_vapoursynth: pass through container FPS value This is basically a hack; but apparently a needed one, since many vapoursynth filters insist on having a FPS set. We need to apply the FPS override before creating the filters. Also change some terminal output related to the FPS value.
2015-01-03 02:37:05 +00:00
Split mplayer.c mplayer.c was a bit too big. Split it into multiple files. I hope the way it's split makes sense. Maybe some things don't make too much sense, or go against intuition. These will fixed as soon as I notice them. Some files are a bit questionable (misc.c, osd.c, configfiles.c), and suggestions how to organize this better are welcome. Regressions are possible due to reorganized include statements. Obviously I didn't just copy mplayer.c's orgy of include statements, but recreated them for each file. It's easily possible that there are oversights and mistakes, which will show up on other platforms. There is one actual change: the public avutil.h include is removed from encode.h, and I tried to replace most FFMIN/FFMAX/av_clip uses. I consider using libavutil too much as dangerous, because the set of include files they recursively pull in is rather arbitrary and is different between FFmpeg and Libav.
2013-10-29 21:38:29 +00:00
recreate_video_filters(mpctx);
player: make --stop-screensaver runtime-changeable Move the screensaver enable/disable determination to a central place, and call it if the stop-screensaver property is changed. Also, do not stop the screensaver when in idle mode (i.e. no file is loaded). Fixes #3615.
2016-10-02 10:33:34 +00:00
update_screensaver_state(mpctx);
Split mplayer.c mplayer.c was a bit too big. Split it into multiple files. I hope the way it's split makes sense. Maybe some things don't make too much sense, or go against intuition. These will fixed as soon as I notice them. Some files are a bit questionable (misc.c, osd.c, configfiles.c), and suggestions how to organize this better are welcome. Regressions are possible due to reorganized include statements. Obviously I didn't just copy mplayer.c's orgy of include statements, but recreated them for each file. It's easily possible that there are oversights and mistakes, which will show up on other platforms. There is one actual change: the public avutil.h include is removed from encode.h, and I tried to replace most FFMIN/FFMAX/av_clip uses. I consider using libavutil too much as dangerous, because the set of include files they recursively pull in is rather arbitrary and is different between FFmpeg and Libav.
2013-10-29 21:38:29 +00:00
player: refactor: eliminate MPContext.d_video Eventually we want the VO be driven by a A->V filter, so a decoder doesn't even have to exist. Some features definitely require a decoder though (like reporting the decoder in use, hardware decoding, etc.), so for each thing which accessed d_video, it has to be redecided if and how it can access decoder state. At least the "framedrop" property slightly changes semantics: you can now always set this property, even if no video is active. Some untested changes in this commit, but our bio-based distributed test suite has to take care of this.
2016-01-17 17:07:50 +00:00
vo_set_paused(vo_c->vo, mpctx->paused);
Split mplayer.c mplayer.c was a bit too big. Split it into multiple files. I hope the way it's split makes sense. Maybe some things don't make too much sense, or go against intuition. These will fixed as soon as I notice them. Some files are a bit questionable (misc.c, osd.c, configfiles.c), and suggestions how to organize this better are welcome. Regressions are possible due to reorganized include statements. Obviously I didn't just copy mplayer.c's orgy of include statements, but recreated them for each file. It's easily possible that there are oversights and mistakes, which will show up on other platforms. There is one actual change: the public avutil.h include is removed from encode.h, and I tried to replace most FFMIN/FFMAX/av_clip uses. I consider using libavutil too much as dangerous, because the set of include files they recursively pull in is rather arbitrary and is different between FFmpeg and Libav.
2013-10-29 21:38:29 +00:00
audio: change playback restart and resyncing This commit makes audio decoding non-blocking. If e.g. the network is too slow the playloop will just go to sleep, instead of blocking until enough data is available. For video, this was already done with commit 7083f88c. For audio, it's unfortunately much more complicated, because the audio decoder was used in a blocking manner. Large changes are required to get around this. The whole playback restart mechanism must be turned into a statemachine, especially since it has close interactions with video restart. Lots of video code is thus also changed. (For the record, I don't think switching this code to threads would make this conceptually easier: the code would still have to deal with external input while blocked, so these in-between states do get visible [and thus need to be handled] anyway. On the other hand, it certainly should be possible to modularize this code a bit better.) This will probably cause a bunch of regressions.
2014-07-28 18:40:43 +00:00
// If we switch on video again, ensure audio position matches up.
player: refactor: eliminate MPContext.d_audio
2016-01-21 21:24:20 +00:00
if (mpctx->ao_chain)
audio: change playback restart and resyncing This commit makes audio decoding non-blocking. If e.g. the network is too slow the playloop will just go to sleep, instead of blocking until enough data is available. For video, this was already done with commit 7083f88c. For audio, it's unfortunately much more complicated, because the audio decoder was used in a blocking manner. Large changes are required to get around this. The whole playback restart mechanism must be turned into a statemachine, especially since it has close interactions with video restart. Lots of video code is thus also changed. (For the record, I don't think switching this code to threads would make this conceptually easier: the code would still have to deal with external input while blocked, so these in-between states do get visible [and thus need to be handled] anyway. On the other hand, it certainly should be possible to modularize this code a bit better.) This will probably cause a bunch of regressions.
2014-07-28 18:40:43 +00:00
mpctx->audio_status = STATUS_SYNCING;
player: split seek_reset() This also reduces some code duplication with other parts of the code. The changfe is mostly cosmetic, although there are also some subtle changes in behavior. At least one change is that the big desync message is now printed after every seek.
2014-07-30 21:01:55 +00:00
reset_video_state(mpctx);
reset_subtitle_state(mpctx);
Split mplayer.c mplayer.c was a bit too big. Split it into multiple files. I hope the way it's split makes sense. Maybe some things don't make too much sense, or go against intuition. These will fixed as soon as I notice them. Some files are a bit questionable (misc.c, osd.c, configfiles.c), and suggestions how to organize this better are welcome. Regressions are possible due to reorganized include statements. Obviously I didn't just copy mplayer.c's orgy of include statements, but recreated them for each file. It's easily possible that there are oversights and mistakes, which will show up on other platforms. There is one actual change: the public avutil.h include is removed from encode.h, and I tried to replace most FFMIN/FFMAX/av_clip uses. I consider using libavutil too much as dangerous, because the set of include files they recursively pull in is rather arbitrary and is different between FFmpeg and Libav.
2013-10-29 21:38:29 +00:00
return 1;
err_out:
no_video:
player: remove central uninit_player() function and flags mess Each subsystem (or similar thing) had an INITIALIZED_ flag assigned. The main use of this was that you could pass a bitmask of these flags to uninit_player(). Except in some situations where you wanted to uninitialize nearly everything, this wasn't really useful. Moreover, it was quite annoying that subsystems had most of the code in a specific file, but the uninit code in loadfile.c (because that's where uninit_player() was implemented). Simplify all this. Remove the flags; e.g. instead of testing for the INITIALIZED_AO flag, test whether mpctx->ao is set. Move uninit code to separate functions, e.g. uninit_audio_out().
2014-10-03 17:57:49 +00:00
uninit_video_chain(mpctx);
player: remove some further current_track dependencies Now it's used for initialization only for audio and video.
2016-02-05 22:41:44 +00:00
error_on_track(mpctx, track);
Split mplayer.c mplayer.c was a bit too big. Split it into multiple files. I hope the way it's split makes sense. Maybe some things don't make too much sense, or go against intuition. These will fixed as soon as I notice them. Some files are a bit questionable (misc.c, osd.c, configfiles.c), and suggestions how to organize this better are welcome. Regressions are possible due to reorganized include statements. Obviously I didn't just copy mplayer.c's orgy of include statements, but recreated them for each file. It's easily possible that there are oversights and mistakes, which will show up on other platforms. There is one actual change: the public avutil.h include is removed from encode.h, and I tried to replace most FFMIN/FFMAX/av_clip uses. I consider using libavutil too much as dangerous, because the set of include files they recursively pull in is rather arbitrary and is different between FFmpeg and Libav.
2013-10-29 21:38:29 +00:00
handle_force_window(mpctx, true);
return 0;
}
// Try to refresh the video by doing a precise seek to the currently displayed
// frame. This can go wrong in all sorts of ways, so use sparingly.
void mp_force_video_refresh(struct MPContext *mpctx)
{
struct MPOpts *opts = mpctx->opts;
video: refactor: disentangle decoding/filtering some more This moves some code related to decoding from video.c to dec_video.c, and also removes some accesses to dec_video.c from the filtering code. dec_video.ch is starting to make sense, and simply returns video frames from a demuxer stream. The API exposed is also somewhat intended to be easily changeable to move decoding to a separate thread, if we ever want this (due to libavcodec already being threaded, I don't see much of a reason, but it might still be helpful).
2016-01-16 20:19:52 +00:00
struct vo_chain *vo_c = mpctx->vo_chain;
command: allow changing filters before video chain initialization Apparently this is more intuitive. Somewhat tricky, because of the odd state after loading a file but before initializing the VO.
2014-03-30 17:21:54 +00:00
video: refactor: disentangle decoding/filtering some more This moves some code related to decoding from video.c to dec_video.c, and also removes some accesses to dec_video.c from the filtering code. dec_video.ch is starting to make sense, and simply returns video frames from a demuxer stream. The API exposed is also somewhat intended to be easily changeable to move decoding to a separate thread, if we ever want this (due to libavcodec already being threaded, I don't see much of a reason, but it might still be helpful).
2016-01-16 20:19:52 +00:00
if (!vo_c || !vo_c->input_format.imgfmt)
command: allow changing filters before video chain initialization Apparently this is more intuitive. Somewhat tricky, because of the odd state after loading a file but before initializing the VO.
2014-03-30 17:21:54 +00:00
return;
Split mplayer.c mplayer.c was a bit too big. Split it into multiple files. I hope the way it's split makes sense. Maybe some things don't make too much sense, or go against intuition. These will fixed as soon as I notice them. Some files are a bit questionable (misc.c, osd.c, configfiles.c), and suggestions how to organize this better are welcome. Regressions are possible due to reorganized include statements. Obviously I didn't just copy mplayer.c's orgy of include statements, but recreated them for each file. It's easily possible that there are oversights and mistakes, which will show up on other platforms. There is one actual change: the public avutil.h include is removed from encode.h, and I tried to replace most FFMIN/FFMAX/av_clip uses. I consider using libavutil too much as dangerous, because the set of include files they recursively pull in is rather arbitrary and is different between FFmpeg and Libav.
2013-10-29 21:38:29 +00:00
// If not paused, the next frame should come soon enough.
player: refresh very low framerate video on filter changes Limit the max. time the refresh is delayed. Make it refresh at all if image mode is enabled. Fixes #3435.
2016-08-19 07:38:38 +00:00
if ((opts->pause || mpctx->time_frame >= 0.5) &&
(mpctx->video_status >= STATUS_PLAYING ||
mpctx->video_status <= STATUS_DRAINING) &&
video: don't force video refresh if video is restarting
2015-07-10 16:07:39 +00:00
mpctx->last_vo_pts != MP_NOPTS_VALUE)
{
player: use symbolic constant for seek precision Meh.
2015-03-04 16:21:05 +00:00
queue_seek(mpctx, MPSEEK_ABSOLUTE, mpctx->last_vo_pts,
player: allow passing flags to queue_seek() Change the last parameter from a bool to an int, which is supposed to take bit-flags. The at this point only flag is MPSEEK_FLAG_DELAY, which replaces the previous bool parameter. The old false parameter becomes 0, the old true parameter becomes MPSEEK_FLAG_DELAY. Since the old "immediate" parameter is now essentially inverted, two coalesced immediate and delayed seeks end up as delayed instead of immediate. This change doesn't matter, since there are no relative immediate seeks anyway.
2016-08-15 19:07:32 +00:00
MPSEEK_VERY_EXACT, 0);
player: use symbolic constant for seek precision Meh.
2015-03-04 16:21:05 +00:00
}
Split mplayer.c mplayer.c was a bit too big. Split it into multiple files. I hope the way it's split makes sense. Maybe some things don't make too much sense, or go against intuition. These will fixed as soon as I notice them. Some files are a bit questionable (misc.c, osd.c, configfiles.c), and suggestions how to organize this better are welcome. Regressions are possible due to reorganized include statements. Obviously I didn't just copy mplayer.c's orgy of include statements, but recreated them for each file. It's easily possible that there are oversights and mistakes, which will show up on other platforms. There is one actual change: the public avutil.h include is removed from encode.h, and I tried to replace most FFMIN/FFMAX/av_clip uses. I consider using libavutil too much as dangerous, because the set of include files they recursively pull in is rather arbitrary and is different between FFmpeg and Libav.
2013-10-29 21:38:29 +00:00
}
player: refactor: eliminate MPContext.d_video Eventually we want the VO be driven by a A->V filter, so a decoder doesn't even have to exist. Some features definitely require a decoder though (like reporting the decoder in use, hardware decoding, etc.), so for each thing which accessed d_video, it has to be redecided if and how it can access decoder state. At least the "framedrop" property slightly changes semantics: you can now always set this property, even if no video is active. Some untested changes in this commit, but our bio-based distributed test suite has to take care of this.
2016-01-17 17:07:50 +00:00
static bool check_framedrop(struct MPContext *mpctx, struct vo_chain *vo_c)
video: change everything Change how the video decoding loop works. The structure should now be a bit easier to follow. The interactions on format changes are (probably) simpler. This also aligns the decoding loop with future planned changes, such as moving various things to separate threads.
2014-05-01 21:53:18 +00:00
{
struct MPOpts *opts = mpctx->opts;
// check for frame-drop:
video: minor simplification of the old framedrop code No changes in functionality, other than being slightly more correct at stream EOF.
2014-08-13 19:50:23 +00:00
if (mpctx->video_status == STATUS_PLAYING && !mpctx->paused &&
player: refactor: some more minor decoder/output decoupling These changes don't make too much sense without context, but are preparation for later. Then the audio_src/video_src fields will be actually be NULL under circumstances.
2016-01-29 21:46:28 +00:00
mpctx->audio_status == STATUS_PLAYING && !ao_untimed(mpctx->ao) &&
vo_c->video_src)
video: change everything Change how the video decoding loop works. The structure should now be a bit easier to follow. The interactions on format changes are (probably) simpler. This also aligns the decoding loop with future planned changes, such as moving various things to separate threads.
2014-05-01 21:53:18 +00:00
{
player: refactor: eliminate MPContext.d_video Eventually we want the VO be driven by a A->V filter, so a decoder doesn't even have to exist. Some features definitely require a decoder though (like reporting the decoder in use, hardware decoding, etc.), so for each thing which accessed d_video, it has to be redecided if and how it can access decoder state. At least the "framedrop" property slightly changes semantics: you can now always set this property, even if no video is active. Some untested changes in this commit, but our bio-based distributed test suite has to take care of this.
2016-01-17 17:07:50 +00:00
float fps = vo_c->container_fps;
video: minor simplification of the old framedrop code No changes in functionality, other than being slightly more correct at stream EOF.
2014-08-13 19:50:23 +00:00
double frame_time = fps > 0 ? 1.0 / fps : 0;
video: change everything Change how the video decoding loop works. The structure should now be a bit easier to follow. The interactions on format changes are (probably) simpler. This also aligns the decoding loop with future planned changes, such as moving various things to separate threads.
2014-05-01 21:53:18 +00:00
// we should avoid dropping too many frames in sequence unless we
// are too late. and we allow 100ms A-V delay here:
video: refactor: disentangle decoding/filtering some more This moves some code related to decoding from video.c to dec_video.c, and also removes some accesses to dec_video.c from the filtering code. dec_video.ch is starting to make sense, and simply returns video frames from a demuxer stream. The API exposed is also somewhat intended to be easily changeable to move decoding to a separate thread, if we ever want this (due to libavcodec already being threaded, I don't see much of a reason, but it might still be helpful).
2016-01-16 20:19:52 +00:00
int dropped_frames =
player: refactor: eliminate MPContext.d_video Eventually we want the VO be driven by a A->V filter, so a decoder doesn't even have to exist. Some features definitely require a decoder though (like reporting the decoder in use, hardware decoding, etc.), so for each thing which accessed d_video, it has to be redecided if and how it can access decoder state. At least the "framedrop" property slightly changes semantics: you can now always set this property, even if no video is active. Some untested changes in this commit, but our bio-based distributed test suite has to take care of this.
2016-01-17 17:07:50 +00:00
vo_c->video_src->dropped_frames - mpctx->dropped_frames_start;
video: refactor: disentangle decoding/filtering some more This moves some code related to decoding from video.c to dec_video.c, and also removes some accesses to dec_video.c from the filtering code. dec_video.ch is starting to make sense, and simply returns video frames from a demuxer stream. The API exposed is also somewhat intended to be easily changeable to move decoding to a separate thread, if we ever want this (due to libavcodec already being threaded, I don't see much of a reason, but it might still be helpful).
2016-01-16 20:19:52 +00:00
if (mpctx->last_av_difference - 0.100 > dropped_frames * frame_time)
video: improve decoder-based framedropping mode This is the "old" framedropping mode (derived from MPlayer). At least in the mplayer2/mpv source base, it stopped working properly years ago (or maybe it never worked properly). For one, it depends on the video framerate, which assume constant framerate. Another problem was that it could lead to freezing video display: video could get so much behind that it couldn't recover from framedrop. Make some small changes to improve this. Don't use the current audio position to check how much we are behind. Instead, use the last known A/V difference. last_av_difference is updated only when a video frame is scheduled for display. This means we can keep stop dropping once we're done catching up, even if video is technically still behind. What helps us here that this forces a video frame to be displayed after a while. Likewise, we reset the dropped_frames count only when scheduling a new frame for display as well. Some inspiration was taken from earlier work by xnor (see issue #620), although the implementation turned out quite different. This still uses the demuxer-reported (possibly broken) FPS value. It also doesn't account for filters changing FPS. We can't do much about this, because without decoding _and_ filtering, we just can't know how long a frame is. In theory, you could derive that from the raw packet timestamps and the filter chain contents, but actually doing this is too involved. Fortunately, the main thing the FPS affects is actually the displayed framedrop count.
2014-09-19 21:46:20 +00:00
return !!(opts->frame_dropping & 2);
video: change everything Change how the video decoding loop works. The structure should now be a bit easier to follow. The interactions on format changes are (probably) simpler. This also aligns the decoding loop with future planned changes, such as moving various things to separate threads.
2014-05-01 21:53:18 +00:00
}
video: refactor: disentangle decoding/filtering some more This moves some code related to decoding from video.c to dec_video.c, and also removes some accesses to dec_video.c from the filtering code. dec_video.ch is starting to make sense, and simply returns video frames from a demuxer stream. The API exposed is also somewhat intended to be easily changeable to move decoding to a separate thread, if we ever want this (due to libavcodec already being threaded, I don't see much of a reason, but it might still be helpful).
2016-01-16 20:19:52 +00:00
return false;
player: make decoding cover art more robust When showing cover art, the decoding logic pretends that the source has an infinite number of frames. This slightly simplifies dealing with filter data flow. It was done by feeding the same packet repeatedly to the decoder (each decode run produces new output). Change this by decoding once at the video initialization. This is easier to follow, and increases robustness in case of broken images. Usually, we try to tolerate decoding errors, so decoding normally continues, but in this case it would just burn the CPU for no reason. Fixes #2056.
2015-06-18 16:39:46 +00:00
}
video: change everything Change how the video decoding loop works. The structure should now be a bit easier to follow. The interactions on format changes are (probably) simpler. This also aligns the decoding loop with future planned changes, such as moving various things to separate threads.
2014-05-01 21:53:18 +00:00
// Read a packet, store decoded image into d_video->waiting_decoded_mpi
video: use symbolic constants instead of magic integers In my opinion this is not really necessary, since there's only a single user of update_video(), but others reading this code would probably hate me for using magic integer values instead of symbolic constants. This should be a purely cosmetic commit; any changes in behavior are bugs.
2014-07-18 13:11:21 +00:00
// returns VD_* code
video: change everything Change how the video decoding loop works. The structure should now be a bit easier to follow. The interactions on format changes are (probably) simpler. This also aligns the decoding loop with future planned changes, such as moving various things to separate threads.
2014-05-01 21:53:18 +00:00
static int decode_image(struct MPContext *mpctx)
Split mplayer.c mplayer.c was a bit too big. Split it into multiple files. I hope the way it's split makes sense. Maybe some things don't make too much sense, or go against intuition. These will fixed as soon as I notice them. Some files are a bit questionable (misc.c, osd.c, configfiles.c), and suggestions how to organize this better are welcome. Regressions are possible due to reorganized include statements. Obviously I didn't just copy mplayer.c's orgy of include statements, but recreated them for each file. It's easily possible that there are oversights and mistakes, which will show up on other platforms. There is one actual change: the public avutil.h include is removed from encode.h, and I tried to replace most FFMIN/FFMAX/av_clip uses. I consider using libavutil too much as dangerous, because the set of include files they recursively pull in is rather arbitrary and is different between FFmpeg and Libav.
2013-10-29 21:38:29 +00:00
{
video: refactor: disentangle decoding/filtering some more This moves some code related to decoding from video.c to dec_video.c, and also removes some accesses to dec_video.c from the filtering code. dec_video.ch is starting to make sense, and simply returns video frames from a demuxer stream. The API exposed is also somewhat intended to be easily changeable to move decoding to a separate thread, if we ever want this (due to libavcodec already being threaded, I don't see much of a reason, but it might still be helpful).
2016-01-16 20:19:52 +00:00
struct vo_chain *vo_c = mpctx->vo_chain;
player: add complex filter graph support See --lavfi-complex option. This is still quite rough. There's no support for dynamic configuration of any kind. There are probably corner cases where playback might freeze or burn 100% CPU (due to dataflow problems when interaction with libavfilter). Future possible plans might include: - freely switch tracks by providing some sort of default track graph label - automatically enabling audio visualization - automatically mix audio or stack video when multiple tracks are selected at once (similar to how multiple sub tracks can be selected)
2016-02-05 22:19:56 +00:00
if (vo_c->input_mpi)
return VD_PROGRESS;
int res = DATA_EOF;
if (vo_c->filter_src) {
res = lavfi_request_frame_v(vo_c->filter_src, &vo_c->input_mpi);
} else if (vo_c->video_src) {
struct dec_video *d_video = vo_c->video_src;
bool hrseek = mpctx->hrseek_active && mpctx->hrseek_framedrop &&
mpctx->video_status == STATUS_SYNCING;
video_set_start(d_video, hrseek ? mpctx->hrseek_pts : MP_NOPTS_VALUE);
video_set_framedrop(d_video, check_framedrop(mpctx, vo_c));
video_work(d_video);
res = video_get_frame(d_video, &vo_c->input_mpi);
}
Split mplayer.c mplayer.c was a bit too big. Split it into multiple files. I hope the way it's split makes sense. Maybe some things don't make too much sense, or go against intuition. These will fixed as soon as I notice them. Some files are a bit questionable (misc.c, osd.c, configfiles.c), and suggestions how to organize this better are welcome. Regressions are possible due to reorganized include statements. Obviously I didn't just copy mplayer.c's orgy of include statements, but recreated them for each file. It's easily possible that there are oversights and mistakes, which will show up on other platforms. There is one actual change: the public avutil.h include is removed from encode.h, and I tried to replace most FFMIN/FFMAX/av_clip uses. I consider using libavutil too much as dangerous, because the set of include files they recursively pull in is rather arbitrary and is different between FFmpeg and Libav.
2013-10-29 21:38:29 +00:00
player: add complex filter graph support See --lavfi-complex option. This is still quite rough. There's no support for dynamic configuration of any kind. There are probably corner cases where playback might freeze or burn 100% CPU (due to dataflow problems when interaction with libavfilter). Future possible plans might include: - freely switch tracks by providing some sort of default track graph label - automatically enabling audio visualization - automatically mix audio or stack video when multiple tracks are selected at once (similar to how multiple sub tracks can be selected)
2016-02-05 22:19:56 +00:00
switch (res) {
audio/video: merge decoder return values Will be helpful for the coming filter support. I planned on merging audio/video decoding, but this will have to wait a bit longer, so only remove the duplicate status codes.
2016-02-01 20:32:01 +00:00
case DATA_WAIT: return VD_WAIT;
player: add complex filter graph support See --lavfi-complex option. This is still quite rough. There's no support for dynamic configuration of any kind. There are probably corner cases where playback might freeze or burn 100% CPU (due to dataflow problems when interaction with libavfilter). Future possible plans might include: - freely switch tracks by providing some sort of default track graph label - automatically enabling audio visualization - automatically mix audio or stack video when multiple tracks are selected at once (similar to how multiple sub tracks can be selected)
2016-02-05 22:19:56 +00:00
case DATA_OK:
case DATA_AGAIN: return VD_PROGRESS;
audio/video: merge decoder return values Will be helpful for the coming filter support. I planned on merging audio/video decoding, but this will have to wait a bit longer, so only remove the duplicate status codes.
2016-02-01 20:32:01 +00:00
case DATA_EOF: return VD_EOF;
player: add complex filter graph support See --lavfi-complex option. This is still quite rough. There's no support for dynamic configuration of any kind. There are probably corner cases where playback might freeze or burn 100% CPU (due to dataflow problems when interaction with libavfilter). Future possible plans might include: - freely switch tracks by providing some sort of default track graph label - automatically enabling audio visualization - automatically mix audio or stack video when multiple tracks are selected at once (similar to how multiple sub tracks can be selected)
2016-02-05 22:19:56 +00:00
default: abort();
video: refactor: disentangle decoding/filtering some more This moves some code related to decoding from video.c to dec_video.c, and also removes some accesses to dec_video.c from the filtering code. dec_video.ch is starting to make sense, and simply returns video frames from a demuxer stream. The API exposed is also somewhat intended to be easily changeable to move decoding to a separate thread, if we ever want this (due to libavcodec already being threaded, I don't see much of a reason, but it might still be helpful).
2016-01-16 20:19:52 +00:00
}
}
video: change everything Change how the video decoding loop works. The structure should now be a bit easier to follow. The interactions on format changes are (probably) simpler. This also aligns the decoding loop with future planned changes, such as moving various things to separate threads.
2014-05-01 21:53:18 +00:00
video: separate calling decoder/filter Rename video_decode_and_filter to video_filter, and add a new video_decode_and_filter function. This function now calls the decoder. This is done so that we can check filters a second time after decoding, which avoids a useless playloop iteration. (This and the previous commits are really just microoptimizations, which simply reduce the number of times the playloop has to recheck everything.)
2014-09-18 17:22:39 +00:00
// Feed newly decoded frames to the filter, take care of format changes.
// If eof=true, drain the filter chain, and return VD_EOF if empty.
static int video_filter(struct MPContext *mpctx, bool eof)
video: move video filter chain initialization from decoder to player This should help fixing some issues (like not draining video frames correctly on reinit), as well as decoupling the decoder, filter chain, and VO code. I also wanted to make the hardware video decoding fallback work properly if software-only video filters are inserted. This currently has the issue that the fallback is too violent, and throws away a bunch of demuxer packets needed to restart software decoding properly. But keeping "backup" packets turned out as too hacky, so I'm not doing this, at least not yet.
2013-12-10 18:07:29 +00:00
{
video: refactor: disentangle decoding/filtering some more This moves some code related to decoding from video.c to dec_video.c, and also removes some accesses to dec_video.c from the filtering code. dec_video.ch is starting to make sense, and simply returns video frames from a demuxer stream. The API exposed is also somewhat intended to be easily changeable to move decoding to a separate thread, if we ever want this (due to libavcodec already being threaded, I don't see much of a reason, but it might still be helpful).
2016-01-16 20:19:52 +00:00
struct vo_chain *vo_c = mpctx->vo_chain;
struct vf_chain *vf = vo_c->vf;
Split mplayer.c mplayer.c was a bit too big. Split it into multiple files. I hope the way it's split makes sense. Maybe some things don't make too much sense, or go against intuition. These will fixed as soon as I notice them. Some files are a bit questionable (misc.c, osd.c, configfiles.c), and suggestions how to organize this better are welcome. Regressions are possible due to reorganized include statements. Obviously I didn't just copy mplayer.c's orgy of include statements, but recreated them for each file. It's easily possible that there are oversights and mistakes, which will show up on other platforms. There is one actual change: the public avutil.h include is removed from encode.h, and I tried to replace most FFMIN/FFMAX/av_clip uses. I consider using libavutil too much as dangerous, because the set of include files they recursively pull in is rather arbitrary and is different between FFmpeg and Libav.
2013-10-29 21:38:29 +00:00
video: change everything Change how the video decoding loop works. The structure should now be a bit easier to follow. The interactions on format changes are (probably) simpler. This also aligns the decoding loop with future planned changes, such as moving various things to separate threads.
2014-05-01 21:53:18 +00:00
if (vf->initialized < 0)
video: use symbolic constants instead of magic integers In my opinion this is not really necessary, since there's only a single user of update_video(), but others reading this code would probably hate me for using magic integer values instead of symbolic constants. This should be a purely cosmetic commit; any changes in behavior are bugs.
2014-07-18 13:11:21 +00:00
return VD_ERROR;
video: display last frame, drain frames on video reconfig Until now, the player didn't care to drain frames on video reconfig. Instead, the VO was reconfigured (i.e. resized) before the queued frames finished displaying. This can for example be observed by passing multiple images with different size as mf:// filename. Then the window would resize one frame before image with the new size is displayed. With --vo=vdpau, the effect is worse, because this VO queues more than 1 frame internally. Fix this by explicitly draining buffered frames before video reconfig. Raise the display time of the last frame. Otherwise, the last frame would be shown for a very short time only. This usually doesn't matter, but helps when playing image files. This is a byproduct of frame draining, because normally, video timing is based on the frames queued to the VO, and we can't do that with frames of different size or format. So we pretend that the frame before the change is the last frame in order to time it. This code is incorrect though: it tries to use the framerate, which often doesn't make sense. But it's good enough to test this code with mf://.
2013-12-10 18:33:11 +00:00
video: change everything Change how the video decoding loop works. The structure should now be a bit easier to follow. The interactions on format changes are (probably) simpler. This also aligns the decoding loop with future planned changes, such as moving various things to separate threads.
2014-05-01 21:53:18 +00:00
// There is already a filtered frame available.
video: better pipelining with vf_vapoursynth Most of this is explained in the code comments. This change should improve performance with vapoursynth, especially if concurrent requests are used. This should change nothing if vf_vapoursynth is not in the filter chain, since non-threaded filters obviously can not asynchronously finish filtering of frames.
2015-01-03 02:01:58 +00:00
// If vf_needs_input() returns > 0, the filter wants input anyway.
if (vf_output_frame(vf, eof) > 0 && vf_needs_input(vf) < 1)
video: use symbolic constants instead of magic integers In my opinion this is not really necessary, since there's only a single user of update_video(), but others reading this code would probably hate me for using magic integer values instead of symbolic constants. This should be a purely cosmetic commit; any changes in behavior are bugs.
2014-07-18 13:11:21 +00:00
return VD_PROGRESS;
video: change everything Change how the video decoding loop works. The structure should now be a bit easier to follow. The interactions on format changes are (probably) simpler. This also aligns the decoding loop with future planned changes, such as moving various things to separate threads.
2014-05-01 21:53:18 +00:00
// Decoder output is different from filter input?
bool need_vf_reconfig = !vf->input_params.imgfmt || vf->initialized < 1 ||
video: refactor: disentangle decoding/filtering some more This moves some code related to decoding from video.c to dec_video.c, and also removes some accesses to dec_video.c from the filtering code. dec_video.ch is starting to make sense, and simply returns video frames from a demuxer stream. The API exposed is also somewhat intended to be easily changeable to move decoding to a separate thread, if we ever want this (due to libavcodec already being threaded, I don't see much of a reason, but it might still be helpful).
2016-01-16 20:19:52 +00:00
!mp_image_params_equal(&vo_c->input_format, &vf->input_params);
video: change everything Change how the video decoding loop works. The structure should now be a bit easier to follow. The interactions on format changes are (probably) simpler. This also aligns the decoding loop with future planned changes, such as moving various things to separate threads.
2014-05-01 21:53:18 +00:00
// (If imgfmt==0, nothing was decoded yet, and the format is unknown.)
video: refactor: disentangle decoding/filtering some more This moves some code related to decoding from video.c to dec_video.c, and also removes some accesses to dec_video.c from the filtering code. dec_video.ch is starting to make sense, and simply returns video frames from a demuxer stream. The API exposed is also somewhat intended to be easily changeable to move decoding to a separate thread, if we ever want this (due to libavcodec already being threaded, I don't see much of a reason, but it might still be helpful).
2016-01-16 20:19:52 +00:00
if (need_vf_reconfig && vo_c->input_format.imgfmt) {
video: change everything Change how the video decoding loop works. The structure should now be a bit easier to follow. The interactions on format changes are (probably) simpler. This also aligns the decoding loop with future planned changes, such as moving various things to separate threads.
2014-05-01 21:53:18 +00:00
// Drain the filter chain.
if (vf_output_frame(vf, true) > 0)
video: use symbolic constants instead of magic integers In my opinion this is not really necessary, since there's only a single user of update_video(), but others reading this code would probably hate me for using magic integer values instead of symbolic constants. This should be a purely cosmetic commit; any changes in behavior are bugs.
2014-07-18 13:11:21 +00:00
return VD_PROGRESS;
video: change everything Change how the video decoding loop works. The structure should now be a bit easier to follow. The interactions on format changes are (probably) simpler. This also aligns the decoding loop with future planned changes, such as moving various things to separate threads.
2014-05-01 21:53:18 +00:00
// The filter chain is drained; execute the filter format change.
video: fix midstream video configuration changes Commit 771a8bf5 added code to avoid unnecessary vf_reconfig() calls for unrelated reasons, but forget to consider that it has to be called at least once if the input format changes. As a consequence it got "stuck" due to not being able to decode more frames.
2016-07-08 06:22:51 +00:00
vf->initialized = 0;
player: rewrite deinterlace filter auto-insertion Instead of using the "vf" command code (which changes filters at runtime on user input), use the general filter-insertion code. The latter was added later, and is more suitable for automatically inserted filters. The old code failed in particular when using watch-later saving, which stored the filter list in the resume config file. If a user changed the hardware decoding mode via command line, the stored filter chain was out of date and could cause failure due to not working with hardware or software decoding mode. Storing the deinterlace filter in the filter list was unavoidable, because it was part of the user state. (The new code only edits the actually instantiated filters.)
2016-07-05 19:10:26 +00:00
filter_reconfig(mpctx, mpctx->vo_chain);
video: refactor: slightly disentangle video filtering
2016-01-15 21:23:01 +00:00
mp_notify(mpctx, MPV_EVENT_VIDEO_RECONFIG, NULL);
// Most video filters don't work with hardware decoding, so this
// might be the reason why filter reconfig failed.
player: refactor: some more minor decoder/output decoupling These changes don't make too much sense without context, but are preparation for later. Then the audio_src/video_src fields will be actually be NULL under circumstances.
2016-01-29 21:46:28 +00:00
if (vf->initialized < 0 && vo_c->video_src &&
player: refactor: eliminate MPContext.d_video Eventually we want the VO be driven by a A->V filter, so a decoder doesn't even have to exist. Some features definitely require a decoder though (like reporting the decoder in use, hardware decoding, etc.), so for each thing which accessed d_video, it has to be redecided if and how it can access decoder state. At least the "framedrop" property slightly changes semantics: you can now always set this property, even if no video is active. Some untested changes in this commit, but our bio-based distributed test suite has to take care of this.
2016-01-17 17:07:50 +00:00
video_vd_control(vo_c->video_src, VDCTRL_FORCE_HWDEC_FALLBACK, NULL)
video: refactor: disentangle decoding/filtering some more This moves some code related to decoding from video.c to dec_video.c, and also removes some accesses to dec_video.c from the filtering code. dec_video.ch is starting to make sense, and simply returns video frames from a demuxer stream. The API exposed is also somewhat intended to be easily changeable to move decoding to a separate thread, if we ever want this (due to libavcodec already being threaded, I don't see much of a reason, but it might still be helpful).
2016-01-16 20:19:52 +00:00
== CONTROL_OK)
video: refactor: slightly disentangle video filtering
2016-01-15 21:23:01 +00:00
{
// Fallback active; decoder will return software format next
// time. Don't abort video decoding.
vf->initialized = 0;
video: refactor: disentangle decoding/filtering some more This moves some code related to decoding from video.c to dec_video.c, and also removes some accesses to dec_video.c from the filtering code. dec_video.ch is starting to make sense, and simply returns video frames from a demuxer stream. The API exposed is also somewhat intended to be easily changeable to move decoding to a separate thread, if we ever want this (due to libavcodec already being threaded, I don't see much of a reason, but it might still be helpful).
2016-01-16 20:19:52 +00:00
mp_image_unrefp(&vo_c->input_mpi);
vo_c->input_format = (struct mp_image_params){0};
video: refactor: slightly disentangle video filtering
2016-01-15 21:23:01 +00:00
MP_VERBOSE(mpctx, "hwdec falback due to filters.\n");
return VD_PROGRESS; // try again
}
if (vf->initialized < 1) {
MP_FATAL(mpctx, "Cannot initialize video filters.\n");
video: use symbolic constants instead of magic integers In my opinion this is not really necessary, since there's only a single user of update_video(), but others reading this code would probably hate me for using magic integer values instead of symbolic constants. This should be a purely cosmetic commit; any changes in behavior are bugs.
2014-07-18 13:11:21 +00:00
return VD_ERROR;
video: refactor: slightly disentangle video filtering
2016-01-15 21:23:01 +00:00
}
video: try harder to decode cover art picture only once For cover art, we pretend that the video stream is infinite, but also stop decoding once we have an image on the VO (this seems advantageous for the case when strange filters are inserted or the VO image gets lost). Since a while ago, the video chain started decoding 2 images though ("Non-monotonic video pts: 0.000000 <= 0.000000"), which is annoying and wasteful. Improve this by handling a certain corner case at initialization, which will decode a second image while the first one is still stuck in the filter chain. Also, just in case there are filters which buffer a lot, also force EOF filtering (which means we tell the filters to flush buffered frames). CC: @mpv-player/stable
2014-10-09 15:57:17 +00:00
return VD_RECONFIG;
Split mplayer.c mplayer.c was a bit too big. Split it into multiple files. I hope the way it's split makes sense. Maybe some things don't make too much sense, or go against intuition. These will fixed as soon as I notice them. Some files are a bit questionable (misc.c, osd.c, configfiles.c), and suggestions how to organize this better are welcome. Regressions are possible due to reorganized include statements. Obviously I didn't just copy mplayer.c's orgy of include statements, but recreated them for each file. It's easily possible that there are oversights and mistakes, which will show up on other platforms. There is one actual change: the public avutil.h include is removed from encode.h, and I tried to replace most FFMIN/FFMAX/av_clip uses. I consider using libavutil too much as dangerous, because the set of include files they recursively pull in is rather arbitrary and is different between FFmpeg and Libav.
2013-10-29 21:38:29 +00:00
}
video: move video filter chain initialization from decoder to player This should help fixing some issues (like not draining video frames correctly on reinit), as well as decoupling the decoder, filter chain, and VO code. I also wanted to make the hardware video decoding fallback work properly if software-only video filters are inserted. This currently has the issue that the fallback is too violent, and throws away a bunch of demuxer packets needed to restart software decoding properly. But keeping "backup" packets turned out as too hacky, so I'm not doing this, at least not yet.
2013-12-10 18:07:29 +00:00
video: change everything Change how the video decoding loop works. The structure should now be a bit easier to follow. The interactions on format changes are (probably) simpler. This also aligns the decoding loop with future planned changes, such as moving various things to separate threads.
2014-05-01 21:53:18 +00:00
// If something was decoded, and the filter chain is ready, filter it.
video: refactor: disentangle decoding/filtering some more This moves some code related to decoding from video.c to dec_video.c, and also removes some accesses to dec_video.c from the filtering code. dec_video.ch is starting to make sense, and simply returns video frames from a demuxer stream. The API exposed is also somewhat intended to be easily changeable to move decoding to a separate thread, if we ever want this (due to libavcodec already being threaded, I don't see much of a reason, but it might still be helpful).
2016-01-16 20:19:52 +00:00
if (!need_vf_reconfig && vo_c->input_mpi) {
vf_filter_frame(vf, vo_c->input_mpi);
vo_c->input_mpi = NULL;
video: use symbolic constants instead of magic integers In my opinion this is not really necessary, since there's only a single user of update_video(), but others reading this code would probably hate me for using magic integer values instead of symbolic constants. This should be a purely cosmetic commit; any changes in behavior are bugs.
2014-07-18 13:11:21 +00:00
return VD_PROGRESS;
video: move video filter chain initialization from decoder to player This should help fixing some issues (like not draining video frames correctly on reinit), as well as decoupling the decoder, filter chain, and VO code. I also wanted to make the hardware video decoding fallback work properly if software-only video filters are inserted. This currently has the issue that the fallback is too violent, and throws away a bunch of demuxer packets needed to restart software decoding properly. But keeping "backup" packets turned out as too hacky, so I'm not doing this, at least not yet.
2013-12-10 18:07:29 +00:00
}
Split mplayer.c mplayer.c was a bit too big. Split it into multiple files. I hope the way it's split makes sense. Maybe some things don't make too much sense, or go against intuition. These will fixed as soon as I notice them. Some files are a bit questionable (misc.c, osd.c, configfiles.c), and suggestions how to organize this better are welcome. Regressions are possible due to reorganized include statements. Obviously I didn't just copy mplayer.c's orgy of include statements, but recreated them for each file. It's easily possible that there are oversights and mistakes, which will show up on other platforms. There is one actual change: the public avutil.h include is removed from encode.h, and I tried to replace most FFMIN/FFMAX/av_clip uses. I consider using libavutil too much as dangerous, because the set of include files they recursively pull in is rather arbitrary and is different between FFmpeg and Libav.
2013-10-29 21:38:29 +00:00
video: separate calling decoder/filter Rename video_decode_and_filter to video_filter, and add a new video_decode_and_filter function. This function now calls the decoder. This is done so that we can check filters a second time after decoding, which avoids a useless playloop iteration. (This and the previous commits are really just microoptimizations, which simply reduce the number of times the playloop has to recheck everything.)
2014-09-18 17:22:39 +00:00
return eof ? VD_EOF : VD_PROGRESS;
}
// Make sure at least 1 filtered image is available, decode new video if needed.
// returns VD_* code
// A return value of VD_PROGRESS doesn't necessarily output a frame, but makes
// the promise that calling this function again will eventually do something.
static int video_decode_and_filter(struct MPContext *mpctx)
{
video: refactor: disentangle decoding/filtering some more This moves some code related to decoding from video.c to dec_video.c, and also removes some accesses to dec_video.c from the filtering code. dec_video.ch is starting to make sense, and simply returns video frames from a demuxer stream. The API exposed is also somewhat intended to be easily changeable to move decoding to a separate thread, if we ever want this (due to libavcodec already being threaded, I don't see much of a reason, but it might still be helpful).
2016-01-16 20:19:52 +00:00
struct vo_chain *vo_c = mpctx->vo_chain;
video: separate calling decoder/filter Rename video_decode_and_filter to video_filter, and add a new video_decode_and_filter function. This function now calls the decoder. This is done so that we can check filters a second time after decoding, which avoids a useless playloop iteration. (This and the previous commits are really just microoptimizations, which simply reduce the number of times the playloop has to recheck everything.)
2014-09-18 17:22:39 +00:00
int r = video_filter(mpctx, false);
if (r < 0)
return r;
video: refactor: disentangle decoding/filtering some more This moves some code related to decoding from video.c to dec_video.c, and also removes some accesses to dec_video.c from the filtering code. dec_video.ch is starting to make sense, and simply returns video frames from a demuxer stream. The API exposed is also somewhat intended to be easily changeable to move decoding to a separate thread, if we ever want this (due to libavcodec already being threaded, I don't see much of a reason, but it might still be helpful).
2016-01-16 20:19:52 +00:00
if (!vo_c->input_mpi) {
player: change aspects of cover art handling Cover art handling is a disgusting hack that causes a mess in all components. And this will stay this way. This is the Xth time I've changed cover art handling, and that will probably also continue. But change the code such that cover art is injected into the demux packet stream, instead of having an explicit special case it in the decoder glue code. (This is somewhat more similar to the cover art hack in libavformat.) To avoid that the over art picture is decoded again on each seek, we need some additional "caching" in player/video.c. Decoding it after each seek would work as well, but since cover art pictures can be pretty huge, it's probably ok to invest some lines of code into caching it. One weird thing is that the cover art packet will remain queued after seeks, but that is probably not an issue. In exchange, we can drop the dec_video.c code, which is pretty convenient for one of the following commits. This code duplicates a bunch of lower-level decode calls and does icky messing with this weird state stuff, so I'm glad it goes away.
2017-01-10 13:03:41 +00:00
if (vo_c->cached_coverart) {
// Don't ever decode it twice, not even after seek resets.
// (On seek resets, input_mpi is set to the cached image.)
r = VD_EOF;
} else {
// Decode a new image, or at least feed the decoder a packet.
r = decode_image(mpctx);
if (r == VD_WAIT)
return r;
}
video: change everything Change how the video decoding loop works. The structure should now be a bit easier to follow. The interactions on format changes are (probably) simpler. This also aligns the decoding loop with future planned changes, such as moving various things to separate threads.
2014-05-01 21:53:18 +00:00
}
player: change aspects of cover art handling Cover art handling is a disgusting hack that causes a mess in all components. And this will stay this way. This is the Xth time I've changed cover art handling, and that will probably also continue. But change the code such that cover art is injected into the demux packet stream, instead of having an explicit special case it in the decoder glue code. (This is somewhat more similar to the cover art hack in libavformat.) To avoid that the over art picture is decoded again on each seek, we need some additional "caching" in player/video.c. Decoding it after each seek would work as well, but since cover art pictures can be pretty huge, it's probably ok to invest some lines of code into caching it. One weird thing is that the cover art packet will remain queued after seeks, but that is probably not an issue. In exchange, we can drop the dec_video.c code, which is pretty convenient for one of the following commits. This code duplicates a bunch of lower-level decode calls and does icky messing with this weird state stuff, so I'm glad it goes away.
2017-01-10 13:03:41 +00:00
if (vo_c->input_mpi) {
player: refactor: some more minor decoder/output decoupling These changes don't make too much sense without context, but are preparation for later. Then the audio_src/video_src fields will be actually be NULL under circumstances.
2016-01-29 21:46:28 +00:00
vo_c->input_format = vo_c->input_mpi->params;
video: support filtering hardware frames via libavfilter Requires a bunch of hacks: - we access AVFilterLink.hw_frames_ctx. This is not a public API in FFmpeg and Libav. Newer FFmpeg provides an accessor (av_buffersink_get_hw_frames_ctx), but it's not available in Libav or the current FFmpeg release or Libav. We need this value after filter graph creation, so We have no choice but to access this. One alternative is making filter creation and format negotiation fully lazy (i.e. delay it and do it as filters are output), but this would be a huge change. So for now, we knowingly violate FFmpeg's and Libav's ABI and API constraints because they don't provide anything better. On newer FFmpeg, we use the (quite ugly) accessor, though. - mp_image_params doesn't (and can't) have a field for the frames context AVBufferRef. So we pass it via vf_set_proto_frame(), and even more hacks. - if a filter needs a hw context, but we haven't created one yet (because normally we create them lazily), it will fail at init. - we allow any hw format now, although this could go horrible wrong. Why all this effort? We could move hw deinterlacing filters etc. to FFmpeg, which is a very worthy goal.
2017-01-16 14:52:13 +00:00
vf_set_proto_frame(vo_c->vf, vo_c->input_mpi);
Split mplayer.c mplayer.c was a bit too big. Split it into multiple files. I hope the way it's split makes sense. Maybe some things don't make too much sense, or go against intuition. These will fixed as soon as I notice them. Some files are a bit questionable (misc.c, osd.c, configfiles.c), and suggestions how to organize this better are welcome. Regressions are possible due to reorganized include statements. Obviously I didn't just copy mplayer.c's orgy of include statements, but recreated them for each file. It's easily possible that there are oversights and mistakes, which will show up on other platforms. There is one actual change: the public avutil.h include is removed from encode.h, and I tried to replace most FFMIN/FFMAX/av_clip uses. I consider using libavutil too much as dangerous, because the set of include files they recursively pull in is rather arbitrary and is different between FFmpeg and Libav.
2013-10-29 21:38:29 +00:00
player: change aspects of cover art handling Cover art handling is a disgusting hack that causes a mess in all components. And this will stay this way. This is the Xth time I've changed cover art handling, and that will probably also continue. But change the code such that cover art is injected into the demux packet stream, instead of having an explicit special case it in the decoder glue code. (This is somewhat more similar to the cover art hack in libavformat.) To avoid that the over art picture is decoded again on each seek, we need some additional "caching" in player/video.c. Decoding it after each seek would work as well, but since cover art pictures can be pretty huge, it's probably ok to invest some lines of code into caching it. One weird thing is that the cover art packet will remain queued after seeks, but that is probably not an issue. In exchange, we can drop the dec_video.c code, which is pretty convenient for one of the following commits. This code duplicates a bunch of lower-level decode calls and does icky messing with this weird state stuff, so I'm glad it goes away.
2017-01-10 13:03:41 +00:00
if (vo_c->is_coverart && !vo_c->cached_coverart)
vo_c->cached_coverart = mp_image_new_ref(vo_c->input_mpi);
}
video: refactor: disentangle decoding/filtering some more This moves some code related to decoding from video.c to dec_video.c, and also removes some accesses to dec_video.c from the filtering code. dec_video.ch is starting to make sense, and simply returns video frames from a demuxer stream. The API exposed is also somewhat intended to be easily changeable to move decoding to a separate thread, if we ever want this (due to libavcodec already being threaded, I don't see much of a reason, but it might still be helpful).
2016-01-16 20:19:52 +00:00
bool eof = !vo_c->input_mpi && (r == VD_EOF || r < 0);
video: try harder to decode cover art picture only once For cover art, we pretend that the video stream is infinite, but also stop decoding once we have an image on the VO (this seems advantageous for the case when strange filters are inserted or the VO image gets lost). Since a while ago, the video chain started decoding 2 images though ("Non-monotonic video pts: 0.000000 <= 0.000000"), which is annoying and wasteful. Improve this by handling a certain corner case at initialization, which will decode a second image while the first one is still stuck in the filter chain. Also, just in case there are filters which buffer a lot, also force EOF filtering (which means we tell the filters to flush buffered frames). CC: @mpv-player/stable
2014-10-09 15:57:17 +00:00
r = video_filter(mpctx, eof);
if (r == VD_RECONFIG) // retry feeding decoded image
r = video_filter(mpctx, eof);
return r;
video: display last frame, drain frames on video reconfig Until now, the player didn't care to drain frames on video reconfig. Instead, the VO was reconfigured (i.e. resized) before the queued frames finished displaying. This can for example be observed by passing multiple images with different size as mf:// filename. Then the window would resize one frame before image with the new size is displayed. With --vo=vdpau, the effect is worse, because this VO queues more than 1 frame internally. Fix this by explicitly draining buffered frames before video reconfig. Raise the display time of the last frame. Otherwise, the last frame would be shown for a very short time only. This usually doesn't matter, but helps when playing image files. This is a byproduct of frame draining, because normally, video timing is based on the frames queued to the VO, and we can't do that with frames of different size or format. So we pretend that the frame before the change is the last frame in order to time it. This code is incorrect though: it tries to use the framerate, which often doesn't make sense. But it's good enough to test this code with mf://.
2013-12-10 18:33:11 +00:00
}
video: better pipelining with vf_vapoursynth Most of this is explained in the code comments. This change should improve performance with vapoursynth, especially if concurrent requests are used. This should change nothing if vf_vapoursynth is not in the filter chain, since non-threaded filters obviously can not asynchronously finish filtering of frames.
2015-01-03 02:01:58 +00:00
static int video_feed_async_filter(struct MPContext *mpctx)
{
video: decouple filtering/decoding slightly more Lots of noise to remove the vfilter/vo fields from dec_video. From now on, video filtering and output will still be done together, summarized under struct vo_chain. There is the question where exactly the vf_chain should go in such a decoupled architecture. The end goal is being able to place a "complex" filter between video decoders and output (which will culminate in natural integration of A->V filters for natural integration of libavfilter audio visualizations). The vf_chain is still useful for "final" processing, such as format conversions and deinterlacing. Also, there's only 1 VO and 1 --vf option. So having 1 vf_chain for a VO seems ideal, since otherwise there would be no natural way to handle all these existing options and mechanisms. There is still some work required to truly decouple decoding.
2016-01-13 23:18:48 +00:00
struct vf_chain *vf = mpctx->vo_chain->vf;
video: better pipelining with vf_vapoursynth Most of this is explained in the code comments. This change should improve performance with vapoursynth, especially if concurrent requests are used. This should change nothing if vf_vapoursynth is not in the filter chain, since non-threaded filters obviously can not asynchronously finish filtering of frames.
2015-01-03 02:01:58 +00:00
if (vf->initialized < 0)
return VD_ERROR;
if (vf_needs_input(vf) < 1)
return 0;
player: litter code with explicit wakeup calls This does 3 kinds of changes: - change sleeptime=x to mp_set_timeout() - change sleeptime=0 to mp_wakeup_core() calls (to be more explicit) - change commands etc. to call mp_wakeup_core() if they do changes that require the playloop to be rerun This is preparation for the following changes. The goal is to process client API requests without having to rerun the playloop every time. As of this commit, the changes should not change behavior. In particular, the playloop is still implicitly woken up on every command.
2016-09-16 12:24:15 +00:00
mp_wakeup_core(mpctx); // retry until done
video: better pipelining with vf_vapoursynth Most of this is explained in the code comments. This change should improve performance with vapoursynth, especially if concurrent requests are used. This should change nothing if vf_vapoursynth is not in the filter chain, since non-threaded filters obviously can not asynchronously finish filtering of frames.
2015-01-03 02:01:58 +00:00
return video_decode_and_filter(mpctx);
}
video: move some code around No functional changes. init_vo() is now needed a bit further down, and moving it keeps definition and use close. adjust_sync() will be used by a function further up in one of the following commits.
2014-08-22 12:19:01 +00:00
/* Modify video timing to match the audio timeline. There are two main
* reasons this is needed. First, video and audio can start from different
* positions at beginning of file or after a seek (MPlayer starts both
* immediately even if they have different pts). Second, the file can have
* audio timestamps that are inconsistent with the duration of the audio
* packets, for example two consecutive timestamp values differing by
* one second but only a packet with enough samples for half a second
* of playback between them.
*/
static void adjust_sync(struct MPContext *mpctx, double v_pts, double frame_time)
Split mplayer.c mplayer.c was a bit too big. Split it into multiple files. I hope the way it's split makes sense. Maybe some things don't make too much sense, or go against intuition. These will fixed as soon as I notice them. Some files are a bit questionable (misc.c, osd.c, configfiles.c), and suggestions how to organize this better are welcome. Regressions are possible due to reorganized include statements. Obviously I didn't just copy mplayer.c's orgy of include statements, but recreated them for each file. It's easily possible that there are oversights and mistakes, which will show up on other platforms. There is one actual change: the public avutil.h include is removed from encode.h, and I tried to replace most FFMIN/FFMAX/av_clip uses. I consider using libavutil too much as dangerous, because the set of include files they recursively pull in is rather arbitrary and is different between FFmpeg and Libav.
2013-10-29 21:38:29 +00:00
{
struct MPOpts *opts = mpctx->opts;
video: change everything Change how the video decoding loop works. The structure should now be a bit easier to follow. The interactions on format changes are (probably) simpler. This also aligns the decoding loop with future planned changes, such as moving various things to separate threads.
2014-05-01 21:53:18 +00:00
video: move some code around No functional changes. init_vo() is now needed a bit further down, and moving it keeps definition and use close. adjust_sync() will be used by a function further up in one of the following commits.
2014-08-22 12:19:01 +00:00
if (mpctx->audio_status != STATUS_PLAYING)
return;
client API, video: signal reconfig at the right time Filter reconfig can now happen a few frames before VO reconfig.
2014-05-01 23:28:52 +00:00
player: remove redundant variable mpctx->audio_delay always has the same value as opts->audio_delay. (This was not the case a long time ago, when the audio-delay property didn't actually write to opts->audio_delay. I think.)
2015-01-29 11:10:07 +00:00
double a_pts = written_audio_pts(mpctx) + opts->audio_delay - mpctx->delay;
video: move some code around No functional changes. init_vo() is now needed a bit further down, and moving it keeps definition and use close. adjust_sync() will be used by a function further up in one of the following commits.
2014-08-22 12:19:01 +00:00
double av_delay = a_pts - v_pts;
double change = av_delay * 0.1;
player: speed up audio/video re-sync when there is a huge delay when there is a huge delay between audio/video sync, it can take a really long time to converge back. this speeds up the resync time by increasing the max_change allowed per iteration. Signed-off-by: wm4 <wm4@nowhere>
2016-10-20 18:20:23 +00:00
double factor = fabs(av_delay) < 0.3 ? 0.1 : 0.4;
video: move some code around No functional changes. init_vo() is now needed a bit further down, and moving it keeps definition and use close. adjust_sync() will be used by a function further up in one of the following commits.
2014-08-22 12:19:01 +00:00
double max_change = opts->default_max_pts_correction >= 0 ?
player: speed up audio/video re-sync when there is a huge delay when there is a huge delay between audio/video sync, it can take a really long time to converge back. this speeds up the resync time by increasing the max_change allowed per iteration. Signed-off-by: wm4 <wm4@nowhere>
2016-10-20 18:20:23 +00:00
opts->default_max_pts_correction : frame_time * factor;
video: move some code around No functional changes. init_vo() is now needed a bit further down, and moving it keeps definition and use close. adjust_sync() will be used by a function further up in one of the following commits.
2014-08-22 12:19:01 +00:00
if (change < -max_change)
change = -max_change;
else if (change > max_change)
change = max_change;
mpctx->delay += change;
mpctx->total_avsync_change += change;
player: disable total-avsync-change update in display-sync mode The total-avsync-change property made no sense in display-sync mode (in addition to making not all that much sense in general).
2015-10-27 19:56:58 +00:00
if (mpctx->display_sync_active)
mpctx->total_avsync_change = 0;
Split mplayer.c mplayer.c was a bit too big. Split it into multiple files. I hope the way it's split makes sense. Maybe some things don't make too much sense, or go against intuition. These will fixed as soon as I notice them. Some files are a bit questionable (misc.c, osd.c, configfiles.c), and suggestions how to organize this better are welcome. Regressions are possible due to reorganized include statements. Obviously I didn't just copy mplayer.c's orgy of include statements, but recreated them for each file. It's easily possible that there are oversights and mistakes, which will show up on other platforms. There is one actual change: the public avutil.h include is removed from encode.h, and I tried to replace most FFMIN/FFMAX/av_clip uses. I consider using libavutil too much as dangerous, because the set of include files they recursively pull in is rather arbitrary and is different between FFmpeg and Libav.
2013-10-29 21:38:29 +00:00
}
video: pass future frames to VO Now the VO can request a number of future frames with the last parameter of vo_set_queue_params(). This will be helpful to fix the interpolation code. Note that the first frame (after playback start or seeking) will usually not have any future frames (to make seeking fast). Near the end of the file, the number of future frames will become lower as well.
2015-07-01 17:22:40 +00:00
// Make the frame at position 0 "known" to the playback logic. This must happen
// only once for each frame, so this function has to be called carefully.
// Generally, if position 0 gets a new frame, this must be called.
static void handle_new_frame(struct MPContext *mpctx)
video: refactor aspects of queue and EOF handling Instead of touching the 2-entry queue in mpctx->next_frame directly, move some of it to functions.
2015-03-25 12:33:34 +00:00
{
video: pass future frames to VO Now the VO can request a number of future frames with the last parameter of vo_set_queue_params(). This will be helpful to fix the interpolation code. Note that the first frame (after playback start or seeking) will usually not have any future frames (to make seeking fast). Near the end of the file, the number of future frames will become lower as well.
2015-07-01 17:22:40 +00:00
assert(mpctx->num_next_frames >= 1);
video: refactor aspects of queue and EOF handling Instead of touching the 2-entry queue in mpctx->next_frame directly, move some of it to functions.
2015-03-25 12:33:34 +00:00
double frame_time = 0;
video: pass future frames to VO Now the VO can request a number of future frames with the last parameter of vo_set_queue_params(). This will be helpful to fix the interpolation code. Note that the first frame (after playback start or seeking) will usually not have any future frames (to make seeking fast). Near the end of the file, the number of future frames will become lower as well.
2015-07-01 17:22:40 +00:00
double pts = mpctx->next_frames[0]->pts;
video: refactor aspects of queue and EOF handling Instead of touching the 2-entry queue in mpctx->next_frame directly, move some of it to functions.
2015-03-25 12:33:34 +00:00
if (mpctx->video_pts != MP_NOPTS_VALUE) {
frame_time = pts - mpctx->video_pts;
player: assume video forwards timestamps jumps only with some formats Another crappy fix for timestamp reset issues. This time, we try to fix files which have very weird but legitimate frame durations, such as cdgraphics. It can have many short frames, but once in a while there are potentially very long frames. Fixes #3027.
2016-04-24 16:24:46 +00:00
double tolerance = mpctx->demuxer->ts_resets_possible ? 5 : 1e4;
video: better heuristic for timestamp resets Reduce the default tolerance for timestamp jumps from 60 to 15 seconds. For .ts files, where ts_resets_possible coming from AVFMT_TS_DISCONT is set, apply a more sophisticated heuristic. It's clear that such a file wouldn't have a framerate below, say, 23hz. If the demuxer reports a lower fps, we allow longer PTS jumps. This should replace long pauses on discontinuities with .ts files with at most a short stutter. Of course, all kinds of things could go wrong anyway if the source is VFR, or FFmpeg's frame rate detection fails in some other way. I haven't found such a file yet, though.
2015-05-20 12:24:55 +00:00
if (frame_time <= 0 || frame_time >= tolerance) {
// Assume a discontinuity.
video: refactor aspects of queue and EOF handling Instead of touching the 2-entry queue in mpctx->next_frame directly, move some of it to functions.
2015-03-25 12:33:34 +00:00
MP_WARN(mpctx, "Invalid video timestamp: %f -> %f\n",
mpctx->video_pts, pts);
player: resync audio only on larger timestamp discontinuities Helps with files that have occasional broken timestamps. For larger discontinuities, e.g. caused by actual timestamp resets, we still want to realign audio. (I guess in general, this should be removed and replaced by a more general resync-on-desync logic, but not now.)
2015-12-04 15:55:55 +00:00
frame_time = 0;
video: refactor aspects of queue and EOF handling Instead of touching the 2-entry queue in mpctx->next_frame directly, move some of it to functions.
2015-03-25 12:33:34 +00:00
}
}
mpctx->delay -= frame_time;
if (mpctx->video_status >= STATUS_PLAYING) {
player: separate controls for user and video controlled speed For video sync, we want separate playback speed controls for user- requested speed and the "correction" speed for video timing. Further, we use this separation to make sure only a resampler is inserted if playback speed is only changed for video sync correction. As of this commit, this is basically inactive code. It's just preparation for the video sync code (the following commit).
2015-08-10 16:40:16 +00:00
mpctx->time_frame += frame_time / mpctx->video_speed;
video: refactor aspects of queue and EOF handling Instead of touching the 2-entry queue in mpctx->next_frame directly, move some of it to functions.
2015-03-25 12:33:34 +00:00
adjust_sync(mpctx, pts, frame_time);
}
player: refactor: some more minor decoder/output decoupling These changes don't make too much sense without context, but are preparation for later. Then the audio_src/video_src fields will be actually be NULL under circumstances.
2016-01-29 21:46:28 +00:00
struct dec_video *d_video = mpctx->vo_chain->video_src;
if (d_video)
mpctx->dropped_frames_start = d_video->dropped_frames;
video: refactor aspects of queue and EOF handling Instead of touching the 2-entry queue in mpctx->next_frame directly, move some of it to functions.
2015-03-25 12:33:34 +00:00
MP_TRACE(mpctx, "frametime=%5.3f\n", frame_time);
}
vo: change internal API for drawing frames draw_image_timed is renamed to draw_frame. struct frame_timing is renamed to vo_frame. flip_page_timed is merged into draw_frame (the additional parameters are part of struct vo_frame). draw_frame also deprecates VOCTRL_REDRAW_FRAME, and replaces it with a method that works for both VOs which can cache the current frame, and VOs which need to redraw it anyway. This is preparation to making the interpolation and (work in progress) display sync code saner. Lots of other refactoring, and also some simplifications.
2015-07-01 17:24:28 +00:00
// Remove the first frame in mpctx->next_frames
static void shift_frames(struct MPContext *mpctx)
{
if (mpctx->num_next_frames < 1)
return;
talloc_free(mpctx->next_frames[0]);
for (int n = 0; n < mpctx->num_next_frames - 1; n++)
mpctx->next_frames[n] = mpctx->next_frames[n + 1];
mpctx->num_next_frames -= 1;
}
video: pass future frames to VO Now the VO can request a number of future frames with the last parameter of vo_set_queue_params(). This will be helpful to fix the interpolation code. Note that the first frame (after playback start or seeking) will usually not have any future frames (to make seeking fast). Near the end of the file, the number of future frames will become lower as well.
2015-07-01 17:22:40 +00:00
static int get_req_frames(struct MPContext *mpctx, bool eof)
{
// On EOF, drain all frames.
player: always require a future frame with display-sync enabled We need a frame duration even on start, because the number of vsyncs the frame is shown is predetermined. (vo_opengl actually makes use of this property in certain cases.)
2015-11-13 21:42:42 +00:00
if (eof)
video: pass future frames to VO Now the VO can request a number of future frames with the last parameter of vo_set_queue_params(). This will be helpful to fix the interpolation code. Note that the first frame (after playback start or seeking) will usually not have any future frames (to make seeking fast). Near the end of the file, the number of future frames will become lower as well.
2015-07-01 17:22:40 +00:00
return 1;
player: always require a future frame with display-sync enabled We need a frame duration even on start, because the number of vsyncs the frame is shown is predetermined. (vo_opengl actually makes use of this property in certain cases.)
2015-11-13 21:42:42 +00:00
// On the first frame, output a new frame as quickly as possible.
// But display-sync likes to have a correct frame duration always.
if (mpctx->video_pts == MP_NOPTS_VALUE)
return mpctx->opts->video_sync == VS_DEFAULT ? 1 : 2;
vo: minor simplification for queue size handling Instead of calling it "future frames" and adding or subtracting 1 from it, always call it "requested frames". This simplifies it a bit. MPContext.next_frames had 2 added to it; this was mainly to ensure a minimum size of 2. Drop it and assume VO_MAX_REQ_FRAMES is at least 2; together with the other changes, this can be the exact size of the array.
2015-07-20 19:12:46 +00:00
int req = vo_get_num_req_frames(mpctx->video_out);
player: simplify backstepping Basically reimplement it. The old implementation was quite stupid, and was probably done this way because video filtering and output used to be way less decoupled. Now we can reimplement it in a very simple way: when backstepping, seek to current time, but keep the last frame that was supposed to be discarded when reaching the target time. When the seek finishes, prepend the saved frame to the video frame queue. A disadvantage is that the new implementation fails to skip over timeline boundaries (ordered chapters etc.), but this never worked properly anyway. It's possible that this will be fixed some time in the future.
2016-01-12 22:49:00 +00:00
return MPCLAMP(req, 2, MP_ARRAY_SIZE(mpctx->next_frames) - 1);
video: pass future frames to VO Now the VO can request a number of future frames with the last parameter of vo_set_queue_params(). This will be helpful to fix the interpolation code. Note that the first frame (after playback start or seeking) will usually not have any future frames (to make seeking fast). Near the end of the file, the number of future frames will become lower as well.
2015-07-01 17:22:40 +00:00
}
video: refactor aspects of queue and EOF handling Instead of touching the 2-entry queue in mpctx->next_frame directly, move some of it to functions.
2015-03-25 12:33:34 +00:00
// Whether it's fine to call add_new_frame() now.
static bool needs_new_frame(struct MPContext *mpctx)
{
video: pass future frames to VO Now the VO can request a number of future frames with the last parameter of vo_set_queue_params(). This will be helpful to fix the interpolation code. Note that the first frame (after playback start or seeking) will usually not have any future frames (to make seeking fast). Near the end of the file, the number of future frames will become lower as well.
2015-07-01 17:22:40 +00:00
return mpctx->num_next_frames < get_req_frames(mpctx, false);
video: refactor aspects of queue and EOF handling Instead of touching the 2-entry queue in mpctx->next_frame directly, move some of it to functions.
2015-03-25 12:33:34 +00:00
}
video: pass future frames to VO Now the VO can request a number of future frames with the last parameter of vo_set_queue_params(). This will be helpful to fix the interpolation code. Note that the first frame (after playback start or seeking) will usually not have any future frames (to make seeking fast). Near the end of the file, the number of future frames will become lower as well.
2015-07-01 17:22:40 +00:00
// Queue a frame to mpctx->next_frames[]. Call only if needs_new_frame() signals ok.
video: refactor aspects of queue and EOF handling Instead of touching the 2-entry queue in mpctx->next_frame directly, move some of it to functions.
2015-03-25 12:33:34 +00:00
static void add_new_frame(struct MPContext *mpctx, struct mp_image *frame)
{
player: simplify backstepping Basically reimplement it. The old implementation was quite stupid, and was probably done this way because video filtering and output used to be way less decoupled. Now we can reimplement it in a very simple way: when backstepping, seek to current time, but keep the last frame that was supposed to be discarded when reaching the target time. When the seek finishes, prepend the saved frame to the video frame queue. A disadvantage is that the new implementation fails to skip over timeline boundaries (ordered chapters etc.), but this never worked properly anyway. It's possible that this will be fixed some time in the future.
2016-01-12 22:49:00 +00:00
assert(mpctx->num_next_frames < MP_ARRAY_SIZE(mpctx->next_frames));
video: refactor aspects of queue and EOF handling Instead of touching the 2-entry queue in mpctx->next_frame directly, move some of it to functions.
2015-03-25 12:33:34 +00:00
assert(frame);
video: pass future frames to VO Now the VO can request a number of future frames with the last parameter of vo_set_queue_params(). This will be helpful to fix the interpolation code. Note that the first frame (after playback start or seeking) will usually not have any future frames (to make seeking fast). Near the end of the file, the number of future frames will become lower as well.
2015-07-01 17:22:40 +00:00
mpctx->next_frames[mpctx->num_next_frames++] = frame;
if (mpctx->num_next_frames == 1)
handle_new_frame(mpctx);
video: refactor aspects of queue and EOF handling Instead of touching the 2-entry queue in mpctx->next_frame directly, move some of it to functions.
2015-03-25 12:33:34 +00:00
}
video: check whether there are enough frames after filtering Move the check to a function. Run the check a second time after decoding/filtering. This second check is strictly speaking redundant (which is why it wasn't done until now), but it avoids a useless playloop iteration.
2014-09-18 17:20:22 +00:00
// Enough video filtered already to push one frame to the VO?
video: refactor aspects of queue and EOF handling Instead of touching the 2-entry queue in mpctx->next_frame directly, move some of it to functions.
2015-03-25 12:33:34 +00:00
// Set eof to true if no new frames are to be expected.
static bool have_new_frame(struct MPContext *mpctx, bool eof)
video: check whether there are enough frames after filtering Move the check to a function. Run the check a second time after decoding/filtering. This second check is strictly speaking redundant (which is why it wasn't done until now), but it avoids a useless playloop iteration.
2014-09-18 17:20:22 +00:00
{
video: pass future frames to VO Now the VO can request a number of future frames with the last parameter of vo_set_queue_params(). This will be helpful to fix the interpolation code. Note that the first frame (after playback start or seeking) will usually not have any future frames (to make seeking fast). Near the end of the file, the number of future frames will become lower as well.
2015-07-01 17:22:40 +00:00
return mpctx->num_next_frames >= get_req_frames(mpctx, eof);
video: check whether there are enough frames after filtering Move the check to a function. Run the check a second time after decoding/filtering. This second check is strictly speaking redundant (which is why it wasn't done until now), but it avoids a useless playloop iteration.
2014-09-18 17:20:22 +00:00
}
video: pass future frames to VO Now the VO can request a number of future frames with the last parameter of vo_set_queue_params(). This will be helpful to fix the interpolation code. Note that the first frame (after playback start or seeking) will usually not have any future frames (to make seeking fast). Near the end of the file, the number of future frames will become lower as well.
2015-07-01 17:22:40 +00:00
// Fill mpctx->next_frames[] with a newly filtered or decoded image.
video: use symbolic constants instead of magic integers In my opinion this is not really necessary, since there's only a single user of update_video(), but others reading this code would probably hate me for using magic integer values instead of symbolic constants. This should be a purely cosmetic commit; any changes in behavior are bugs.
2014-07-18 13:11:21 +00:00
// returns VD_* code
player: minor simplification No need to pass endpts down in such a dumb way. Also remove an outdated comment somewhere.
2016-02-27 19:02:51 +00:00
static int video_output_image(struct MPContext *mpctx)
Split mplayer.c mplayer.c was a bit too big. Split it into multiple files. I hope the way it's split makes sense. Maybe some things don't make too much sense, or go against intuition. These will fixed as soon as I notice them. Some files are a bit questionable (misc.c, osd.c, configfiles.c), and suggestions how to organize this better are welcome. Regressions are possible due to reorganized include statements. Obviously I didn't just copy mplayer.c's orgy of include statements, but recreated them for each file. It's easily possible that there are oversights and mistakes, which will show up on other platforms. There is one actual change: the public avutil.h include is removed from encode.h, and I tried to replace most FFMIN/FFMAX/av_clip uses. I consider using libavutil too much as dangerous, because the set of include files they recursively pull in is rather arbitrary and is different between FFmpeg and Libav.
2013-10-29 21:38:29 +00:00
{
player: use different variable to indicate coverart Slightly better.
2016-02-01 21:14:32 +00:00
struct vo_chain *vo_c = mpctx->vo_chain;
player: when seeking past EOF with --keep-open, seek to last frame It feels strange that seeking past EOF with --keep-open actually leaves the player at a random position. You can't even unpause, because the demuxer is in the EOF state, and what you see on screen is just what was around before the seek. Improve this by attempting to seek to the last video frame if EOF happens. We explicitly don't do this if EOF was reached normally to increase robustness (if the VO got a frame since the last seek, it obviously means we had normal playback before EOF). If an error happens when trying to find the last frame (such as not actually finding a last frame because e.g. the demuxer misbehaves), this will probably turn your CPU into a heater. There is no logic to prevent reinitiating the last-frame search if the last-frame search reached EOF. (Pausing usually prevents that EOF is reached again after a successful last-frame search.) Fixes #819.
2014-12-07 01:47:09 +00:00
bool hrseek = mpctx->hrseek_active && mpctx->video_status == STATUS_SYNCING;
player: use different variable to indicate coverart Slightly better.
2016-02-01 21:14:32 +00:00
if (vo_c->is_coverart) {
video: refactor queue handling This simplifies the code, and fixes an odd bug: the second-last frame was displayed for a very short duration if framedrop was enabled. The reason was that basically the time difference between second-last and last frame were skipped, because at this point EOF was already signaled. Also see commit b0959488 for a similar issue in the same code. This removes the messiness of the next_frame 2-frame queue, and strictly runs the "new frame" code when a frame is moved to the first position of the queue, instead of somehow messing with return codes. This also merges update_video() into video_output_image().
2014-08-22 12:20:30 +00:00
if (vo_has_frame(mpctx->video_out))
video: use symbolic constants instead of magic integers In my opinion this is not really necessary, since there's only a single user of update_video(), but others reading this code would probably hate me for using magic integer values instead of symbolic constants. This should be a purely cosmetic commit; any changes in behavior are bugs.
2014-07-18 13:11:21 +00:00
return VD_EOF;
video: refactor: disentangle decoding/filtering some more This moves some code related to decoding from video.c to dec_video.c, and also removes some accesses to dec_video.c from the filtering code. dec_video.ch is starting to make sense, and simply returns video frames from a demuxer stream. The API exposed is also somewhat intended to be easily changeable to move decoding to a separate thread, if we ever want this (due to libavcodec already being threaded, I don't see much of a reason, but it might still be helpful).
2016-01-16 20:19:52 +00:00
hrseek = false;
video: reduce non-sense messages when playing coverart Don't print PTS warnings by skipping the normal video path.
2014-08-13 19:55:46 +00:00
}
video: refactor aspects of queue and EOF handling Instead of touching the 2-entry queue in mpctx->next_frame directly, move some of it to functions.
2015-03-25 12:33:34 +00:00
if (have_new_frame(mpctx, false))
video: filter new frames at a better time (2) We generally want 2 things: 1. minimal wakeups for decoding each frame 2. minimal number of frames decoded on continuous seeking Commit 35810cb8 changed this a bit, and fixed 1. But it broke 2., and now it decodes 2 frames instead of 1 when you keep seeking (arrow key held down or such). This made seeking appear slower. Fix this by making the logic more explicit. In particular, call the filters only if we actually try to get a new frame. When playing with --no-audio and all other distractions disabled (like OSC), it still wakes up 2 times per frame - but the second time is merely because the VO didn't accept the new frame yet.
2014-09-22 16:06:59 +00:00
return VD_NEW_FRAME;
video: filter new frames at a better time Move this code below the code that "shifts" the newly filtered frame. This allows us to skip a useless playloop iteration later, because obviously we need to filter a new frame after the previous frame has been "shifted", and not before that.
2014-09-18 17:17:38 +00:00
// Get a new frame if we need one.
video: filter new frames at a better time (2) We generally want 2 things: 1. minimal wakeups for decoding each frame 2. minimal number of frames decoded on continuous seeking Commit 35810cb8 changed this a bit, and fixed 1. But it broke 2., and now it decodes 2 frames instead of 1 when you keep seeking (arrow key held down or such). This made seeking appear slower. Fix this by making the logic more explicit. In particular, call the filters only if we actually try to get a new frame. When playing with --no-audio and all other distractions disabled (like OSC), it still wakes up 2 times per frame - but the second time is merely because the VO didn't accept the new frame yet.
2014-09-22 16:06:59 +00:00
int r = VD_PROGRESS;
video: refactor aspects of queue and EOF handling Instead of touching the 2-entry queue in mpctx->next_frame directly, move some of it to functions.
2015-03-25 12:33:34 +00:00
if (needs_new_frame(mpctx)) {
video: filter new frames at a better time (2) We generally want 2 things: 1. minimal wakeups for decoding each frame 2. minimal number of frames decoded on continuous seeking Commit 35810cb8 changed this a bit, and fixed 1. But it broke 2., and now it decodes 2 frames instead of 1 when you keep seeking (arrow key held down or such). This made seeking appear slower. Fix this by making the logic more explicit. In particular, call the filters only if we actually try to get a new frame. When playing with --no-audio and all other distractions disabled (like OSC), it still wakes up 2 times per frame - but the second time is merely because the VO didn't accept the new frame yet.
2014-09-22 16:06:59 +00:00
// Filter a new frame.
r = video_decode_and_filter(mpctx);
if (r < 0)
return r; // error
player: use different variable to indicate coverart Slightly better.
2016-02-01 21:14:32 +00:00
struct mp_image *img = vf_read_output_frame(vo_c->vf);
video: filter new frames at a better time Move this code below the code that "shifts" the newly filtered frame. This allows us to skip a useless playloop iteration later, because obviously we need to filter a new frame after the previous frame has been "shifted", and not before that.
2014-09-18 17:17:38 +00:00
if (img) {
player: minor simplification No need to pass endpts down in such a dumb way. Also remove an outdated comment somewhere.
2016-02-27 19:02:51 +00:00
double endpts = get_play_end_pts(mpctx);
video: don't discard video frames after endpts Instead of letting it keep decoding by trying to find a new frame, "plug" the frame queue by not removing it. (Or actually, by putting it back instead of discarding it.) Matters for seamless looping (following commits), and possibly some other corner cases. The added function vf_unread_output_frame() is a bit of a sin, but still reasonable, since its implementation is trivial.
2016-08-18 18:37:25 +00:00
if ((endpts != MP_NOPTS_VALUE && img->pts >= endpts) ||
mpctx->max_frames == 0)
{
vf_unread_output_frame(vo_c->vf, img);
img = NULL;
video: make frame skipping code slightly more readable
2015-03-25 12:58:14 +00:00
r = VD_EOF;
} else if (hrseek && mpctx->hrseek_lastframe) {
player: when seeking past EOF with --keep-open, seek to last frame It feels strange that seeking past EOF with --keep-open actually leaves the player at a random position. You can't even unpause, because the demuxer is in the EOF state, and what you see on screen is just what was around before the seek. Improve this by attempting to seek to the last video frame if EOF happens. We explicitly don't do this if EOF was reached normally to increase robustness (if the VO got a frame since the last seek, it obviously means we had normal playback before EOF). If an error happens when trying to find the last frame (such as not actually finding a last frame because e.g. the demuxer misbehaves), this will probably turn your CPU into a heater. There is no logic to prevent reinitiating the last-frame search if the last-frame search reached EOF. (Pausing usually prevents that EOF is reached again after a successful last-frame search.) Fixes #819.
2014-12-07 01:47:09 +00:00
mp_image_setrefp(&mpctx->saved_frame, img);
video: make frame skipping code slightly more readable
2015-03-25 12:58:14 +00:00
} else if (hrseek && img->pts < mpctx->hrseek_pts - .005) {
player: simplify backstepping Basically reimplement it. The old implementation was quite stupid, and was probably done this way because video filtering and output used to be way less decoupled. Now we can reimplement it in a very simple way: when backstepping, seek to current time, but keep the last frame that was supposed to be discarded when reaching the target time. When the seek finishes, prepend the saved frame to the video frame queue. A disadvantage is that the new implementation fails to skip over timeline boundaries (ordered chapters etc.), but this never worked properly anyway. It's possible that this will be fixed some time in the future.
2016-01-12 22:49:00 +00:00
/* just skip - but save if backstep active */
if (mpctx->hrseek_backstep)
mp_image_setrefp(&mpctx->saved_frame, img);
video: slightly improve video stream switching Resync newly switched video streams to the current playback position. (Normal seeks will reset playback_pts to NOPTS.)
2016-01-26 13:06:41 +00:00
} else if (mpctx->video_status == STATUS_SYNCING &&
mpctx->playback_pts != MP_NOPTS_VALUE &&
player: use different variable to indicate coverart Slightly better.
2016-02-01 21:14:32 +00:00
img->pts < mpctx->playback_pts && !vo_c->is_coverart)
video: slightly improve video stream switching Resync newly switched video streams to the current playback position. (Normal seeks will reset playback_pts to NOPTS.)
2016-01-26 13:06:41 +00:00
{
/* skip after stream-switching */
video: filter new frames at a better time Move this code below the code that "shifts" the newly filtered frame. This allows us to skip a useless playloop iteration later, because obviously we need to filter a new frame after the previous frame has been "shifted", and not before that.
2014-09-18 17:17:38 +00:00
} else {
video: refactor: disentangle decoding/filtering some more This moves some code related to decoding from video.c to dec_video.c, and also removes some accesses to dec_video.c from the filtering code. dec_video.ch is starting to make sense, and simply returns video frames from a demuxer stream. The API exposed is also somewhat intended to be easily changeable to move decoding to a separate thread, if we ever want this (due to libavcodec already being threaded, I don't see much of a reason, but it might still be helpful).
2016-01-16 20:19:52 +00:00
if (hrseek && mpctx->hrseek_backstep) {
player: simplify backstepping Basically reimplement it. The old implementation was quite stupid, and was probably done this way because video filtering and output used to be way less decoupled. Now we can reimplement it in a very simple way: when backstepping, seek to current time, but keep the last frame that was supposed to be discarded when reaching the target time. When the seek finishes, prepend the saved frame to the video frame queue. A disadvantage is that the new implementation fails to skip over timeline boundaries (ordered chapters etc.), but this never worked properly anyway. It's possible that this will be fixed some time in the future.
2016-01-12 22:49:00 +00:00
if (mpctx->saved_frame) {
add_new_frame(mpctx, mpctx->saved_frame);
mpctx->saved_frame = NULL;
} else {
MP_WARN(mpctx, "Backstep failed.\n");
}
mpctx->hrseek_backstep = false;
}
video: refactor aspects of queue and EOF handling Instead of touching the 2-entry queue in mpctx->next_frame directly, move some of it to functions.
2015-03-25 12:33:34 +00:00
add_new_frame(mpctx, img);
video: make frame skipping code slightly more readable
2015-03-25 12:58:14 +00:00
img = NULL;
video: filter new frames at a better time Move this code below the code that "shifts" the newly filtered frame. This allows us to skip a useless playloop iteration later, because obviously we need to filter a new frame after the previous frame has been "shifted", and not before that.
2014-09-18 17:17:38 +00:00
}
video: make frame skipping code slightly more readable
2015-03-25 12:58:14 +00:00
talloc_free(img);
video: filter new frames at a better time Move this code below the code that "shifts" the newly filtered frame. This allows us to skip a useless playloop iteration later, because obviously we need to filter a new frame after the previous frame has been "shifted", and not before that.
2014-09-18 17:17:38 +00:00
}
}
player: when seeking past EOF with --keep-open, seek to last frame It feels strange that seeking past EOF with --keep-open actually leaves the player at a random position. You can't even unpause, because the demuxer is in the EOF state, and what you see on screen is just what was around before the seek. Improve this by attempting to seek to the last video frame if EOF happens. We explicitly don't do this if EOF was reached normally to increase robustness (if the VO got a frame since the last seek, it obviously means we had normal playback before EOF). If an error happens when trying to find the last frame (such as not actually finding a last frame because e.g. the demuxer misbehaves), this will probably turn your CPU into a heater. There is no logic to prevent reinitiating the last-frame search if the last-frame search reached EOF. (Pausing usually prevents that EOF is reached again after a successful last-frame search.) Fixes #819.
2014-12-07 01:47:09 +00:00
// Last-frame seek
video: fix seek-to-last-frame Accidentally broken in 79779616; we really need to check for true EOF, not just whether there are no frames yet.
2015-03-26 10:57:48 +00:00
if (r <= 0 && hrseek && mpctx->hrseek_lastframe && mpctx->saved_frame) {
video: refactor aspects of queue and EOF handling Instead of touching the 2-entry queue in mpctx->next_frame directly, move some of it to functions.
2015-03-25 12:33:34 +00:00
add_new_frame(mpctx, mpctx->saved_frame);
player: when seeking past EOF with --keep-open, seek to last frame It feels strange that seeking past EOF with --keep-open actually leaves the player at a random position. You can't even unpause, because the demuxer is in the EOF state, and what you see on screen is just what was around before the seek. Improve this by attempting to seek to the last video frame if EOF happens. We explicitly don't do this if EOF was reached normally to increase robustness (if the VO got a frame since the last seek, it obviously means we had normal playback before EOF). If an error happens when trying to find the last frame (such as not actually finding a last frame because e.g. the demuxer misbehaves), this will probably turn your CPU into a heater. There is no logic to prevent reinitiating the last-frame search if the last-frame search reached EOF. (Pausing usually prevents that EOF is reached again after a successful last-frame search.) Fixes #819.
2014-12-07 01:47:09 +00:00
mpctx->saved_frame = NULL;
video: refactor aspects of queue and EOF handling Instead of touching the 2-entry queue in mpctx->next_frame directly, move some of it to functions.
2015-03-25 12:33:34 +00:00
r = VD_PROGRESS;
player: when seeking past EOF with --keep-open, seek to last frame It feels strange that seeking past EOF with --keep-open actually leaves the player at a random position. You can't even unpause, because the demuxer is in the EOF state, and what you see on screen is just what was around before the seek. Improve this by attempting to seek to the last video frame if EOF happens. We explicitly don't do this if EOF was reached normally to increase robustness (if the VO got a frame since the last seek, it obviously means we had normal playback before EOF). If an error happens when trying to find the last frame (such as not actually finding a last frame because e.g. the demuxer misbehaves), this will probably turn your CPU into a heater. There is no logic to prevent reinitiating the last-frame search if the last-frame search reached EOF. (Pausing usually prevents that EOF is reached again after a successful last-frame search.) Fixes #819.
2014-12-07 01:47:09 +00:00
}
video: refactor aspects of queue and EOF handling Instead of touching the 2-entry queue in mpctx->next_frame directly, move some of it to functions.
2015-03-25 12:33:34 +00:00
return have_new_frame(mpctx, r <= 0) ? VD_NEW_FRAME : r;
Split mplayer.c mplayer.c was a bit too big. Split it into multiple files. I hope the way it's split makes sense. Maybe some things don't make too much sense, or go against intuition. These will fixed as soon as I notice them. Some files are a bit questionable (misc.c, osd.c, configfiles.c), and suggestions how to organize this better are welcome. Regressions are possible due to reorganized include statements. Obviously I didn't just copy mplayer.c's orgy of include statements, but recreated them for each file. It's easily possible that there are oversights and mistakes, which will show up on other platforms. There is one actual change: the public avutil.h include is removed from encode.h, and I tried to replace most FFMIN/FFMAX/av_clip uses. I consider using libavutil too much as dangerous, because the set of include files they recursively pull in is rather arbitrary and is different between FFmpeg and Libav.
2013-10-29 21:38:29 +00:00
}
player: move video display code out of the playloop Basically move the code from playloop.c to video.c. The new function write_video() now contains the code that was part of run_playloop(). There are no functional changes, except handling "new_frame_shown" slightly differently. This is done so that we don't need new a new MPContext field or a return value for write_video() to signal this condition. Instead, it's handled indirectly.
2014-07-30 21:24:08 +00:00
video: move some more code around No functional changes.
2014-08-13 19:54:13 +00:00
/* Update avsync before a new video frame is displayed. Actually, this can be
* called arbitrarily often before the actual display.
* This adjusts the time of the next video frame */
static void update_avsync_before_frame(struct MPContext *mpctx)
{
struct MPOpts *opts = mpctx->opts;
struct vo *vo = mpctx->video_out;
player: use different variable to indicate coverart Slightly better.
2016-02-01 21:14:32 +00:00
if (mpctx->vo_chain->is_coverart || mpctx->video_status < STATUS_READY) {
video: move some more code around No functional changes.
2014-08-13 19:54:13 +00:00
mpctx->time_frame = 0;
player: add display sync mode If this mode is enabled, the player tries to strictly synchronize video to display refresh. It will adjust playback speed to match the display, so if you play 23.976 fps video on a 24 Hz screen, playback speed is increased by approximately 1/1000. Audio wll be resampled to keep up with playback. This is different from the default sync mode, which will sync video to audio, with the consequence that video might skip or repeat a frame once in a while to make video keep up with audio. This is still unpolished. There are some major problems as well; in particular, mkv VFR files won't work well. The reason is that Matroska is terrible and rounds timestamps to milliseconds. This makes it rather hard to guess the framerate of a section of video that is playing. We could probably fix this by just accepting jittery timestamps (instead of explicitly disabling the sync code in this case), but I'm not ready to accept such a solution yet. Another issue is that we are extremely reliant on OS video and audio APIs working in an expected manner, which of course is not too often the case. Consequently, the new sync mode is a bit fragile.
2015-08-10 16:43:25 +00:00
} else if (mpctx->display_sync_active || opts->video_sync == VS_NONE) {
// don't touch the timing
video: move some more code around No functional changes.
2014-08-13 19:54:13 +00:00
} else if (mpctx->audio_status == STATUS_PLAYING &&
mpctx->video_status == STATUS_PLAYING &&
!ao_untimed(mpctx->ao))
{
double buffered_audio = ao_get_delay(mpctx->ao);
player: dump audio jitter to stats file This allows us to plot the difference between video timestamps, and the adjusted video timestamps due to syncing video to audio speed.
2015-02-01 17:32:26 +00:00
player: separate controls for user and video controlled speed For video sync, we want separate playback speed controls for user- requested speed and the "correction" speed for video timing. Further, we use this separation to make sure only a resampler is inserted if playback speed is only changed for video sync correction. As of this commit, this is basically inactive code. It's just preparation for the video sync code (the following commit).
2015-08-10 16:40:16 +00:00
double predicted = mpctx->delay / mpctx->video_speed +
player: dump audio jitter to stats file This allows us to plot the difference between video timestamps, and the adjusted video timestamps due to syncing video to audio speed.
2015-02-01 17:32:26 +00:00
mpctx->time_frame;
double difference = buffered_audio - predicted;
MP_STATS(mpctx, "value %f audio-diff", difference);
video: move some more code around No functional changes.
2014-08-13 19:54:13 +00:00
if (opts->autosync) {
/* Smooth reported playback position from AO by averaging
* it with the value expected based on previus value and
* time elapsed since then. May help smooth video timing
* with audio output that have inaccurate position reporting.
* This is badly implemented; the behavior of the smoothing
* now undesirably depends on how often this code runs
* (mainly depends on video frame rate). */
buffered_audio = predicted + difference / opts->autosync;
}
player: separate controls for user and video controlled speed For video sync, we want separate playback speed controls for user- requested speed and the "correction" speed for video timing. Further, we use this separation to make sure only a resampler is inserted if playback speed is only changed for video sync correction. As of this commit, this is basically inactive code. It's just preparation for the video sync code (the following commit).
2015-08-10 16:40:16 +00:00
mpctx->time_frame = buffered_audio - mpctx->delay / mpctx->video_speed;
video: move some more code around No functional changes.
2014-08-13 19:54:13 +00:00
} else {
/* If we're more than 200 ms behind the right playback
* position, don't try to speed up display of following
* frames to catch up; continue with default speed from
* the current frame instead.
* If untimed is set always output frames immediately
* without sleeping.
*/
if (mpctx->time_frame < -0.2 || opts->untimed || vo->driver->untimed)
mpctx->time_frame = 0;
}
}
player: minor refactor for A/V diff computation Calculate the A/V difference directly in the display sync code, instead of the awkward current way, which reuses the fields for audio sync. We still set time_frame, because it makes falling back to audio sync somewhat smoother.
2015-10-28 22:50:54 +00:00
// Update the A/V sync difference when a new video frame is being shown.
static void update_av_diff(struct MPContext *mpctx, double offset)
player: move video display code out of the playloop Basically move the code from playloop.c to video.c. The new function write_video() now contains the code that was part of run_playloop(). There are no functional changes, except handling "new_frame_shown" slightly differently. This is done so that we don't need new a new MPContext field or a return value for write_video() to signal this condition. Instead, it's handled indirectly.
2014-07-30 21:24:08 +00:00
{
player: remove redundant variable mpctx->audio_delay always has the same value as opts->audio_delay. (This was not the case a long time ago, when the audio-delay property didn't actually write to opts->audio_delay. I think.)
2015-01-29 11:10:07 +00:00
struct MPOpts *opts = mpctx->opts;
player: reset last_av_difference if not applicable Don't let stale values linger around. Also fix a slightly related case in audio.c.
2014-09-19 20:38:00 +00:00
mpctx->last_av_difference = 0;
video: refactor queue handling This simplifies the code, and fixes an odd bug: the second-last frame was displayed for a very short duration if framedrop was enabled. The reason was that basically the time difference between second-last and last frame were skipped, because at this point EOF was already signaled. Also see commit b0959488 for a similar issue in the same code. This removes the messiness of the next_frame 2-frame queue, and strictly runs the "new frame" code when a frame is moved to the first position of the queue, instead of somehow messing with return codes. This also merges update_video() into video_output_image().
2014-08-22 12:20:30 +00:00
player: move video display code out of the playloop Basically move the code from playloop.c to video.c. The new function write_video() now contains the code that was part of run_playloop(). There are no functional changes, except handling "new_frame_shown" slightly differently. This is done so that we don't need new a new MPContext field or a return value for write_video() to signal this condition. Instead, it's handled indirectly.
2014-07-30 21:24:08 +00:00
if (mpctx->audio_status != STATUS_PLAYING ||
mpctx->video_status != STATUS_PLAYING)
return;
double a_pos = playing_audio_pts(mpctx);
player: minor refactor for A/V diff computation Calculate the A/V difference directly in the display sync code, instead of the awkward current way, which reuses the fields for audio sync. We still set time_frame, because it makes falling back to audio sync somewhat smoother.
2015-10-28 22:50:54 +00:00
if (a_pos != MP_NOPTS_VALUE && mpctx->video_pts != MP_NOPTS_VALUE) {
mpctx->last_av_difference = a_pos - mpctx->video_pts
+ opts->audio_delay + offset;
}
player: move video display code out of the playloop Basically move the code from playloop.c to video.c. The new function write_video() now contains the code that was part of run_playloop(). There are no functional changes, except handling "new_frame_shown" slightly differently. This is done so that we don't need new a new MPContext field or a return value for write_video() to signal this condition. Instead, it's handled indirectly.
2014-07-30 21:24:08 +00:00
player: minor refactor for A/V diff computation Calculate the A/V difference directly in the display sync code, instead of the awkward current way, which reuses the fields for audio sync. We still set time_frame, because it makes falling back to audio sync somewhat smoother.
2015-10-28 22:50:54 +00:00
if (fabs(mpctx->last_av_difference) > 0.5 && !mpctx->drop_message_shown) {
player: move video display code out of the playloop Basically move the code from playloop.c to video.c. The new function write_video() now contains the code that was part of run_playloop(). There are no functional changes, except handling "new_frame_shown" slightly differently. This is done so that we don't need new a new MPContext field or a return value for write_video() to signal this condition. Instead, it's handled indirectly.
2014-07-30 21:24:08 +00:00
MP_WARN(mpctx, "%s", av_desync_help_text);
mpctx->drop_message_shown = true;
}
}
video: move some code around No functional changes. init_vo() is now needed a bit further down, and moving it keeps definition and use close. adjust_sync() will be used by a function further up in one of the following commits.
2014-08-22 12:19:01 +00:00
static void init_vo(struct MPContext *mpctx)
{
struct MPOpts *opts = mpctx->opts;
video: decouple filtering/decoding slightly more Lots of noise to remove the vfilter/vo fields from dec_video. From now on, video filtering and output will still be done together, summarized under struct vo_chain. There is the question where exactly the vf_chain should go in such a decoupled architecture. The end goal is being able to place a "complex" filter between video decoders and output (which will culminate in natural integration of A->V filters for natural integration of libavfilter audio visualizations). The vf_chain is still useful for "final" processing, such as format conversions and deinterlacing. Also, there's only 1 VO and 1 --vf option. So having 1 vf_chain for a VO seems ideal, since otherwise there would be no natural way to handle all these existing options and mechanisms. There is still some work required to truly decouple decoding.
2016-01-13 23:18:48 +00:00
struct vo_chain *vo_c = mpctx->vo_chain;
video: move some code around No functional changes. init_vo() is now needed a bit further down, and moving it keeps definition and use close. adjust_sync() will be used by a function further up in one of the following commits.
2014-08-22 12:19:01 +00:00
player: disable video equalizer frontend code for WIP LGPL mode Nick and kiriuja could not be reached, and created/changed this in 92c5c274, 6441a5ad, bffd4007, 555c6766, c2c997fd. The video equalizer stuff was redone fully later, but there are still parts that look too similar and basically use the same approach. I'm more comfortable with declaring it GPL only for now. I plan to redo them later in a way that will remove copyright.
2017-06-23 13:13:47 +00:00
#if HAVE_GPL
command: fix or document some property/option consistency issues Make some existing properties behave more like options. This mostly means they don't deny access if the associated component is not active, but redirects to the option. One kind of fishy change is that we apply --brightness etc. only if they're not set to the default value. This won't necessarily work with --vo=xv, but affects only cases where 1. the Xv adapter has been changed to non-defaults, and 2. the user tries to reset them with mpv by passing e.g. --brightness=0. We don't care about Xv, and the noted use-case is dumb, so this change is acceptable.
2016-09-01 18:57:33 +00:00
if (opts->gamma_gamma != 0)
video: decouple filtering/decoding slightly more Lots of noise to remove the vfilter/vo fields from dec_video. From now on, video filtering and output will still be done together, summarized under struct vo_chain. There is the question where exactly the vf_chain should go in such a decoupled architecture. The end goal is being able to place a "complex" filter between video decoders and output (which will culminate in natural integration of A->V filters for natural integration of libavfilter audio visualizations). The vf_chain is still useful for "final" processing, such as format conversions and deinterlacing. Also, there's only 1 VO and 1 --vf option. So having 1 vf_chain for a VO seems ideal, since otherwise there would be no natural way to handle all these existing options and mechanisms. There is still some work required to truly decouple decoding.
2016-01-13 23:18:48 +00:00
video_set_colors(vo_c, "gamma", opts->gamma_gamma);
command: fix or document some property/option consistency issues Make some existing properties behave more like options. This mostly means they don't deny access if the associated component is not active, but redirects to the option. One kind of fishy change is that we apply --brightness etc. only if they're not set to the default value. This won't necessarily work with --vo=xv, but affects only cases where 1. the Xv adapter has been changed to non-defaults, and 2. the user tries to reset them with mpv by passing e.g. --brightness=0. We don't care about Xv, and the noted use-case is dumb, so this change is acceptable.
2016-09-01 18:57:33 +00:00
if (opts->gamma_brightness != 0)
video: decouple filtering/decoding slightly more Lots of noise to remove the vfilter/vo fields from dec_video. From now on, video filtering and output will still be done together, summarized under struct vo_chain. There is the question where exactly the vf_chain should go in such a decoupled architecture. The end goal is being able to place a "complex" filter between video decoders and output (which will culminate in natural integration of A->V filters for natural integration of libavfilter audio visualizations). The vf_chain is still useful for "final" processing, such as format conversions and deinterlacing. Also, there's only 1 VO and 1 --vf option. So having 1 vf_chain for a VO seems ideal, since otherwise there would be no natural way to handle all these existing options and mechanisms. There is still some work required to truly decouple decoding.
2016-01-13 23:18:48 +00:00
video_set_colors(vo_c, "brightness", opts->gamma_brightness);
command: fix or document some property/option consistency issues Make some existing properties behave more like options. This mostly means they don't deny access if the associated component is not active, but redirects to the option. One kind of fishy change is that we apply --brightness etc. only if they're not set to the default value. This won't necessarily work with --vo=xv, but affects only cases where 1. the Xv adapter has been changed to non-defaults, and 2. the user tries to reset them with mpv by passing e.g. --brightness=0. We don't care about Xv, and the noted use-case is dumb, so this change is acceptable.
2016-09-01 18:57:33 +00:00
if (opts->gamma_contrast != 0)
video: decouple filtering/decoding slightly more Lots of noise to remove the vfilter/vo fields from dec_video. From now on, video filtering and output will still be done together, summarized under struct vo_chain. There is the question where exactly the vf_chain should go in such a decoupled architecture. The end goal is being able to place a "complex" filter between video decoders and output (which will culminate in natural integration of A->V filters for natural integration of libavfilter audio visualizations). The vf_chain is still useful for "final" processing, such as format conversions and deinterlacing. Also, there's only 1 VO and 1 --vf option. So having 1 vf_chain for a VO seems ideal, since otherwise there would be no natural way to handle all these existing options and mechanisms. There is still some work required to truly decouple decoding.
2016-01-13 23:18:48 +00:00
video_set_colors(vo_c, "contrast", opts->gamma_contrast);
command: fix or document some property/option consistency issues Make some existing properties behave more like options. This mostly means they don't deny access if the associated component is not active, but redirects to the option. One kind of fishy change is that we apply --brightness etc. only if they're not set to the default value. This won't necessarily work with --vo=xv, but affects only cases where 1. the Xv adapter has been changed to non-defaults, and 2. the user tries to reset them with mpv by passing e.g. --brightness=0. We don't care about Xv, and the noted use-case is dumb, so this change is acceptable.
2016-09-01 18:57:33 +00:00
if (opts->gamma_saturation != 0)
video: decouple filtering/decoding slightly more Lots of noise to remove the vfilter/vo fields from dec_video. From now on, video filtering and output will still be done together, summarized under struct vo_chain. There is the question where exactly the vf_chain should go in such a decoupled architecture. The end goal is being able to place a "complex" filter between video decoders and output (which will culminate in natural integration of A->V filters for natural integration of libavfilter audio visualizations). The vf_chain is still useful for "final" processing, such as format conversions and deinterlacing. Also, there's only 1 VO and 1 --vf option. So having 1 vf_chain for a VO seems ideal, since otherwise there would be no natural way to handle all these existing options and mechanisms. There is still some work required to truly decouple decoding.
2016-01-13 23:18:48 +00:00
video_set_colors(vo_c, "saturation", opts->gamma_saturation);
command: fix or document some property/option consistency issues Make some existing properties behave more like options. This mostly means they don't deny access if the associated component is not active, but redirects to the option. One kind of fishy change is that we apply --brightness etc. only if they're not set to the default value. This won't necessarily work with --vo=xv, but affects only cases where 1. the Xv adapter has been changed to non-defaults, and 2. the user tries to reset them with mpv by passing e.g. --brightness=0. We don't care about Xv, and the noted use-case is dumb, so this change is acceptable.
2016-09-01 18:57:33 +00:00
if (opts->gamma_hue != 0)
video: decouple filtering/decoding slightly more Lots of noise to remove the vfilter/vo fields from dec_video. From now on, video filtering and output will still be done together, summarized under struct vo_chain. There is the question where exactly the vf_chain should go in such a decoupled architecture. The end goal is being able to place a "complex" filter between video decoders and output (which will culminate in natural integration of A->V filters for natural integration of libavfilter audio visualizations). The vf_chain is still useful for "final" processing, such as format conversions and deinterlacing. Also, there's only 1 VO and 1 --vf option. So having 1 vf_chain for a VO seems ideal, since otherwise there would be no natural way to handle all these existing options and mechanisms. There is still some work required to truly decouple decoding.
2016-01-13 23:18:48 +00:00
video_set_colors(vo_c, "hue", opts->gamma_hue);
video_set_colors(vo_c, "output-levels", opts->video_output_levels);
player: disable video equalizer frontend code for WIP LGPL mode Nick and kiriuja could not be reached, and created/changed this in 92c5c274, 6441a5ad, bffd4007, 555c6766, c2c997fd. The video equalizer stuff was redone fully later, but there are still parts that look too similar and basically use the same approach. I'm more comfortable with declaring it GPL only for now. I plan to redo them later in a way that will remove copyright.
2017-06-23 13:13:47 +00:00
#endif
video: move some code around No functional changes. init_vo() is now needed a bit further down, and moving it keeps definition and use close. adjust_sync() will be used by a function further up in one of the following commits.
2014-08-22 12:19:01 +00:00
mp_notify(mpctx, MPV_EVENT_VIDEO_RECONFIG, NULL);
}
player: refactor display-sync frame duration calculations Get rid of get_past_frame_durations(), which was a bit too messy. Add a past_frames array, which contains the same information in a more reasonable way. This also means that we can get the exact current and past frame durations without going through awful stuff. (The main problem is that vo_pts_history contains future frames as well, which is needed for frame backstepping etc., but gets in the way here.) Also disable the automatic disabling of display-sync if the frame duration changes, and extend the frame durations allowed for display sync. To allow arbitrarily high durations, vo.c needs to be changed to pause and potentially redraw OSD while showing a single frame, so they're still limited. In an attempt to deal with VFR, calculate the overall speed using the average FPS. The frame scheduling itself does not use the average FPS, but the duration of the current frame. This does not work too well, but provides a good base for further improvements. Where this commit actually helps a lot is dealing with rounded timestamps, e.g. if the container framerate is wrong or unknown, or if the muxer wrote incorrectly rounded timestamps. While the rounding errors apparently can't be get rid of completely in the general case, this is still much better than e.g. disabling display-sync completely just because some frame durations go out of bounds.
2015-11-13 21:45:40 +00:00
double calc_average_frame_duration(struct MPContext *mpctx)
player: redo estimated-vf-fps calculation Additionally to taking the average, this tries to use the demuxer FPS to eliminate jitter, and applies some other heuristics to check if the result is sane. This code will also be used for the display sync code (it will actually make use of the require_exact parameter). (The value of doing this over keeping the simpler demux_mkv hack is somewhat questionable. But at least it allows us to deal with other container formats that use jittery timestamps, such as mp4 remuxed from mkv.)
2015-08-10 16:38:57 +00:00
{
player: refactor display-sync frame duration calculations Get rid of get_past_frame_durations(), which was a bit too messy. Add a past_frames array, which contains the same information in a more reasonable way. This also means that we can get the exact current and past frame durations without going through awful stuff. (The main problem is that vo_pts_history contains future frames as well, which is needed for frame backstepping etc., but gets in the way here.) Also disable the automatic disabling of display-sync if the frame duration changes, and extend the frame durations allowed for display sync. To allow arbitrarily high durations, vo.c needs to be changed to pause and potentially redraw OSD while showing a single frame, so they're still limited. In an attempt to deal with VFR, calculate the overall speed using the average FPS. The frame scheduling itself does not use the average FPS, but the duration of the current frame. This does not work too well, but provides a good base for further improvements. Where this commit actually helps a lot is dealing with rounded timestamps, e.g. if the container framerate is wrong or unknown, or if the muxer wrote incorrectly rounded timestamps. While the rounding errors apparently can't be get rid of completely in the general case, this is still much better than e.g. disabling display-sync completely just because some frame durations go out of bounds.
2015-11-13 21:45:40 +00:00
double total = 0;
int num = 0;
for (int n = 0; n < mpctx->num_past_frames; n++) {
player: fix average frame duration calculation
2016-09-09 16:56:45 +00:00
double dur = mpctx->past_frames[n].approx_duration;
player: refactor display-sync frame duration calculations Get rid of get_past_frame_durations(), which was a bit too messy. Add a past_frames array, which contains the same information in a more reasonable way. This also means that we can get the exact current and past frame durations without going through awful stuff. (The main problem is that vo_pts_history contains future frames as well, which is needed for frame backstepping etc., but gets in the way here.) Also disable the automatic disabling of display-sync if the frame duration changes, and extend the frame durations allowed for display sync. To allow arbitrarily high durations, vo.c needs to be changed to pause and potentially redraw OSD while showing a single frame, so they're still limited. In an attempt to deal with VFR, calculate the overall speed using the average FPS. The frame scheduling itself does not use the average FPS, but the duration of the current frame. This does not work too well, but provides a good base for further improvements. Where this commit actually helps a lot is dealing with rounded timestamps, e.g. if the container framerate is wrong or unknown, or if the muxer wrote incorrectly rounded timestamps. While the rounding errors apparently can't be get rid of completely in the general case, this is still much better than e.g. disabling display-sync completely just because some frame durations go out of bounds.
2015-11-13 21:45:40 +00:00
if (dur <= 0)
continue;
total += dur;
num += 1;
player: redo estimated-vf-fps calculation Additionally to taking the average, this tries to use the demuxer FPS to eliminate jitter, and applies some other heuristics to check if the result is sane. This code will also be used for the display sync code (it will actually make use of the require_exact parameter). (The value of doing this over keeping the simpler demux_mkv hack is somewhat questionable. But at least it allows us to deal with other container formats that use jittery timestamps, such as mp4 remuxed from mkv.)
2015-08-10 16:38:57 +00:00
}
player: refactor display-sync frame duration calculations Get rid of get_past_frame_durations(), which was a bit too messy. Add a past_frames array, which contains the same information in a more reasonable way. This also means that we can get the exact current and past frame durations without going through awful stuff. (The main problem is that vo_pts_history contains future frames as well, which is needed for frame backstepping etc., but gets in the way here.) Also disable the automatic disabling of display-sync if the frame duration changes, and extend the frame durations allowed for display sync. To allow arbitrarily high durations, vo.c needs to be changed to pause and potentially redraw OSD while showing a single frame, so they're still limited. In an attempt to deal with VFR, calculate the overall speed using the average FPS. The frame scheduling itself does not use the average FPS, but the duration of the current frame. This does not work too well, but provides a good base for further improvements. Where this commit actually helps a lot is dealing with rounded timestamps, e.g. if the container framerate is wrong or unknown, or if the muxer wrote incorrectly rounded timestamps. While the rounding errors apparently can't be get rid of completely in the general case, this is still much better than e.g. disabling display-sync completely just because some frame durations go out of bounds.
2015-11-13 21:45:40 +00:00
return num > 0 ? total / num : 0;
player: redo estimated-vf-fps calculation Additionally to taking the average, this tries to use the demuxer FPS to eliminate jitter, and applies some other heuristics to check if the result is sane. This code will also be used for the display sync code (it will actually make use of the require_exact parameter). (The value of doing this over keeping the simpler demux_mkv hack is somewhat questionable. But at least it allows us to deal with other container formats that use jittery timestamps, such as mp4 remuxed from mkv.)
2015-08-10 16:38:57 +00:00
}
player: compute required display-sync speed change differently Instead of looking at the current frame duration for the intended speedup, look at all past frames, and find a good average speed. This ties in with not wanting to average _all_ frame durations, which doesn't make sense in VFR situations. This is currently done in the most naive way possible, but already sort of works for VFR which switches between frame durations that are integer multiples of a base rate. Certainly more improvements could be made, such as trying to adjust directly on FPS changes, instead of averaging everything, but for now this is not needed at all.
2015-11-13 21:47:14 +00:00
// Find a speed factor such that the display FPS is an integer multiple of the
// effective video FPS. If this is not possible, try to do it for multiples,
// which still leads to an improved end result.
// Both parameters are durations in seconds.
static double calc_best_speed(double vsync, double frame)
{
double ratio = frame / vsync;
double best_scale = -1;
double best_dev = INFINITY;
for (int factor = 1; factor <= 5; factor++) {
double scale = ratio * factor / rint(ratio * factor);
double dev = fabs(scale - 1);
if (dev < best_dev) {
best_scale = scale;
best_dev = dev;
}
}
return best_scale;
}
static double find_best_speed(struct MPContext *mpctx, double vsync)
{
double total = 0;
int num = 0;
for (int n = 0; n < mpctx->num_past_frames; n++) {
double dur = mpctx->past_frames[n].approx_duration;
if (dur <= 0)
continue;
total += calc_best_speed(vsync, dur / mpctx->opts->playback_speed);
num++;
}
return num > 0 ? total / num : 1;
}
player: add display sync mode If this mode is enabled, the player tries to strictly synchronize video to display refresh. It will adjust playback speed to match the display, so if you play 23.976 fps video on a 24 Hz screen, playback speed is increased by approximately 1/1000. Audio wll be resampled to keep up with playback. This is different from the default sync mode, which will sync video to audio, with the consequence that video might skip or repeat a frame once in a while to make video keep up with audio. This is still unpolished. There are some major problems as well; in particular, mkv VFR files won't work well. The reason is that Matroska is terrible and rounds timestamps to milliseconds. This makes it rather hard to guess the framerate of a section of video that is playing. We could probably fix this by just accepting jittery timestamps (instead of explicitly disabling the sync code in this case), but I'm not ready to accept such a solution yet. Another issue is that we are extremely reliant on OS video and audio APIs working in an expected manner, which of course is not too often the case. Consequently, the new sync mode is a bit fragile.
2015-08-10 16:43:25 +00:00
static bool using_spdif_passthrough(struct MPContext *mpctx)
{
player: disable DS with spdif transcoding too Otherwise it behaves dumb. (Although you could argue it shouldn't try to guess whether speed changes work, but instead simply disable DS if they don't work.)
2016-07-24 17:19:24 +00:00
if (mpctx->ao_chain && mpctx->ao_chain->ao) {
struct mp_audio out_format = {0};
ao_get_format(mpctx->ao_chain->ao, &out_format);
return !af_fmt_is_pcm(out_format.format);
}
player: add display sync mode If this mode is enabled, the player tries to strictly synchronize video to display refresh. It will adjust playback speed to match the display, so if you play 23.976 fps video on a 24 Hz screen, playback speed is increased by approximately 1/1000. Audio wll be resampled to keep up with playback. This is different from the default sync mode, which will sync video to audio, with the consequence that video might skip or repeat a frame once in a while to make video keep up with audio. This is still unpolished. There are some major problems as well; in particular, mkv VFR files won't work well. The reason is that Matroska is terrible and rounds timestamps to milliseconds. This makes it rather hard to guess the framerate of a section of video that is playing. We could probably fix this by just accepting jittery timestamps (instead of explicitly disabling the sync code in this case), but I'm not ready to accept such a solution yet. Another issue is that we are extremely reliant on OS video and audio APIs working in an expected manner, which of course is not too often the case. Consequently, the new sync mode is a bit fragile.
2015-08-10 16:43:25 +00:00
return false;
}
player: fix audio drift computation at different playback speeds This computed nonsense if the user set a playback speed other than 1 (in addition to the display-sync speed change).
2015-11-14 20:42:18 +00:00
// Compute the relative audio speed difference by taking A/V dsync into account.
static double compute_audio_drift(struct MPContext *mpctx, double vsync)
player: try to compensate actual audio drift We always let audio slowly desync until a threshold is reached, and then pushed it back by applying a maximum compensation speed. Refine what comes afterwards: instead of playing with the nominal video speed, use the actual required audio speed for keeping sync as measured by the A/V difference. (The "actual" speed is the ideal speed with A/V differences added.) Although this works in theory, it's somewhat questionable how much this works in practice. The ideal time value is actually not exact, but is the time at which the frame is scheduled (could be compensated by using the time_left calculations in handle_display_sync_frame()). It doesn't account for speed changes or catastrophic discontinuities. It uses only 10 past frames.
2015-11-13 21:51:39 +00:00
{
player: fix audio drift computation at different playback speeds This computed nonsense if the user set a playback speed other than 1 (in addition to the display-sync speed change).
2015-11-14 20:42:18 +00:00
// Least-squares linear regression, using relative real time for x, and
// audio desync for y. Assume speed didn't change for the frames we're
player: try to compensate actual audio drift We always let audio slowly desync until a threshold is reached, and then pushed it back by applying a maximum compensation speed. Refine what comes afterwards: instead of playing with the nominal video speed, use the actual required audio speed for keeping sync as measured by the A/V difference. (The "actual" speed is the ideal speed with A/V differences added.) Although this works in theory, it's somewhat questionable how much this works in practice. The ideal time value is actually not exact, but is the time at which the frame is scheduled (could be compensated by using the time_left calculations in handle_display_sync_frame()). It doesn't account for speed changes or catastrophic discontinuities. It uses only 10 past frames.
2015-11-13 21:51:39 +00:00
// looking at for simplicity. This also should actually use the realtime
// (minus paused time) for x, but use vsync scheduling points instead.
if (mpctx->num_past_frames <= 10)
return NAN;
int num = mpctx->num_past_frames - 1;
double sum_x = 0, sum_y = 0, sum_xy = 0, sum_xx = 0;
double x = 0;
for (int n = 0; n < num; n++) {
struct frame_info *frame = &mpctx->past_frames[n + 1];
if (frame->num_vsyncs < 0)
return NAN;
player: fix audio drift computation at different playback speeds This computed nonsense if the user set a playback speed other than 1 (in addition to the display-sync speed change).
2015-11-14 20:42:18 +00:00
double y = frame->av_diff;
player: try to compensate actual audio drift We always let audio slowly desync until a threshold is reached, and then pushed it back by applying a maximum compensation speed. Refine what comes afterwards: instead of playing with the nominal video speed, use the actual required audio speed for keeping sync as measured by the A/V difference. (The "actual" speed is the ideal speed with A/V differences added.) Although this works in theory, it's somewhat questionable how much this works in practice. The ideal time value is actually not exact, but is the time at which the frame is scheduled (could be compensated by using the time_left calculations in handle_display_sync_frame()). It doesn't account for speed changes or catastrophic discontinuities. It uses only 10 past frames.
2015-11-13 21:51:39 +00:00
sum_x += x;
sum_y += y;
sum_xy += x * y;
sum_xx += x * x;
x -= frame->num_vsyncs * vsync;
}
return (sum_x * sum_y - num * sum_xy) / (sum_x * sum_x - num * sum_xx);
}
player: change display-sync audio speed only if needed As long as it's within the desync tolerance, do not change the audio speed at all for resampling. This reduces speed changes which might be caused by jittering timestamps and similar cases. (While in theory you could just not care and change speed every single frame, I'm afraid that such changes could possibly cause audio artifacts. So better just avoid it in the first place.)
2015-11-13 21:50:58 +00:00
static void adjust_audio_resample_speed(struct MPContext *mpctx, double vsync)
player: move audio speed adjustment code Move it (in a cosmetic sense), and also move its invocation to below all the video handling. All other changes remain cosmetic, including moving the framedrop calculation code, and getting rid of the video_speed_correction variable.
2015-11-04 18:37:15 +00:00
{
struct MPOpts *opts = mpctx->opts;
int mode = opts->video_sync;
player: change display-sync audio speed only if needed As long as it's within the desync tolerance, do not change the audio speed at all for resampling. This reduces speed changes which might be caused by jittering timestamps and similar cases. (While in theory you could just not care and change speed every single frame, I'm afraid that such changes could possibly cause audio artifacts. So better just avoid it in the first place.)
2015-11-13 21:50:58 +00:00
if (mode != VS_DISP_RESAMPLE || mpctx->audio_status != STATUS_PLAYING) {
mpctx->speed_factor_a = mpctx->speed_factor_v;
return;
player: move audio speed adjustment code Move it (in a cosmetic sense), and also move its invocation to below all the video handling. All other changes remain cosmetic, including moving the framedrop calculation code, and getting rid of the video_speed_correction variable.
2015-11-04 18:37:15 +00:00
}
player: change display-sync audio speed only if needed As long as it's within the desync tolerance, do not change the audio speed at all for resampling. This reduces speed changes which might be caused by jittering timestamps and similar cases. (While in theory you could just not care and change speed every single frame, I'm afraid that such changes could possibly cause audio artifacts. So better just avoid it in the first place.)
2015-11-13 21:50:58 +00:00
// Try to smooth out audio timing drifts. This can happen if either
// video isn't playing at expected speed, or audio is not playing at
// the requested speed. Both are unavoidable.
// The audio desync is made up of 2 parts: 1. drift due to rounding
// errors and imperfect information, and 2. an offset, due to
// unaligned audio/video start, or disruptive events halting audio
// or video for a small time.
// Instead of trying to be clever, just apply an awfully dumb drift
// compensation with a constant factor, which does what we want. In
// theory we could calculate the exact drift compensation needed,
// but it likely would be wrong anyway, and we'd run into the same
// issues again, except with more complex code.
// 1 means drifts to positive, -1 means drifts to negative
double max_drift = vsync / 2;
double av_diff = mpctx->last_av_difference;
int new = mpctx->display_sync_drift_dir;
if (av_diff * -mpctx->display_sync_drift_dir >= 0)
new = 0;
if (fabs(av_diff) > max_drift)
new = av_diff >= 0 ? 1 : -1;
bool change = mpctx->display_sync_drift_dir != new;
if (new || change) {
if (change)
MP_VERBOSE(mpctx, "Change display sync audio drift: %d\n", new);
mpctx->display_sync_drift_dir = new;
player: move audio speed adjustment code Move it (in a cosmetic sense), and also move its invocation to below all the video handling. All other changes remain cosmetic, including moving the framedrop calculation code, and getting rid of the video_speed_correction variable.
2015-11-04 18:37:15 +00:00
player: change display-sync audio speed only if needed As long as it's within the desync tolerance, do not change the audio speed at all for resampling. This reduces speed changes which might be caused by jittering timestamps and similar cases. (While in theory you could just not care and change speed every single frame, I'm afraid that such changes could possibly cause audio artifacts. So better just avoid it in the first place.)
2015-11-13 21:50:58 +00:00
double max_correct = opts->sync_max_audio_change / 100;
double audio_factor = 1 + max_correct * -mpctx->display_sync_drift_dir;
player: try to compensate actual audio drift We always let audio slowly desync until a threshold is reached, and then pushed it back by applying a maximum compensation speed. Refine what comes afterwards: instead of playing with the nominal video speed, use the actual required audio speed for keeping sync as measured by the A/V difference. (The "actual" speed is the ideal speed with A/V differences added.) Although this works in theory, it's somewhat questionable how much this works in practice. The ideal time value is actually not exact, but is the time at which the frame is scheduled (could be compensated by using the time_left calculations in handle_display_sync_frame()). It doesn't account for speed changes or catastrophic discontinuities. It uses only 10 past frames.
2015-11-13 21:51:39 +00:00
if (new == 0) {
// If we're resetting, actually try to be clever and pick a speed
// which compensates the general drift we're getting.
player: fix audio drift computation at different playback speeds This computed nonsense if the user set a playback speed other than 1 (in addition to the display-sync speed change).
2015-11-14 20:42:18 +00:00
double drift = compute_audio_drift(mpctx, vsync);
player: try to compensate actual audio drift We always let audio slowly desync until a threshold is reached, and then pushed it back by applying a maximum compensation speed. Refine what comes afterwards: instead of playing with the nominal video speed, use the actual required audio speed for keeping sync as measured by the A/V difference. (The "actual" speed is the ideal speed with A/V differences added.) Although this works in theory, it's somewhat questionable how much this works in practice. The ideal time value is actually not exact, but is the time at which the frame is scheduled (could be compensated by using the time_left calculations in handle_display_sync_frame()). It doesn't account for speed changes or catastrophic discontinuities. It uses only 10 past frames.
2015-11-13 21:51:39 +00:00
if (isnormal(drift)) {
player: fix audio drift computation at different playback speeds This computed nonsense if the user set a playback speed other than 1 (in addition to the display-sync speed change).
2015-11-14 20:42:18 +00:00
// other = will be multiplied with audio_factor for final speed
double other = mpctx->opts->playback_speed * mpctx->speed_factor_v;
audio_factor = (mpctx->audio_speed - drift) / other;
player: try to compensate actual audio drift We always let audio slowly desync until a threshold is reached, and then pushed it back by applying a maximum compensation speed. Refine what comes afterwards: instead of playing with the nominal video speed, use the actual required audio speed for keeping sync as measured by the A/V difference. (The "actual" speed is the ideal speed with A/V differences added.) Although this works in theory, it's somewhat questionable how much this works in practice. The ideal time value is actually not exact, but is the time at which the frame is scheduled (could be compensated by using the time_left calculations in handle_display_sync_frame()). It doesn't account for speed changes or catastrophic discontinuities. It uses only 10 past frames.
2015-11-13 21:51:39 +00:00
MP_VERBOSE(mpctx, "Compensation factor: %f\n", audio_factor);
}
}
audio_factor = MPCLAMP(audio_factor, 1 - max_correct, 1 + max_correct);
player: change display-sync audio speed only if needed As long as it's within the desync tolerance, do not change the audio speed at all for resampling. This reduces speed changes which might be caused by jittering timestamps and similar cases. (While in theory you could just not care and change speed every single frame, I'm afraid that such changes could possibly cause audio artifacts. So better just avoid it in the first place.)
2015-11-13 21:50:58 +00:00
mpctx->speed_factor_a = audio_factor * mpctx->speed_factor_v;
}
player: move audio speed adjustment code Move it (in a cosmetic sense), and also move its invocation to below all the video handling. All other changes remain cosmetic, including moving the framedrop calculation code, and getting rid of the video_speed_correction variable.
2015-11-04 18:37:15 +00:00
}
player: add display sync mode If this mode is enabled, the player tries to strictly synchronize video to display refresh. It will adjust playback speed to match the display, so if you play 23.976 fps video on a 24 Hz screen, playback speed is increased by approximately 1/1000. Audio wll be resampled to keep up with playback. This is different from the default sync mode, which will sync video to audio, with the consequence that video might skip or repeat a frame once in a while to make video keep up with audio. This is still unpolished. There are some major problems as well; in particular, mkv VFR files won't work well. The reason is that Matroska is terrible and rounds timestamps to milliseconds. This makes it rather hard to guess the framerate of a section of video that is playing. We could probably fix this by just accepting jittery timestamps (instead of explicitly disabling the sync code in this case), but I'm not ready to accept such a solution yet. Another issue is that we are extremely reliant on OS video and audio APIs working in an expected manner, which of course is not too often the case. Consequently, the new sync mode is a bit fragile.
2015-08-10 16:43:25 +00:00
// Manipulate frame timing for display sync, or do nothing for normal timing.
static void handle_display_sync_frame(struct MPContext *mpctx,
struct vo_frame *frame)
{
struct MPOpts *opts = mpctx->opts;
struct vo *vo = mpctx->video_out;
int mode = opts->video_sync;
if (!mpctx->display_sync_active) {
mpctx->display_sync_error = 0.0;
mpctx->display_sync_drift_dir = 0;
}
mpctx->display_sync_active = false;
player: fix commit 50bb209a Well, this was stupid.
2015-11-28 01:24:03 +00:00
if (!VS_IS_DISP(mode) || mpctx->display_sync_broken)
player: always disable display-sync on desyncs Instead of periodically trying to enable it again. There are two cases that can happen: 1. A random discontinuity messed everything up, 2. Things are just broken and will desync all the time Until now, it tried to deal with case 1 - but maybe this is really rare, and we don't really need to care about it. On the other hand, case 2 is kind of hard to diagnose if the user doesn't use the terminal. Seeking will reenable display-sync, so you can fix playback if case 1 happens, but still get predictable behavior in case 2.
2015-11-27 13:40:52 +00:00
return;
player: add display sync mode If this mode is enabled, the player tries to strictly synchronize video to display refresh. It will adjust playback speed to match the display, so if you play 23.976 fps video on a 24 Hz screen, playback speed is increased by approximately 1/1000. Audio wll be resampled to keep up with playback. This is different from the default sync mode, which will sync video to audio, with the consequence that video might skip or repeat a frame once in a while to make video keep up with audio. This is still unpolished. There are some major problems as well; in particular, mkv VFR files won't work well. The reason is that Matroska is terrible and rounds timestamps to milliseconds. This makes it rather hard to guess the framerate of a section of video that is playing. We could probably fix this by just accepting jittery timestamps (instead of explicitly disabling the sync code in this case), but I'm not ready to accept such a solution yet. Another issue is that we are extremely reliant on OS video and audio APIs working in an expected manner, which of course is not too often the case. Consequently, the new sync mode is a bit fragile.
2015-08-10 16:43:25 +00:00
bool resample = mode == VS_DISP_RESAMPLE || mode == VS_DISP_RESAMPLE_VDROP ||
mode == VS_DISP_RESAMPLE_NONE;
bool drop = mode == VS_DISP_VDROP || mode == VS_DISP_RESAMPLE ||
player: add audio drop/duplicate mode Not very robust in the moment.
2015-10-27 19:56:46 +00:00
mode == VS_DISP_ADROP || mode == VS_DISP_RESAMPLE_VDROP;
player: add display sync mode If this mode is enabled, the player tries to strictly synchronize video to display refresh. It will adjust playback speed to match the display, so if you play 23.976 fps video on a 24 Hz screen, playback speed is increased by approximately 1/1000. Audio wll be resampled to keep up with playback. This is different from the default sync mode, which will sync video to audio, with the consequence that video might skip or repeat a frame once in a while to make video keep up with audio. This is still unpolished. There are some major problems as well; in particular, mkv VFR files won't work well. The reason is that Matroska is terrible and rounds timestamps to milliseconds. This makes it rather hard to guess the framerate of a section of video that is playing. We could probably fix this by just accepting jittery timestamps (instead of explicitly disabling the sync code in this case), but I'm not ready to accept such a solution yet. Another issue is that we are extremely reliant on OS video and audio APIs working in an expected manner, which of course is not too often the case. Consequently, the new sync mode is a bit fragile.
2015-08-10 16:43:25 +00:00
drop &= (opts->frame_dropping & 1);
if (resample && using_spdif_passthrough(mpctx))
player: always disable display-sync on desyncs Instead of periodically trying to enable it again. There are two cases that can happen: 1. A random discontinuity messed everything up, 2. Things are just broken and will desync all the time Until now, it tried to deal with case 1 - but maybe this is really rare, and we don't really need to care about it. On the other hand, case 2 is kind of hard to diagnose if the user doesn't use the terminal. Seeking will reenable display-sync, so you can fix playback if case 1 happens, but still get predictable behavior in case 2.
2015-11-27 13:40:52 +00:00
return;
player: add display sync mode If this mode is enabled, the player tries to strictly synchronize video to display refresh. It will adjust playback speed to match the display, so if you play 23.976 fps video on a 24 Hz screen, playback speed is increased by approximately 1/1000. Audio wll be resampled to keep up with playback. This is different from the default sync mode, which will sync video to audio, with the consequence that video might skip or repeat a frame once in a while to make video keep up with audio. This is still unpolished. There are some major problems as well; in particular, mkv VFR files won't work well. The reason is that Matroska is terrible and rounds timestamps to milliseconds. This makes it rather hard to guess the framerate of a section of video that is playing. We could probably fix this by just accepting jittery timestamps (instead of explicitly disabling the sync code in this case), but I'm not ready to accept such a solution yet. Another issue is that we are extremely reliant on OS video and audio APIs working in an expected manner, which of course is not too often the case. Consequently, the new sync mode is a bit fragile.
2015-08-10 16:43:25 +00:00
double vsync = vo_get_vsync_interval(vo) / 1e6;
if (vsync <= 0)
player: always disable display-sync on desyncs Instead of periodically trying to enable it again. There are two cases that can happen: 1. A random discontinuity messed everything up, 2. Things are just broken and will desync all the time Until now, it tried to deal with case 1 - but maybe this is really rare, and we don't really need to care about it. On the other hand, case 2 is kind of hard to diagnose if the user doesn't use the terminal. Seeking will reenable display-sync, so you can fix playback if case 1 happens, but still get predictable behavior in case 2.
2015-11-27 13:40:52 +00:00
return;
player: add display sync mode If this mode is enabled, the player tries to strictly synchronize video to display refresh. It will adjust playback speed to match the display, so if you play 23.976 fps video on a 24 Hz screen, playback speed is increased by approximately 1/1000. Audio wll be resampled to keep up with playback. This is different from the default sync mode, which will sync video to audio, with the consequence that video might skip or repeat a frame once in a while to make video keep up with audio. This is still unpolished. There are some major problems as well; in particular, mkv VFR files won't work well. The reason is that Matroska is terrible and rounds timestamps to milliseconds. This makes it rather hard to guess the framerate of a section of video that is playing. We could probably fix this by just accepting jittery timestamps (instead of explicitly disabling the sync code in this case), but I'm not ready to accept such a solution yet. Another issue is that we are extremely reliant on OS video and audio APIs working in an expected manner, which of course is not too often the case. Consequently, the new sync mode is a bit fragile.
2015-08-10 16:43:25 +00:00
player: don't make display-sync panic on timestamp discontinuities
2015-12-04 16:07:50 +00:00
double adjusted_duration = MPMAX(0, mpctx->past_frames[0].approx_duration);
player: refactor display-sync frame duration calculations Get rid of get_past_frame_durations(), which was a bit too messy. Add a past_frames array, which contains the same information in a more reasonable way. This also means that we can get the exact current and past frame durations without going through awful stuff. (The main problem is that vo_pts_history contains future frames as well, which is needed for frame backstepping etc., but gets in the way here.) Also disable the automatic disabling of display-sync if the frame duration changes, and extend the frame durations allowed for display sync. To allow arbitrarily high durations, vo.c needs to be changed to pause and potentially redraw OSD while showing a single frame, so they're still limited. In an attempt to deal with VFR, calculate the overall speed using the average FPS. The frame scheduling itself does not use the average FPS, but the duration of the current frame. This does not work too well, but provides a good base for further improvements. Where this commit actually helps a lot is dealing with rounded timestamps, e.g. if the container framerate is wrong or unknown, or if the muxer wrote incorrectly rounded timestamps. While the rounding errors apparently can't be get rid of completely in the general case, this is still much better than e.g. disabling display-sync completely just because some frame durations go out of bounds.
2015-11-13 21:45:40 +00:00
adjusted_duration /= opts->playback_speed;
player: don't make display-sync panic on timestamp discontinuities
2015-12-04 16:07:50 +00:00
if (adjusted_duration > 0.5)
player: always disable display-sync on desyncs Instead of periodically trying to enable it again. There are two cases that can happen: 1. A random discontinuity messed everything up, 2. Things are just broken and will desync all the time Until now, it tried to deal with case 1 - but maybe this is really rare, and we don't really need to care about it. On the other hand, case 2 is kind of hard to diagnose if the user doesn't use the terminal. Seeking will reenable display-sync, so you can fix playback if case 1 happens, but still get predictable behavior in case 2.
2015-11-27 13:40:52 +00:00
return;
player: add display sync mode If this mode is enabled, the player tries to strictly synchronize video to display refresh. It will adjust playback speed to match the display, so if you play 23.976 fps video on a 24 Hz screen, playback speed is increased by approximately 1/1000. Audio wll be resampled to keep up with playback. This is different from the default sync mode, which will sync video to audio, with the consequence that video might skip or repeat a frame once in a while to make video keep up with audio. This is still unpolished. There are some major problems as well; in particular, mkv VFR files won't work well. The reason is that Matroska is terrible and rounds timestamps to milliseconds. This makes it rather hard to guess the framerate of a section of video that is playing. We could probably fix this by just accepting jittery timestamps (instead of explicitly disabling the sync code in this case), but I'm not ready to accept such a solution yet. Another issue is that we are extremely reliant on OS video and audio APIs working in an expected manner, which of course is not too often the case. Consequently, the new sync mode is a bit fragile.
2015-08-10 16:43:25 +00:00
player: make display-vdrop mode do what the manpage claims Don't change video speed in this mode, which is closer to the claim on the manpage that it's close to the behavior of the "audio" mode.
2015-11-26 17:53:32 +00:00
mpctx->speed_factor_v = 1.0;
if (mode != VS_DISP_VDROP) {
double best = find_best_speed(mpctx, vsync);
// If it doesn't work, play at normal speed.
if (fabs(best - 1.0) <= opts->sync_max_video_change / 100)
mpctx->speed_factor_v = best;
}
player: add display sync mode If this mode is enabled, the player tries to strictly synchronize video to display refresh. It will adjust playback speed to match the display, so if you play 23.976 fps video on a 24 Hz screen, playback speed is increased by approximately 1/1000. Audio wll be resampled to keep up with playback. This is different from the default sync mode, which will sync video to audio, with the consequence that video might skip or repeat a frame once in a while to make video keep up with audio. This is still unpolished. There are some major problems as well; in particular, mkv VFR files won't work well. The reason is that Matroska is terrible and rounds timestamps to milliseconds. This makes it rather hard to guess the framerate of a section of video that is playing. We could probably fix this by just accepting jittery timestamps (instead of explicitly disabling the sync code in this case), but I'm not ready to accept such a solution yet. Another issue is that we are extremely reliant on OS video and audio APIs working in an expected manner, which of course is not too often the case. Consequently, the new sync mode is a bit fragile.
2015-08-10 16:43:25 +00:00
double av_diff = mpctx->last_av_difference;
player: always disable display-sync on desyncs Instead of periodically trying to enable it again. There are two cases that can happen: 1. A random discontinuity messed everything up, 2. Things are just broken and will desync all the time Until now, it tried to deal with case 1 - but maybe this is really rare, and we don't really need to care about it. On the other hand, case 2 is kind of hard to diagnose if the user doesn't use the terminal. Seeking will reenable display-sync, so you can fix playback if case 1 happens, but still get predictable behavior in case 2.
2015-11-27 13:40:52 +00:00
if (fabs(av_diff) > 0.5) {
mpctx->display_sync_broken = true;
return;
player: add display sync mode If this mode is enabled, the player tries to strictly synchronize video to display refresh. It will adjust playback speed to match the display, so if you play 23.976 fps video on a 24 Hz screen, playback speed is increased by approximately 1/1000. Audio wll be resampled to keep up with playback. This is different from the default sync mode, which will sync video to audio, with the consequence that video might skip or repeat a frame once in a while to make video keep up with audio. This is still unpolished. There are some major problems as well; in particular, mkv VFR files won't work well. The reason is that Matroska is terrible and rounds timestamps to milliseconds. This makes it rather hard to guess the framerate of a section of video that is playing. We could probably fix this by just accepting jittery timestamps (instead of explicitly disabling the sync code in this case), but I'm not ready to accept such a solution yet. Another issue is that we are extremely reliant on OS video and audio APIs working in an expected manner, which of course is not too often the case. Consequently, the new sync mode is a bit fragile.
2015-08-10 16:43:25 +00:00
}
// Determine for how many vsyncs a frame should be displayed. This can be
// e.g. 2 for 30hz on a 60hz display. It can also be 0 if the video
// framerate is higher than the display framerate.
// We use the speed-adjusted (i.e. real) frame duration for this.
player: move audio speed adjustment code Move it (in a cosmetic sense), and also move its invocation to below all the video handling. All other changes remain cosmetic, including moving the framedrop calculation code, and getting rid of the video_speed_correction variable.
2015-11-04 18:37:15 +00:00
double frame_duration = adjusted_duration / mpctx->speed_factor_v;
player: add display sync mode If this mode is enabled, the player tries to strictly synchronize video to display refresh. It will adjust playback speed to match the display, so if you play 23.976 fps video on a 24 Hz screen, playback speed is increased by approximately 1/1000. Audio wll be resampled to keep up with playback. This is different from the default sync mode, which will sync video to audio, with the consequence that video might skip or repeat a frame once in a while to make video keep up with audio. This is still unpolished. There are some major problems as well; in particular, mkv VFR files won't work well. The reason is that Matroska is terrible and rounds timestamps to milliseconds. This makes it rather hard to guess the framerate of a section of video that is playing. We could probably fix this by just accepting jittery timestamps (instead of explicitly disabling the sync code in this case), but I'm not ready to accept such a solution yet. Another issue is that we are extremely reliant on OS video and audio APIs working in an expected manner, which of course is not too often the case. Consequently, the new sync mode is a bit fragile.
2015-08-10 16:43:25 +00:00
double ratio = (frame_duration + mpctx->display_sync_error) / vsync;
player: less naive rounding
2015-11-11 18:52:35 +00:00
int num_vsyncs = MPMAX(lrint(ratio), 0);
video: fix another A/V difference bug in display-sync mode This didn't show up with cases where the frame pattern has a cycle of 1 or 2 like it is the case with 24-on-24 fps, or 24-on-60 fps. It did show up with 25-on-60 fps. (We don't slow down 25 fps video to 24 on default settings.) In this case, we must not add the timing error of the next frame to the A/V difference estimation of the current frame. Use the previous timing error instead. This is another bug resulting from the confusion about whether we calculate parameters for the currently playing frame, or the one we're about to queue.
2015-10-31 12:19:38 +00:00
double prev_error = mpctx->display_sync_error;
player: add display sync mode If this mode is enabled, the player tries to strictly synchronize video to display refresh. It will adjust playback speed to match the display, so if you play 23.976 fps video on a 24 Hz screen, playback speed is increased by approximately 1/1000. Audio wll be resampled to keep up with playback. This is different from the default sync mode, which will sync video to audio, with the consequence that video might skip or repeat a frame once in a while to make video keep up with audio. This is still unpolished. There are some major problems as well; in particular, mkv VFR files won't work well. The reason is that Matroska is terrible and rounds timestamps to milliseconds. This makes it rather hard to guess the framerate of a section of video that is playing. We could probably fix this by just accepting jittery timestamps (instead of explicitly disabling the sync code in this case), but I'm not ready to accept such a solution yet. Another issue is that we are extremely reliant on OS video and audio APIs working in an expected manner, which of course is not too often the case. Consequently, the new sync mode is a bit fragile.
2015-08-10 16:43:25 +00:00
mpctx->display_sync_error += frame_duration - num_vsyncs * vsync;
player: log some more display-sync information
2015-11-25 21:07:47 +00:00
MP_DBG(mpctx, "s=%f vsyncs=%d dur=%f ratio=%f err=%.20f (%f/%f)\n",
player: move audio speed adjustment code Move it (in a cosmetic sense), and also move its invocation to below all the video handling. All other changes remain cosmetic, including moving the framedrop calculation code, and getting rid of the video_speed_correction variable.
2015-11-04 18:37:15 +00:00
mpctx->speed_factor_v, num_vsyncs, adjusted_duration, ratio,
player: log some more display-sync information
2015-11-25 21:07:47 +00:00
mpctx->display_sync_error, mpctx->display_sync_error / vsync,
mpctx->display_sync_error / frame_duration);
player: add display sync mode If this mode is enabled, the player tries to strictly synchronize video to display refresh. It will adjust playback speed to match the display, so if you play 23.976 fps video on a 24 Hz screen, playback speed is increased by approximately 1/1000. Audio wll be resampled to keep up with playback. This is different from the default sync mode, which will sync video to audio, with the consequence that video might skip or repeat a frame once in a while to make video keep up with audio. This is still unpolished. There are some major problems as well; in particular, mkv VFR files won't work well. The reason is that Matroska is terrible and rounds timestamps to milliseconds. This makes it rather hard to guess the framerate of a section of video that is playing. We could probably fix this by just accepting jittery timestamps (instead of explicitly disabling the sync code in this case), but I'm not ready to accept such a solution yet. Another issue is that we are extremely reliant on OS video and audio APIs working in an expected manner, which of course is not too often the case. Consequently, the new sync mode is a bit fragile.
2015-08-10 16:43:25 +00:00
player: move audio speed adjustment code Move it (in a cosmetic sense), and also move its invocation to below all the video handling. All other changes remain cosmetic, including moving the framedrop calculation code, and getting rid of the video_speed_correction variable.
2015-11-04 18:37:15 +00:00
MP_STATS(mpctx, "value %f avdiff", av_diff);
// Intended number of additional display frames to drop (<0) or repeat (>0)
int drop_repeat = 0;
player: stricter framedrop threshold 80ms allowable desync was a bit too much. It'd allow for a range of 160ms, which everyone can notice. It might also be a bother to apply compensation resampling speed for that long.
2015-11-13 21:53:38 +00:00
// If we are too far ahead/behind, attempt to drop/repeat frames.
player: move audio speed adjustment code Move it (in a cosmetic sense), and also move its invocation to below all the video handling. All other changes remain cosmetic, including moving the framedrop calculation code, and getting rid of the video_speed_correction variable.
2015-11-04 18:37:15 +00:00
// Tolerate some desync to avoid frame dropping due to jitter.
player: stricter framedrop threshold 80ms allowable desync was a bit too much. It'd allow for a range of 160ms, which everyone can notice. It might also be a bother to apply compensation resampling speed for that long.
2015-11-13 21:53:38 +00:00
if (drop && fabs(av_diff) >= 0.020 && fabs(av_diff) / vsync >= 1)
player: move audio speed adjustment code Move it (in a cosmetic sense), and also move its invocation to below all the video handling. All other changes remain cosmetic, including moving the framedrop calculation code, and getting rid of the video_speed_correction variable.
2015-11-04 18:37:15 +00:00
drop_repeat = -av_diff / vsync; // round towards 0
player: add display sync mode If this mode is enabled, the player tries to strictly synchronize video to display refresh. It will adjust playback speed to match the display, so if you play 23.976 fps video on a 24 Hz screen, playback speed is increased by approximately 1/1000. Audio wll be resampled to keep up with playback. This is different from the default sync mode, which will sync video to audio, with the consequence that video might skip or repeat a frame once in a while to make video keep up with audio. This is still unpolished. There are some major problems as well; in particular, mkv VFR files won't work well. The reason is that Matroska is terrible and rounds timestamps to milliseconds. This makes it rather hard to guess the framerate of a section of video that is playing. We could probably fix this by just accepting jittery timestamps (instead of explicitly disabling the sync code in this case), but I'm not ready to accept such a solution yet. Another issue is that we are extremely reliant on OS video and audio APIs working in an expected manner, which of course is not too often the case. Consequently, the new sync mode is a bit fragile.
2015-08-10 16:43:25 +00:00
// We can only drop all frames at most. We can repeat much more frames,
// but we still limit it to 10 times the original frames to avoid that
// corner cases or exceptional situations cause too much havoc.
drop_repeat = MPCLAMP(drop_repeat, -num_vsyncs, num_vsyncs * 10);
num_vsyncs += drop_repeat;
player: always show the first frame in DS mode Fixes bogus frame drop counter in cover art mode.
2016-04-24 18:14:51 +00:00
// Always show the first frame.
if (mpctx->num_past_frames <= 1 && num_vsyncs < 1)
num_vsyncs = 1;
player: add display sync mode If this mode is enabled, the player tries to strictly synchronize video to display refresh. It will adjust playback speed to match the display, so if you play 23.976 fps video on a 24 Hz screen, playback speed is increased by approximately 1/1000. Audio wll be resampled to keep up with playback. This is different from the default sync mode, which will sync video to audio, with the consequence that video might skip or repeat a frame once in a while to make video keep up with audio. This is still unpolished. There are some major problems as well; in particular, mkv VFR files won't work well. The reason is that Matroska is terrible and rounds timestamps to milliseconds. This makes it rather hard to guess the framerate of a section of video that is playing. We could probably fix this by just accepting jittery timestamps (instead of explicitly disabling the sync code in this case), but I'm not ready to accept such a solution yet. Another issue is that we are extremely reliant on OS video and audio APIs working in an expected manner, which of course is not too often the case. Consequently, the new sync mode is a bit fragile.
2015-08-10 16:43:25 +00:00
// Estimate the video position, so we can calculate a good A/V difference
player: minor refactor for A/V diff computation Calculate the A/V difference directly in the display sync code, instead of the awkward current way, which reuses the fields for audio sync. We still set time_frame, because it makes falling back to audio sync somewhat smoother.
2015-10-28 22:50:54 +00:00
// value below. This is used to estimate A/V drift.
player: account for minor VO underruns If the player sends a frame with duration==0 to the VO, it can trivially underrun. Don't panic, but keep the correct time. Also, returning the absolute time from vo_get_next_frame_start_time() just to turn it into a float with relative time was silly. Rename it and make it return what the caller needs.
2015-11-14 20:44:59 +00:00
double time_left = vo_get_delay(vo);
player: add display sync mode If this mode is enabled, the player tries to strictly synchronize video to display refresh. It will adjust playback speed to match the display, so if you play 23.976 fps video on a 24 Hz screen, playback speed is increased by approximately 1/1000. Audio wll be resampled to keep up with playback. This is different from the default sync mode, which will sync video to audio, with the consequence that video might skip or repeat a frame once in a while to make video keep up with audio. This is still unpolished. There are some major problems as well; in particular, mkv VFR files won't work well. The reason is that Matroska is terrible and rounds timestamps to milliseconds. This makes it rather hard to guess the framerate of a section of video that is playing. We could probably fix this by just accepting jittery timestamps (instead of explicitly disabling the sync code in this case), but I'm not ready to accept such a solution yet. Another issue is that we are extremely reliant on OS video and audio APIs working in an expected manner, which of course is not too often the case. Consequently, the new sync mode is a bit fragile.
2015-08-10 16:43:25 +00:00
// We also know that the timing is (necessarily) off, because we have to
// align frame timings on the vsync boundaries. This is unavoidable, and
video: fix another A/V difference bug in display-sync mode This didn't show up with cases where the frame pattern has a cycle of 1 or 2 like it is the case with 24-on-24 fps, or 24-on-60 fps. It did show up with 25-on-60 fps. (We don't slow down 25 fps video to 24 on default settings.) In this case, we must not add the timing error of the next frame to the A/V difference estimation of the current frame. Use the previous timing error instead. This is another bug resulting from the confusion about whether we calculate parameters for the currently playing frame, or the one we're about to queue.
2015-10-31 12:19:38 +00:00
// for the sake of the A/V sync calculations we pretend it's perfect.
time_left += prev_error;
player: fix display sync A/V difference estimation on drops When dropping or repeating frames, we essentially influence when the frame after the next frame will be shown, not the next frame. This led to dropping/repeating frames 2 times, because the A/V difference had a delay of one frame. Compensate it with the expected value.
2015-10-28 22:50:37 +00:00
// Likewise, we know sync is off, but is going to be compensated.
player: minor refactor for A/V diff computation Calculate the A/V difference directly in the display sync code, instead of the awkward current way, which reuses the fields for audio sync. We still set time_frame, because it makes falling back to audio sync somewhat smoother.
2015-10-28 22:50:54 +00:00
time_left += drop_repeat * vsync;
player: fix display-sync timing if audio take long on resume In display-sync mode, the very first video frame is idiotically fully timed, even though audio has not been synced yet at this point, and the video frame is more like a "preview" frame. But since it's fully timed, an underflow is detected if audio takes longer than the display time of the frame (we send the second frame only after audio is done). The timing code will try to compensate for the determined desync, but it really shouldn't. So explicitly discard the timing info in this specific case. On the other hand, if the first frame still hasn't finished display, we can pretend everything is ok. This is a hack - ideally, we either would send a frame without timing info (and then send it again or so when playback starts properly), or we would add real pause support to the VO, and pause it during syncing.
2016-08-07 12:06:54 +00:00
// If syncing took too long, disregard timing of the first frame.
if (mpctx->num_past_frames == 2 && time_left < 0) {
vo_discard_timing_info(vo);
time_left = 0;
}
player: log some more display-sync information
2015-11-25 21:07:47 +00:00
if (drop_repeat) {
command: add mistimed-frame-count property Does what the manpage says. This is a replacement incrementing the dropped frame counter (see previous commit).
2015-10-30 13:05:41 +00:00
mpctx->mistimed_frames_total += 1;
player: log some more display-sync information
2015-11-25 21:07:47 +00:00
MP_STATS(mpctx, "mistimed");
}
command: add mistimed-frame-count property Does what the manpage says. This is a replacement incrementing the dropped frame counter (see previous commit).
2015-10-30 13:05:41 +00:00
player: disable total-avsync-change update in display-sync mode The total-avsync-change property made no sense in display-sync mode (in addition to making not all that much sense in general).
2015-10-27 19:56:58 +00:00
mpctx->total_avsync_change = 0;
player: another fix to A/V difference calculation in display-sync mode update_av_diff() works on the timestamps, while time_left is in real time. When playing at not-1 speed, these are very different, and cause the A/V difference to jitter. Fix this by scaling the expected A/V desync to the correct range.
2015-11-01 18:53:48 +00:00
update_av_diff(mpctx, time_left * opts->playback_speed);
player: minor refactor for A/V diff computation Calculate the A/V difference directly in the display sync code, instead of the awkward current way, which reuses the fields for audio sync. We still set time_frame, because it makes falling back to audio sync somewhat smoother.
2015-10-28 22:50:54 +00:00
command: add vsync-ratio property This is very "illustrative", unlike the video-speed-correction property, and thus useful. It can also be used to observe scheduling errors, which are not detected by the core. (These happen due to rounding errors; possibly not evne our fault, but coming from files with rounded timestamps and so on.)
2015-11-13 21:48:32 +00:00
mpctx->past_frames[0].num_vsyncs = num_vsyncs;
player: try to compensate actual audio drift We always let audio slowly desync until a threshold is reached, and then pushed it back by applying a maximum compensation speed. Refine what comes afterwards: instead of playing with the nominal video speed, use the actual required audio speed for keeping sync as measured by the A/V difference. (The "actual" speed is the ideal speed with A/V differences added.) Although this works in theory, it's somewhat questionable how much this works in practice. The ideal time value is actually not exact, but is the time at which the frame is scheduled (could be compensated by using the time_left calculations in handle_display_sync_frame()). It doesn't account for speed changes or catastrophic discontinuities. It uses only 10 past frames.
2015-11-13 21:51:39 +00:00
mpctx->past_frames[0].av_diff = mpctx->last_av_difference;
command: add vsync-ratio property This is very "illustrative", unlike the video-speed-correction property, and thus useful. It can also be used to observe scheduling errors, which are not detected by the core. (These happen due to rounding errors; possibly not evne our fault, but coming from files with rounded timestamps and so on.)
2015-11-13 21:48:32 +00:00
player: change display-sync audio speed only if needed As long as it's within the desync tolerance, do not change the audio speed at all for resampling. This reduces speed changes which might be caused by jittering timestamps and similar cases. (While in theory you could just not care and change speed every single frame, I'm afraid that such changes could possibly cause audio artifacts. So better just avoid it in the first place.)
2015-11-13 21:50:58 +00:00
if (resample) {
adjust_audio_resample_speed(mpctx, vsync);
} else {
mpctx->speed_factor_a = 1.0;
}
player: move audio speed adjustment code Move it (in a cosmetic sense), and also move its invocation to below all the video handling. All other changes remain cosmetic, including moving the framedrop calculation code, and getting rid of the video_speed_correction variable.
2015-11-04 18:37:15 +00:00
player: minor refactor for A/V diff computation Calculate the A/V difference directly in the display sync code, instead of the awkward current way, which reuses the fields for audio sync. We still set time_frame, because it makes falling back to audio sync somewhat smoother.
2015-10-28 22:50:54 +00:00
// A bad guess, only needed when reverting to audio sync.
mpctx->time_frame = time_left;
player: add display sync mode If this mode is enabled, the player tries to strictly synchronize video to display refresh. It will adjust playback speed to match the display, so if you play 23.976 fps video on a 24 Hz screen, playback speed is increased by approximately 1/1000. Audio wll be resampled to keep up with playback. This is different from the default sync mode, which will sync video to audio, with the consequence that video might skip or repeat a frame once in a while to make video keep up with audio. This is still unpolished. There are some major problems as well; in particular, mkv VFR files won't work well. The reason is that Matroska is terrible and rounds timestamps to milliseconds. This makes it rather hard to guess the framerate of a section of video that is playing. We could probably fix this by just accepting jittery timestamps (instead of explicitly disabling the sync code in this case), but I'm not ready to accept such a solution yet. Another issue is that we are extremely reliant on OS video and audio APIs working in an expected manner, which of course is not too often the case. Consequently, the new sync mode is a bit fragile.
2015-08-10 16:43:25 +00:00
vo: change vo_frame field units This was just converting back and forth between int64_t/microseconds and double/seconds. Remove this stupidity. The pts/duration fields are still in microseconds, but they have no meaning in the display-sync case (also drop printing the pts field from opengl/video.c - it's always 0).
2015-11-27 21:04:44 +00:00
frame->vsync_interval = vsync;
vo_opengl: fix interpolation with display-sync At least I hope so. Deriving the duration from the pts was not really correct. It doesn't include speed adjustments, and becomes completely wrong of the user e.g. changes the playback speed by a huge amount. Pass through the accurate duration value by adding a new vo_frame field. The value for vsync_offset was not correct either. We don't need the error for the next frame, but the error for the current one. This wasn't noticed because it makes no difference in symmetric cases, like 24 fps on 60 Hz. I'm still not entirely confident in the correctness of this, but it sure is an improvement. Also, remove the MP_STATS() calls - they're not really useful to debug anything anymore.
2015-11-28 14:45:35 +00:00
frame->vsync_offset = -prev_error;
frame->ideal_frame_duration = frame_duration;
player: add display sync mode If this mode is enabled, the player tries to strictly synchronize video to display refresh. It will adjust playback speed to match the display, so if you play 23.976 fps video on a 24 Hz screen, playback speed is increased by approximately 1/1000. Audio wll be resampled to keep up with playback. This is different from the default sync mode, which will sync video to audio, with the consequence that video might skip or repeat a frame once in a while to make video keep up with audio. This is still unpolished. There are some major problems as well; in particular, mkv VFR files won't work well. The reason is that Matroska is terrible and rounds timestamps to milliseconds. This makes it rather hard to guess the framerate of a section of video that is playing. We could probably fix this by just accepting jittery timestamps (instead of explicitly disabling the sync code in this case), but I'm not ready to accept such a solution yet. Another issue is that we are extremely reliant on OS video and audio APIs working in an expected manner, which of course is not too often the case. Consequently, the new sync mode is a bit fragile.
2015-08-10 16:43:25 +00:00
frame->num_vsyncs = num_vsyncs;
frame->display_synced = true;
mpctx->display_sync_active = true;
player: change display-sync audio speed only if needed As long as it's within the desync tolerance, do not change the audio speed at all for resampling. This reduces speed changes which might be caused by jittering timestamps and similar cases. (While in theory you could just not care and change speed every single frame, I'm afraid that such changes could possibly cause audio artifacts. So better just avoid it in the first place.)
2015-11-13 21:50:58 +00:00
update_playback_speed(mpctx);
player: add display sync mode If this mode is enabled, the player tries to strictly synchronize video to display refresh. It will adjust playback speed to match the display, so if you play 23.976 fps video on a 24 Hz screen, playback speed is increased by approximately 1/1000. Audio wll be resampled to keep up with playback. This is different from the default sync mode, which will sync video to audio, with the consequence that video might skip or repeat a frame once in a while to make video keep up with audio. This is still unpolished. There are some major problems as well; in particular, mkv VFR files won't work well. The reason is that Matroska is terrible and rounds timestamps to milliseconds. This makes it rather hard to guess the framerate of a section of video that is playing. We could probably fix this by just accepting jittery timestamps (instead of explicitly disabling the sync code in this case), but I'm not ready to accept such a solution yet. Another issue is that we are extremely reliant on OS video and audio APIs working in an expected manner, which of course is not too often the case. Consequently, the new sync mode is a bit fragile.
2015-08-10 16:43:25 +00:00
player: change display-sync audio speed only if needed As long as it's within the desync tolerance, do not change the audio speed at all for resampling. This reduces speed changes which might be caused by jittering timestamps and similar cases. (While in theory you could just not care and change speed every single frame, I'm afraid that such changes could possibly cause audio artifacts. So better just avoid it in the first place.)
2015-11-13 21:50:58 +00:00
MP_STATS(mpctx, "value %f aspeed", mpctx->speed_factor_a - 1);
MP_STATS(mpctx, "value %f vspeed", mpctx->speed_factor_v - 1);
player: always disable display-sync on desyncs Instead of periodically trying to enable it again. There are two cases that can happen: 1. A random discontinuity messed everything up, 2. Things are just broken and will desync all the time Until now, it tried to deal with case 1 - but maybe this is really rare, and we don't really need to care about it. On the other hand, case 2 is kind of hard to diagnose if the user doesn't use the terminal. Seeking will reenable display-sync, so you can fix playback if case 1 happens, but still get predictable behavior in case 2.
2015-11-27 13:40:52 +00:00
}
player: add display sync mode If this mode is enabled, the player tries to strictly synchronize video to display refresh. It will adjust playback speed to match the display, so if you play 23.976 fps video on a 24 Hz screen, playback speed is increased by approximately 1/1000. Audio wll be resampled to keep up with playback. This is different from the default sync mode, which will sync video to audio, with the consequence that video might skip or repeat a frame once in a while to make video keep up with audio. This is still unpolished. There are some major problems as well; in particular, mkv VFR files won't work well. The reason is that Matroska is terrible and rounds timestamps to milliseconds. This makes it rather hard to guess the framerate of a section of video that is playing. We could probably fix this by just accepting jittery timestamps (instead of explicitly disabling the sync code in this case), but I'm not ready to accept such a solution yet. Another issue is that we are extremely reliant on OS video and audio APIs working in an expected manner, which of course is not too often the case. Consequently, the new sync mode is a bit fragile.
2015-08-10 16:43:25 +00:00
player: always disable display-sync on desyncs Instead of periodically trying to enable it again. There are two cases that can happen: 1. A random discontinuity messed everything up, 2. Things are just broken and will desync all the time Until now, it tried to deal with case 1 - but maybe this is really rare, and we don't really need to care about it. On the other hand, case 2 is kind of hard to diagnose if the user doesn't use the terminal. Seeking will reenable display-sync, so you can fix playback if case 1 happens, but still get predictable behavior in case 2.
2015-11-27 13:40:52 +00:00
static void schedule_frame(struct MPContext *mpctx, struct vo_frame *frame)
{
handle_display_sync_frame(mpctx, frame);
player: add display sync mode If this mode is enabled, the player tries to strictly synchronize video to display refresh. It will adjust playback speed to match the display, so if you play 23.976 fps video on a 24 Hz screen, playback speed is increased by approximately 1/1000. Audio wll be resampled to keep up with playback. This is different from the default sync mode, which will sync video to audio, with the consequence that video might skip or repeat a frame once in a while to make video keep up with audio. This is still unpolished. There are some major problems as well; in particular, mkv VFR files won't work well. The reason is that Matroska is terrible and rounds timestamps to milliseconds. This makes it rather hard to guess the framerate of a section of video that is playing. We could probably fix this by just accepting jittery timestamps (instead of explicitly disabling the sync code in this case), but I'm not ready to accept such a solution yet. Another issue is that we are extremely reliant on OS video and audio APIs working in an expected manner, which of course is not too often the case. Consequently, the new sync mode is a bit fragile.
2015-08-10 16:43:25 +00:00
player: remove display_sync_disable_counter We can implement it differently and drop a tiny bit of state.
2015-11-13 21:49:50 +00:00
if (mpctx->num_past_frames > 1 &&
((mpctx->past_frames[1].num_vsyncs >= 0) != mpctx->display_sync_active))
{
player: add display sync mode If this mode is enabled, the player tries to strictly synchronize video to display refresh. It will adjust playback speed to match the display, so if you play 23.976 fps video on a 24 Hz screen, playback speed is increased by approximately 1/1000. Audio wll be resampled to keep up with playback. This is different from the default sync mode, which will sync video to audio, with the consequence that video might skip or repeat a frame once in a while to make video keep up with audio. This is still unpolished. There are some major problems as well; in particular, mkv VFR files won't work well. The reason is that Matroska is terrible and rounds timestamps to milliseconds. This makes it rather hard to guess the framerate of a section of video that is playing. We could probably fix this by just accepting jittery timestamps (instead of explicitly disabling the sync code in this case), but I'm not ready to accept such a solution yet. Another issue is that we are extremely reliant on OS video and audio APIs working in an expected manner, which of course is not too often the case. Consequently, the new sync mode is a bit fragile.
2015-08-10 16:43:25 +00:00
MP_VERBOSE(mpctx, "Video sync mode %s.\n",
mpctx->display_sync_active ? "enabled" : "disabled");
}
player: minor refactor for A/V diff computation Calculate the A/V difference directly in the display sync code, instead of the awkward current way, which reuses the fields for audio sync. We still set time_frame, because it makes falling back to audio sync somewhat smoother.
2015-10-28 22:50:54 +00:00
if (!mpctx->display_sync_active) {
player: change display-sync audio speed only if needed As long as it's within the desync tolerance, do not change the audio speed at all for resampling. This reduces speed changes which might be caused by jittering timestamps and similar cases. (While in theory you could just not care and change speed every single frame, I'm afraid that such changes could possibly cause audio artifacts. So better just avoid it in the first place.)
2015-11-13 21:50:58 +00:00
mpctx->speed_factor_a = 1.0;
mpctx->speed_factor_v = 1.0;
update_playback_speed(mpctx);
player: minor refactor for A/V diff computation Calculate the A/V difference directly in the display sync code, instead of the awkward current way, which reuses the fields for audio sync. We still set time_frame, because it makes falling back to audio sync somewhat smoother.
2015-10-28 22:50:54 +00:00
update_av_diff(mpctx, mpctx->time_frame > 0 ?
mpctx->time_frame * mpctx->video_speed : 0);
}
}
player: refactor display-sync frame duration calculations Get rid of get_past_frame_durations(), which was a bit too messy. Add a past_frames array, which contains the same information in a more reasonable way. This also means that we can get the exact current and past frame durations without going through awful stuff. (The main problem is that vo_pts_history contains future frames as well, which is needed for frame backstepping etc., but gets in the way here.) Also disable the automatic disabling of display-sync if the frame duration changes, and extend the frame durations allowed for display sync. To allow arbitrarily high durations, vo.c needs to be changed to pause and potentially redraw OSD while showing a single frame, so they're still limited. In an attempt to deal with VFR, calculate the overall speed using the average FPS. The frame scheduling itself does not use the average FPS, but the duration of the current frame. This does not work too well, but provides a good base for further improvements. Where this commit actually helps a lot is dealing with rounded timestamps, e.g. if the container framerate is wrong or unknown, or if the muxer wrote incorrectly rounded timestamps. While the rounding errors apparently can't be get rid of completely in the general case, this is still much better than e.g. disabling display-sync completely just because some frame durations go out of bounds.
2015-11-13 21:45:40 +00:00
// Determine the mpctx->past_frames[0] frame duration.
static void calculate_frame_duration(struct MPContext *mpctx)
video: move frame duration code to a separate function Minor preparation for something else.
2015-08-01 20:17:09 +00:00
{
player: refactor display-sync frame duration calculations Get rid of get_past_frame_durations(), which was a bit too messy. Add a past_frames array, which contains the same information in a more reasonable way. This also means that we can get the exact current and past frame durations without going through awful stuff. (The main problem is that vo_pts_history contains future frames as well, which is needed for frame backstepping etc., but gets in the way here.) Also disable the automatic disabling of display-sync if the frame duration changes, and extend the frame durations allowed for display sync. To allow arbitrarily high durations, vo.c needs to be changed to pause and potentially redraw OSD while showing a single frame, so they're still limited. In an attempt to deal with VFR, calculate the overall speed using the average FPS. The frame scheduling itself does not use the average FPS, but the duration of the current frame. This does not work too well, but provides a good base for further improvements. Where this commit actually helps a lot is dealing with rounded timestamps, e.g. if the container framerate is wrong or unknown, or if the muxer wrote incorrectly rounded timestamps. While the rounding errors apparently can't be get rid of completely in the general case, this is still much better than e.g. disabling display-sync completely just because some frame durations go out of bounds.
2015-11-13 21:45:40 +00:00
assert(mpctx->num_past_frames >= 1 && mpctx->num_next_frames >= 1);
video: move frame duration code to a separate function Minor preparation for something else.
2015-08-01 20:17:09 +00:00
player: refactor: eliminate MPContext.d_video Eventually we want the VO be driven by a A->V filter, so a decoder doesn't even have to exist. Some features definitely require a decoder though (like reporting the decoder in use, hardware decoding, etc.), so for each thing which accessed d_video, it has to be redecided if and how it can access decoder state. At least the "framedrop" property slightly changes semantics: you can now always set this property, even if no video is active. Some untested changes in this commit, but our bio-based distributed test suite has to take care of this.
2016-01-17 17:07:50 +00:00
double demux_duration = mpctx->vo_chain->container_fps > 0
? 1.0 / mpctx->vo_chain->container_fps : -1;
video: trust demuxer framerate on invalid timestamps If the PTS goes backwards (whether it's a timestamp reset or some other problem) would just use 0 as frame duration. (At least until the logic for detecting divergence with the timestamps gets active.) Trust the demuxer framerate in these cases instead, if it's available. I think this improves behavior slightly with some broken files.
2016-09-26 15:55:00 +00:00
double duration = demux_duration;
player: refactor display-sync frame duration calculations Get rid of get_past_frame_durations(), which was a bit too messy. Add a past_frames array, which contains the same information in a more reasonable way. This also means that we can get the exact current and past frame durations without going through awful stuff. (The main problem is that vo_pts_history contains future frames as well, which is needed for frame backstepping etc., but gets in the way here.) Also disable the automatic disabling of display-sync if the frame duration changes, and extend the frame durations allowed for display sync. To allow arbitrarily high durations, vo.c needs to be changed to pause and potentially redraw OSD while showing a single frame, so they're still limited. In an attempt to deal with VFR, calculate the overall speed using the average FPS. The frame scheduling itself does not use the average FPS, but the duration of the current frame. This does not work too well, but provides a good base for further improvements. Where this commit actually helps a lot is dealing with rounded timestamps, e.g. if the container framerate is wrong or unknown, or if the muxer wrote incorrectly rounded timestamps. While the rounding errors apparently can't be get rid of completely in the general case, this is still much better than e.g. disabling display-sync completely just because some frame durations go out of bounds.
2015-11-13 21:45:40 +00:00
if (mpctx->num_next_frames >= 2) {
double pts0 = mpctx->next_frames[0]->pts;
double pts1 = mpctx->next_frames[1]->pts;
if (pts0 != MP_NOPTS_VALUE && pts1 != MP_NOPTS_VALUE && pts1 >= pts0)
duration = pts1 - pts0;
video: move frame duration code to a separate function Minor preparation for something else.
2015-08-01 20:17:09 +00:00
}
player: refactor display-sync frame duration calculations Get rid of get_past_frame_durations(), which was a bit too messy. Add a past_frames array, which contains the same information in a more reasonable way. This also means that we can get the exact current and past frame durations without going through awful stuff. (The main problem is that vo_pts_history contains future frames as well, which is needed for frame backstepping etc., but gets in the way here.) Also disable the automatic disabling of display-sync if the frame duration changes, and extend the frame durations allowed for display sync. To allow arbitrarily high durations, vo.c needs to be changed to pause and potentially redraw OSD while showing a single frame, so they're still limited. In an attempt to deal with VFR, calculate the overall speed using the average FPS. The frame scheduling itself does not use the average FPS, but the duration of the current frame. This does not work too well, but provides a good base for further improvements. Where this commit actually helps a lot is dealing with rounded timestamps, e.g. if the container framerate is wrong or unknown, or if the muxer wrote incorrectly rounded timestamps. While the rounding errors apparently can't be get rid of completely in the general case, this is still much better than e.g. disabling display-sync completely just because some frame durations go out of bounds.
2015-11-13 21:45:40 +00:00
// The following code tries to compensate for rounded Matroska timestamps
// by "unrounding" frame durations, or if not possible, approximating them.
player: slightly adjust framerate guessing heuristic Some files not only use rounded timestamps, but they also do it incorrectly. They may jitter between up to 4 specific frame durations. In this case, I found a file that mostly used 41ms and 42ms, but also had 40ms and 43ms outliers (often but not always following each other). This breaks the assumption of the framerate estimation code that the frame duration can deviate up to 1ms. If it jitters around 4 possible frame durations, the maximum deviation is 3ms. Increase it accordingly. The change might make playback of "true VFR" video via display-sync mode worse, but it's not like it was particularly good in the first place. Also, the check whether to usen the container FPS needs to be stricter. In the worst case, num_dur is 1, which doesn't really indicate any evidence that the framerate is correct. Only if there are "enough" frames the deviation check will become meaningful. 16 is an arbitrary value that has been designated "enough" by myself. Also otuput the frame duration values for --dump-stats.
2016-08-29 15:15:06 +00:00
// These formats usually round on 1ms. Some muxers do this incorrectly,
// and might go off by 1ms more, and compensate for it later by an equal
// rounding error into the opposite direction.
double tolerance = 0.001 * 3 + 0.0001;
player: refactor display-sync frame duration calculations Get rid of get_past_frame_durations(), which was a bit too messy. Add a past_frames array, which contains the same information in a more reasonable way. This also means that we can get the exact current and past frame durations without going through awful stuff. (The main problem is that vo_pts_history contains future frames as well, which is needed for frame backstepping etc., but gets in the way here.) Also disable the automatic disabling of display-sync if the frame duration changes, and extend the frame durations allowed for display sync. To allow arbitrarily high durations, vo.c needs to be changed to pause and potentially redraw OSD while showing a single frame, so they're still limited. In an attempt to deal with VFR, calculate the overall speed using the average FPS. The frame scheduling itself does not use the average FPS, but the duration of the current frame. This does not work too well, but provides a good base for further improvements. Where this commit actually helps a lot is dealing with rounded timestamps, e.g. if the container framerate is wrong or unknown, or if the muxer wrote incorrectly rounded timestamps. While the rounding errors apparently can't be get rid of completely in the general case, this is still much better than e.g. disabling display-sync completely just because some frame durations go out of bounds.
2015-11-13 21:45:40 +00:00
double total = 0;
int num_dur = 0;
for (int n = 1; n < mpctx->num_past_frames; n++) {
// Eliminate likely outliers using a really dumb heuristic.
double dur = mpctx->past_frames[n].duration;
if (dur <= 0 || fabs(dur - duration) >= tolerance)
break;
total += dur;
num_dur += 1;
}
player: smooth out frame durations by averaging them Helps somewhat with muxer-rounded timestamps. There is some danger that this introduces a timestamp drift. But since they are averaged values (unlike as when using an incorrect container framerate hint), any potential drift shouldn't be too brutal, or compensate itself soon. So I won't bother yet with comparing the results with the real timestamp, unless we run into actual problems. Of course we still prefer potentially real timestamps over the approximated ones. But unless the timestamps match the container FPS, we can't know whether they are (no, checking whether the they have microsecond components would be cheating). Perhaps in future, we could let the demuxer export the timebase - if the timebase is not 1000 (or divisible by it), we know that millisecond-rounded timestamps won't happen.
2015-11-13 21:46:55 +00:00
double approx_duration = num_dur > 0 ? total / num_dur : duration;
player: refactor display-sync frame duration calculations Get rid of get_past_frame_durations(), which was a bit too messy. Add a past_frames array, which contains the same information in a more reasonable way. This also means that we can get the exact current and past frame durations without going through awful stuff. (The main problem is that vo_pts_history contains future frames as well, which is needed for frame backstepping etc., but gets in the way here.) Also disable the automatic disabling of display-sync if the frame duration changes, and extend the frame durations allowed for display sync. To allow arbitrarily high durations, vo.c needs to be changed to pause and potentially redraw OSD while showing a single frame, so they're still limited. In an attempt to deal with VFR, calculate the overall speed using the average FPS. The frame scheduling itself does not use the average FPS, but the duration of the current frame. This does not work too well, but provides a good base for further improvements. Where this commit actually helps a lot is dealing with rounded timestamps, e.g. if the container framerate is wrong or unknown, or if the muxer wrote incorrectly rounded timestamps. While the rounding errors apparently can't be get rid of completely in the general case, this is still much better than e.g. disabling display-sync completely just because some frame durations go out of bounds.
2015-11-13 21:45:40 +00:00
// Try if the demuxer frame rate fits - if so, just take it.
if (demux_duration > 0) {
// Note that even if each timestamp is within rounding tolerance, it
// could literally not add up (e.g. if demuxer FPS is rounded itself).
if (fabs(duration - demux_duration) < tolerance &&
player: slightly adjust framerate guessing heuristic Some files not only use rounded timestamps, but they also do it incorrectly. They may jitter between up to 4 specific frame durations. In this case, I found a file that mostly used 41ms and 42ms, but also had 40ms and 43ms outliers (often but not always following each other). This breaks the assumption of the framerate estimation code that the frame duration can deviate up to 1ms. If it jitters around 4 possible frame durations, the maximum deviation is 3ms. Increase it accordingly. The change might make playback of "true VFR" video via display-sync mode worse, but it's not like it was particularly good in the first place. Also, the check whether to usen the container FPS needs to be stricter. In the worst case, num_dur is 1, which doesn't really indicate any evidence that the framerate is correct. Only if there are "enough" frames the deviation check will become meaningful. 16 is an arbitrary value that has been designated "enough" by myself. Also otuput the frame duration values for --dump-stats.
2016-08-29 15:15:06 +00:00
fabs(total - demux_duration * num_dur) < tolerance && num_dur >= 16)
player: refactor display-sync frame duration calculations Get rid of get_past_frame_durations(), which was a bit too messy. Add a past_frames array, which contains the same information in a more reasonable way. This also means that we can get the exact current and past frame durations without going through awful stuff. (The main problem is that vo_pts_history contains future frames as well, which is needed for frame backstepping etc., but gets in the way here.) Also disable the automatic disabling of display-sync if the frame duration changes, and extend the frame durations allowed for display sync. To allow arbitrarily high durations, vo.c needs to be changed to pause and potentially redraw OSD while showing a single frame, so they're still limited. In an attempt to deal with VFR, calculate the overall speed using the average FPS. The frame scheduling itself does not use the average FPS, but the duration of the current frame. This does not work too well, but provides a good base for further improvements. Where this commit actually helps a lot is dealing with rounded timestamps, e.g. if the container framerate is wrong or unknown, or if the muxer wrote incorrectly rounded timestamps. While the rounding errors apparently can't be get rid of completely in the general case, this is still much better than e.g. disabling display-sync completely just because some frame durations go out of bounds.
2015-11-13 21:45:40 +00:00
{
approx_duration = demux_duration;
}
}
mpctx->past_frames[0].duration = duration;
mpctx->past_frames[0].approx_duration = approx_duration;
player: slightly adjust framerate guessing heuristic Some files not only use rounded timestamps, but they also do it incorrectly. They may jitter between up to 4 specific frame durations. In this case, I found a file that mostly used 41ms and 42ms, but also had 40ms and 43ms outliers (often but not always following each other). This breaks the assumption of the framerate estimation code that the frame duration can deviate up to 1ms. If it jitters around 4 possible frame durations, the maximum deviation is 3ms. Increase it accordingly. The change might make playback of "true VFR" video via display-sync mode worse, but it's not like it was particularly good in the first place. Also, the check whether to usen the container FPS needs to be stricter. In the worst case, num_dur is 1, which doesn't really indicate any evidence that the framerate is correct. Only if there are "enough" frames the deviation check will become meaningful. 16 is an arbitrary value that has been designated "enough" by myself. Also otuput the frame duration values for --dump-stats.
2016-08-29 15:15:06 +00:00
MP_STATS(mpctx, "value %f frame-duration", MPMAX(0, duration));
MP_STATS(mpctx, "value %f frame-duration-approx", MPMAX(0, approx_duration));
video: move frame duration code to a separate function Minor preparation for something else.
2015-08-01 20:17:09 +00:00
}
player: minor simplification No need to pass endpts down in such a dumb way. Also remove an outdated comment somewhere.
2016-02-27 19:02:51 +00:00
void write_video(struct MPContext *mpctx)
player: move video display code out of the playloop Basically move the code from playloop.c to video.c. The new function write_video() now contains the code that was part of run_playloop(). There are no functional changes, except handling "new_frame_shown" slightly differently. This is done so that we don't need new a new MPContext field or a return value for write_video() to signal this condition. Instead, it's handled indirectly.
2014-07-30 21:24:08 +00:00
{
struct MPOpts *opts = mpctx->opts;
player: refactor: eliminate MPContext.d_video Eventually we want the VO be driven by a A->V filter, so a decoder doesn't even have to exist. Some features definitely require a decoder though (like reporting the decoder in use, hardware decoding, etc.), so for each thing which accessed d_video, it has to be redecided if and how it can access decoder state. At least the "framedrop" property slightly changes semantics: you can now always set this property, even if no video is active. Some untested changes in this commit, but our bio-based distributed test suite has to take care of this.
2016-01-17 17:07:50 +00:00
if (!mpctx->vo_chain)
player: move video display code out of the playloop Basically move the code from playloop.c to video.c. The new function write_video() now contains the code that was part of run_playloop(). There are no functional changes, except handling "new_frame_shown" slightly differently. This is done so that we don't need new a new MPContext field or a return value for write_video() to signal this condition. Instead, it's handled indirectly.
2014-07-30 21:24:08 +00:00
return;
player: remove some further current_track dependencies Now it's used for initialization only for audio and video.
2016-02-05 22:41:44 +00:00
struct track *track = mpctx->vo_chain->track;
player: refactor: eliminate MPContext.d_video Eventually we want the VO be driven by a A->V filter, so a decoder doesn't even have to exist. Some features definitely require a decoder though (like reporting the decoder in use, hardware decoding, etc.), so for each thing which accessed d_video, it has to be redecided if and how it can access decoder state. At least the "framedrop" property slightly changes semantics: you can now always set this property, even if no video is active. Some untested changes in this commit, but our bio-based distributed test suite has to take care of this.
2016-01-17 17:07:50 +00:00
struct vo *vo = mpctx->vo_chain->vo;
player: move video display code out of the playloop Basically move the code from playloop.c to video.c. The new function write_video() now contains the code that was part of run_playloop(). There are no functional changes, except handling "new_frame_shown" slightly differently. This is done so that we don't need new a new MPContext field or a return value for write_video() to signal this condition. Instead, it's handled indirectly.
2014-07-30 21:24:08 +00:00
video: exit early when nothing to do These cases were probably confusing. Exit early, which makes it much clearer what's going on. Should not change anything functionally.
2014-08-13 19:51:26 +00:00
// Actual playback starts when both audio and video are ready.
if (mpctx->video_status == STATUS_READY)
return;
if (mpctx->paused && mpctx->video_status >= STATUS_READY)
return;
player: minor simplification No need to pass endpts down in such a dumb way. Also remove an outdated comment somewhere.
2016-02-27 19:02:51 +00:00
int r = video_output_image(mpctx);
video: refactor queue handling This simplifies the code, and fixes an odd bug: the second-last frame was displayed for a very short duration if framedrop was enabled. The reason was that basically the time difference between second-last and last frame were skipped, because at this point EOF was already signaled. Also see commit b0959488 for a similar issue in the same code. This removes the messiness of the next_frame 2-frame queue, and strictly runs the "new frame" code when a frame is moved to the first position of the queue, instead of somehow messing with return codes. This also merges update_video() into video_output_image().
2014-08-22 12:20:30 +00:00
MP_TRACE(mpctx, "video_output_image: %d\n", r);
player: move video display code out of the playloop Basically move the code from playloop.c to video.c. The new function write_video() now contains the code that was part of run_playloop(). There are no functional changes, except handling "new_frame_shown" slightly differently. This is done so that we don't need new a new MPContext field or a return value for write_video() to signal this condition. Instead, it's handled indirectly.
2014-07-30 21:24:08 +00:00
video: fix and simplify video format changes and last frame display The previous commit broke these things, and fixing them is separate in this commit in order to reduce the volume of changes. Move the image queue from the VO to the playback core. The image queue is a remnant of the old way how vdpau was implemented, and increasingly became more and more an artifact. In the end, it did only one thing: computing the duration of the current frame. This was done by taking the PTS difference between the current and the future frame. We keep this, but by moving it out of the VO, we don't have to special-case format changes anymore. This simplifies the code a lot. Since we need the queue to compute the duration only, a queue size larger than 2 makes no sense, and we can hardcode that. Also change how the last frame is handled. The last frame is a bit of a problem, because video timing works by showing one frame after another, which makes it a special case. Make the VO provide a function to notify us when the frame is done, instead. The frame duration is used for that. This is not perfect. For example, changing playback speed during the last frame doesn't update the end time. Pausing will not stop the clock that times the last frame. But I don't think this matters for such a corner case.
2014-08-12 21:17:35 +00:00
if (r < 0)
goto error;
player: move video display code out of the playloop Basically move the code from playloop.c to video.c. The new function write_video() now contains the code that was part of run_playloop(). There are no functional changes, except handling "new_frame_shown" slightly differently. This is done so that we don't need new a new MPContext field or a return value for write_video() to signal this condition. Instead, it's handled indirectly.
2014-07-30 21:24:08 +00:00
video: minor simplification This is mostly equivalent, but simpler, and reduces some duplication.
2014-08-22 12:18:40 +00:00
if (r == VD_WAIT) // Demuxer will wake us up for more packets to decode.
video: fix and simplify video format changes and last frame display The previous commit broke these things, and fixing them is separate in this commit in order to reduce the volume of changes. Move the image queue from the VO to the playback core. The image queue is a remnant of the old way how vdpau was implemented, and increasingly became more and more an artifact. In the end, it did only one thing: computing the duration of the current frame. This was done by taking the PTS difference between the current and the future frame. We keep this, but by moving it out of the VO, we don't have to special-case format changes anymore. This simplifies the code a lot. Since we need the queue to compute the duration only, a queue size larger than 2 makes no sense, and we can hardcode that. Also change how the last frame is handled. The last frame is a bit of a problem, because video timing works by showing one frame after another, which makes it a special case. Make the VO provide a function to notify us when the frame is done, instead. The frame duration is used for that. This is not perfect. For example, changing playback speed during the last frame doesn't update the end time. Pausing will not stop the clock that times the last frame. But I don't think this matters for such a corner case.
2014-08-12 21:17:35 +00:00
return;
player: move video display code out of the playloop Basically move the code from playloop.c to video.c. The new function write_video() now contains the code that was part of run_playloop(). There are no functional changes, except handling "new_frame_shown" slightly differently. This is done so that we don't need new a new MPContext field or a return value for write_video() to signal this condition. Instead, it's handled indirectly.
2014-07-30 21:24:08 +00:00
video: minor simplification This is mostly equivalent, but simpler, and reduces some duplication.
2014-08-22 12:18:40 +00:00
if (r == VD_EOF) {
player: move video display code out of the playloop Basically move the code from playloop.c to video.c. The new function write_video() now contains the code that was part of run_playloop(). There are no functional changes, except handling "new_frame_shown" slightly differently. This is done so that we don't need new a new MPContext field or a return value for write_video() to signal this condition. Instead, it's handled indirectly.
2014-07-30 21:24:08 +00:00
mpctx->delay = 0;
mpctx->last_av_difference = 0;
player: add option to control duration of image display The --image-display-duration option controls how long an image is displayed. It's also possible to display the image forever (until manual user interaction stops playback). With this, the core drops the old method to "drain" video (i.e. waiting for the last frame duration on end of playback). Instead, we reuse MPContext.time_frame. The old mechanism was disabled for non-images anyway. Fixes #3425.
2016-08-17 20:45:44 +00:00
if (mpctx->video_status <= STATUS_PLAYING) {
mpctx->video_status = STATUS_DRAINING;
get_relative_time(mpctx);
if (mpctx->num_past_frames == 1 && mpctx->past_frames[0].pts == 0 &&
!mpctx->ao_chain)
{
player: log if video is considered an image It's a heristic that can fail, so better log it.
2016-08-21 10:27:48 +00:00
MP_VERBOSE(mpctx, "assuming this is an image\n");
player: add option to control duration of image display The --image-display-duration option controls how long an image is displayed. It's also possible to display the image forever (until manual user interaction stops playback). With this, the core drops the old method to "drain" video (i.e. waiting for the last frame duration on end of playback). Instead, we reuse MPContext.time_frame. The old mechanism was disabled for non-images anyway. Fixes #3425.
2016-08-17 20:45:44 +00:00
mpctx->time_frame += opts->image_display_duration;
video: use demuxer-signaled duration for last video frame Helps with gif, probably does unwanted things with other formats. This doesn't handle --end quite correctly, but this could be added later. Fixes #3924.
2016-12-21 17:18:24 +00:00
} else if (mpctx->last_frame_duration > 0) {
MP_VERBOSE(mpctx, "using demuxer frame duration for last frame\n");
mpctx->time_frame += mpctx->last_frame_duration;
player: add option to control duration of image display The --image-display-duration option controls how long an image is displayed. It's also possible to display the image forever (until manual user interaction stops playback). With this, the core drops the old method to "drain" video (i.e. waiting for the last frame duration on end of playback). Instead, we reuse MPContext.time_frame. The old mechanism was disabled for non-images anyway. Fixes #3425.
2016-08-17 20:45:44 +00:00
} else {
mpctx->time_frame = 0;
}
}
if (mpctx->video_status == STATUS_DRAINING) {
mpctx->time_frame -= get_relative_time(mpctx);
player: litter code with explicit wakeup calls This does 3 kinds of changes: - change sleeptime=x to mp_set_timeout() - change sleeptime=0 to mp_wakeup_core() calls (to be more explicit) - change commands etc. to call mp_wakeup_core() if they do changes that require the playloop to be rerun This is preparation for the following changes. The goal is to process client API requests without having to rerun the playloop every time. As of this commit, the changes should not change behavior. In particular, the playloop is still implicitly woken up on every command.
2016-09-16 12:24:15 +00:00
mp_set_timeout(mpctx, mpctx->time_frame);
player: add option to control duration of image display The --image-display-duration option controls how long an image is displayed. It's also possible to display the image forever (until manual user interaction stops playback). With this, the core drops the old method to "drain" video (i.e. waiting for the last frame duration on end of playback). Instead, we reuse MPContext.time_frame. The old mechanism was disabled for non-images anyway. Fixes #3425.
2016-08-17 20:45:44 +00:00
if (mpctx->time_frame <= 0) {
MP_VERBOSE(mpctx, "video EOF reached\n");
mpctx->video_status = STATUS_EOF;
}
}
player: silence spam in verbose mode when playing audio with cover art When playing cover art, it conceptually reaches EOF as soon as the image was put on the VO, causing the EOF message to be repeated every time new audio was decoded. Just silence the message.
2015-04-14 12:36:15 +00:00
MP_DBG(mpctx, "video EOF (status=%d)\n", mpctx->video_status);
video: fix attached picture mode Playing audio files with embedded cover art broke due to some of the recent changes. Treat video EOF properly, and don't burn the CPU. Disable hrseek for video in attached picture mode, since the decoder will always produce a new image, which makes hrseek never terminate. Fixes #970.
2014-07-31 19:57:05 +00:00
return;
player: move video display code out of the playloop Basically move the code from playloop.c to video.c. The new function write_video() now contains the code that was part of run_playloop(). There are no functional changes, except handling "new_frame_shown" slightly differently. This is done so that we don't need new a new MPContext field or a return value for write_video() to signal this condition. Instead, it's handled indirectly.
2014-07-30 21:24:08 +00:00
}
video: minor simplification This is mostly equivalent, but simpler, and reduces some duplication.
2014-08-22 12:18:40 +00:00
if (mpctx->video_status > STATUS_PLAYING)
mpctx->video_status = STATUS_PLAYING;
if (r != VD_NEW_FRAME) {
player: litter code with explicit wakeup calls This does 3 kinds of changes: - change sleeptime=x to mp_set_timeout() - change sleeptime=0 to mp_wakeup_core() calls (to be more explicit) - change commands etc. to call mp_wakeup_core() if they do changes that require the playloop to be rerun This is preparation for the following changes. The goal is to process client API requests without having to rerun the playloop every time. As of this commit, the changes should not change behavior. In particular, the playloop is still implicitly woken up on every command.
2016-09-16 12:24:15 +00:00
mp_wakeup_core(mpctx); // Decode more in next iteration.
player: move video display code out of the playloop Basically move the code from playloop.c to video.c. The new function write_video() now contains the code that was part of run_playloop(). There are no functional changes, except handling "new_frame_shown" slightly differently. This is done so that we don't need new a new MPContext field or a return value for write_video() to signal this condition. Instead, it's handled indirectly.
2014-07-30 21:24:08 +00:00
return;
video: minor simplification This is mostly equivalent, but simpler, and reduces some duplication.
2014-08-22 12:18:40 +00:00
}
player: move video display code out of the playloop Basically move the code from playloop.c to video.c. The new function write_video() now contains the code that was part of run_playloop(). There are no functional changes, except handling "new_frame_shown" slightly differently. This is done so that we don't need new a new MPContext field or a return value for write_video() to signal this condition. Instead, it's handled indirectly.
2014-07-30 21:24:08 +00:00
video: fix and simplify video format changes and last frame display The previous commit broke these things, and fixing them is separate in this commit in order to reduce the volume of changes. Move the image queue from the VO to the playback core. The image queue is a remnant of the old way how vdpau was implemented, and increasingly became more and more an artifact. In the end, it did only one thing: computing the duration of the current frame. This was done by taking the PTS difference between the current and the future frame. We keep this, but by moving it out of the VO, we don't have to special-case format changes anymore. This simplifies the code a lot. Since we need the queue to compute the duration only, a queue size larger than 2 makes no sense, and we can hardcode that. Also change how the last frame is handled. The last frame is a bit of a problem, because video timing works by showing one frame after another, which makes it a special case. Make the VO provide a function to notify us when the frame is done, instead. The frame duration is used for that. This is not perfect. For example, changing playback speed during the last frame doesn't update the end time. Pausing will not stop the clock that times the last frame. But I don't think this matters for such a corner case.
2014-08-12 21:17:35 +00:00
// Filter output is different from VO input?
video: pass future frames to VO Now the VO can request a number of future frames with the last parameter of vo_set_queue_params(). This will be helpful to fix the interpolation code. Note that the first frame (after playback start or seeking) will usually not have any future frames (to make seeking fast). Near the end of the file, the number of future frames will become lower as well.
2015-07-01 17:22:40 +00:00
struct mp_image_params p = mpctx->next_frames[0]->params;
video: fix and simplify video format changes and last frame display The previous commit broke these things, and fixing them is separate in this commit in order to reduce the volume of changes. Move the image queue from the VO to the playback core. The image queue is a remnant of the old way how vdpau was implemented, and increasingly became more and more an artifact. In the end, it did only one thing: computing the duration of the current frame. This was done by taking the PTS difference between the current and the future frame. We keep this, but by moving it out of the VO, we don't have to special-case format changes anymore. This simplifies the code a lot. Since we need the queue to compute the duration only, a queue size larger than 2 makes no sense, and we can hardcode that. Also change how the last frame is handled. The last frame is a bit of a problem, because video timing works by showing one frame after another, which makes it a special case. Make the VO provide a function to notify us when the frame is done, instead. The frame duration is used for that. This is not perfect. For example, changing playback speed during the last frame doesn't update the end time. Pausing will not stop the clock that times the last frame. But I don't think this matters for such a corner case.
2014-08-12 21:17:35 +00:00
if (!vo->params || !mp_image_params_equal(&p, vo->params)) {
// Changing config deletes the current frame; wait until it's finished.
player: respect --untimed on last frame The last video frame is another case that has a separate code path, although it's pretty similar to the one in commit 73e5aa87. Fix this in a different way, which also takes care of the last frame case, although without context the code becomes slightly more tricky. As further cleanup, move the decision about framedropping itself to the same place, so the check in vo.c becomes much simpler. The check for the vo->driver->encode flag, which is remvoed completely, was redundant too. Fixes #1480.
2015-01-16 12:45:43 +00:00
if (vo_still_displaying(vo))
video: fix and simplify video format changes and last frame display The previous commit broke these things, and fixing them is separate in this commit in order to reduce the volume of changes. Move the image queue from the VO to the playback core. The image queue is a remnant of the old way how vdpau was implemented, and increasingly became more and more an artifact. In the end, it did only one thing: computing the duration of the current frame. This was done by taking the PTS difference between the current and the future frame. We keep this, but by moving it out of the VO, we don't have to special-case format changes anymore. This simplifies the code a lot. Since we need the queue to compute the duration only, a queue size larger than 2 makes no sense, and we can hardcode that. Also change how the last frame is handled. The last frame is a bit of a problem, because video timing works by showing one frame after another, which makes it a special case. Make the VO provide a function to notify us when the frame is done, instead. The frame duration is used for that. This is not perfect. For example, changing playback speed during the last frame doesn't update the end time. Pausing will not stop the clock that times the last frame. But I don't think this matters for such a corner case.
2014-08-12 21:17:35 +00:00
return;
const struct vo_driver *info = mpctx->video_out->driver;
player: print anamorphic size only if video is anamorphic Has been annoying me since forever.
2014-11-02 16:34:02 +00:00
char extra[20] = {0};
video: switch from using display aspect to sample aspect MPlayer traditionally always used the display aspect ratio, e.g. 16:9, while FFmpeg uses the sample (aka pixel) aspect ratio. Both have a bunch of advantages and disadvantages. Actually, it seems using sample aspect ratio is generally nicer. The main reason for the change is making mpv closer to how FFmpeg works in order to make life easier. It's also nice that everything uses integer fractions instead of floats now (except --video-aspect option/property). Note that there is at least 1 user-visible change: vf_dsize now does not set the display size, only the display aspect ratio. This is because the image_params d_w/d_h fields did not just set the display aspect, but also the size (except in encoding mode).
2015-12-19 19:04:31 +00:00
if (p.p_w != p.p_h) {
int d_w, d_h;
mp_image_params_get_dsize(&p, &d_w, &d_h);
snprintf(extra, sizeof(extra), " => %dx%d", d_w, d_h);
}
player: print hw format on "VO: " line too Useful for distinguishing bit depth when hardware decoding. (To the degree it's useful to show it at all. This just brings the hardware decoding case on the same level of showing information as the software decode call.)
2017-01-29 11:53:58 +00:00
char sfmt[20] = {0};
if (p.hw_subfmt)
snprintf(sfmt, sizeof(sfmt), "[%s]", mp_imgfmt_to_name(p.hw_subfmt));
MP_INFO(mpctx, "VO: [%s] %dx%d%s %s%s\n",
info->name, p.w, p.h, extra, mp_imgfmt_to_name(p.imgfmt), sfmt);
video: fix and simplify video format changes and last frame display The previous commit broke these things, and fixing them is separate in this commit in order to reduce the volume of changes. Move the image queue from the VO to the playback core. The image queue is a remnant of the old way how vdpau was implemented, and increasingly became more and more an artifact. In the end, it did only one thing: computing the duration of the current frame. This was done by taking the PTS difference between the current and the future frame. We keep this, but by moving it out of the VO, we don't have to special-case format changes anymore. This simplifies the code a lot. Since we need the queue to compute the duration only, a queue size larger than 2 makes no sense, and we can hardcode that. Also change how the last frame is handled. The last frame is a bit of a problem, because video timing works by showing one frame after another, which makes it a special case. Make the VO provide a function to notify us when the frame is done, instead. The frame duration is used for that. This is not perfect. For example, changing playback speed during the last frame doesn't update the end time. Pausing will not stop the clock that times the last frame. But I don't think this matters for such a corner case.
2014-08-12 21:17:35 +00:00
MP_VERBOSE(mpctx, "VO: Description: %s\n", info->description);
video/out: remove an unused parameter This parameter has been unused for years (the last flag was removed in commit d658b115). Get rid of it. This affects the general VO API, as well as the vo_opengl backend API, so it touches a lot of files. The VOFLAGs are still used to control OpenGL context creation, so move them to the OpenGL backend code.
2015-10-03 16:20:16 +00:00
int vo_r = vo_reconfig(vo, &p);
client API: better error reporting Give somewhat more information on playback failure.
2014-10-28 15:19:07 +00:00
if (vo_r < 0) {
mpctx->error_playing = MPV_ERROR_VO_INIT_FAILED;
video: fix and simplify video format changes and last frame display The previous commit broke these things, and fixing them is separate in this commit in order to reduce the volume of changes. Move the image queue from the VO to the playback core. The image queue is a remnant of the old way how vdpau was implemented, and increasingly became more and more an artifact. In the end, it did only one thing: computing the duration of the current frame. This was done by taking the PTS difference between the current and the future frame. We keep this, but by moving it out of the VO, we don't have to special-case format changes anymore. This simplifies the code a lot. Since we need the queue to compute the duration only, a queue size larger than 2 makes no sense, and we can hardcode that. Also change how the last frame is handled. The last frame is a bit of a problem, because video timing works by showing one frame after another, which makes it a special case. Make the VO provide a function to notify us when the frame is done, instead. The frame duration is used for that. This is not perfect. For example, changing playback speed during the last frame doesn't update the end time. Pausing will not stop the clock that times the last frame. But I don't think this matters for such a corner case.
2014-08-12 21:17:35 +00:00
goto error;
client API: better error reporting Give somewhat more information on playback failure.
2014-10-28 15:19:07 +00:00
}
video: fix and simplify video format changes and last frame display The previous commit broke these things, and fixing them is separate in this commit in order to reduce the volume of changes. Move the image queue from the VO to the playback core. The image queue is a remnant of the old way how vdpau was implemented, and increasingly became more and more an artifact. In the end, it did only one thing: computing the duration of the current frame. This was done by taking the PTS difference between the current and the future frame. We keep this, but by moving it out of the VO, we don't have to special-case format changes anymore. This simplifies the code a lot. Since we need the queue to compute the duration only, a queue size larger than 2 makes no sense, and we can hardcode that. Also change how the last frame is handled. The last frame is a bit of a problem, because video timing works by showing one frame after another, which makes it a special case. Make the VO provide a function to notify us when the frame is done, instead. The frame duration is used for that. This is not perfect. For example, changing playback speed during the last frame doesn't update the end time. Pausing will not stop the clock that times the last frame. But I don't think this matters for such a corner case.
2014-08-12 21:17:35 +00:00
init_vo(mpctx);
}
player: rearrange some A/V-sync related code Move the update_avsync_before_frame() call further down. Moving it closer to where the time_frame value is used (and which the function updates) should make the code more readable. With this change, there's no need anymore to reset the time_frame value on the video reconfig path. Move the update_avsync_after_frame() up. Now no meaningful amount of time passes since the previous get_relative_time() call anymore, and the second one can be removed.
2015-01-30 20:18:30 +00:00
mpctx->time_frame -= get_relative_time(mpctx);
update_avsync_before_frame(mpctx);
player: make sure non-video subtitle rendering is reset if video resumes If video reaches EOF, subtitle timing will be switched to timing without video frames. This means it calls osd_set_force_video_pts() and overrides the PTS of whatever video frame is current (since the video frame's PTS has nothing to do with the current playback position anymore). This was not reset when seeking back into video. Subtitles wouldn't show up, or if there was a subtitle displayed, it would get stuck with it. In particular, this could happen even if EOF was only temporary (such as with --keep-open). Fix this by clearing the override PTS whenever a video frame is shown. Fixes #3770.
2016-11-18 12:01:46 +00:00
// Enforce timing subtitles to video frames.
osd_set_force_video_pts(mpctx->osd, MP_NOPTS_VALUE);
sub: change how subtitles are read Slightly change how it is decided when a new packet should be read. Switch to demux_read_packet_async(), and let the player "wait properly" until required subtitle packets arrive, instead of blocking everything. Move distinguishing the cases of passive and active reading into the demuxer, where it belongs.
2015-12-29 00:35:52 +00:00
if (!update_subtitles(mpctx, mpctx->next_frames[0]->pts)) {
video: fix debug message Should not be a warning, and the message text was also very useless.
2016-01-06 18:48:55 +00:00
MP_VERBOSE(mpctx, "Video frame delayed due waiting on subtitles.\n");
sub: change how subtitles are read Slightly change how it is decided when a new packet should be read. Switch to demux_read_packet_async(), and let the player "wait properly" until required subtitle packets arrive, instead of blocking everything. Move distinguishing the cases of passive and active reading into the demuxer, where it belongs.
2015-12-29 00:35:52 +00:00
return;
}
video: move display and timing to a separate thread The VO is run inside its own thread. It also does most of video timing. The playloop hands the image data and a realtime timestamp to the VO, and the VO does the rest. In particular, this allows the playloop to do other things, instead of blocking for video redraw. But if anything accesses the VO during video timing, it will block. This also fixes vo_sdl.c event handling; but that is only a side-effect, since reimplementing the broken way would require more effort. Also drop --softsleep. In theory, this option helps if the kernel's sleeping mechanism is too inaccurate for video timing. In practice, I haven't ever encountered a situation where it helps, and it just burns CPU cycles. On the other hand it's probably actively harmful, because it prevents the libavcodec decoder threads from doing real work. Side note: Originally, I intended that multiple frames can be queued to the VO. But this is not done, due to problems with OSD and other certain features. OSD in particular is simply designed in a way that it can be neither timed nor copied, so you do have to render it into the video frame before you can draw the next frame. (Subtitles have no such restriction. sd_lavc was even updated to fix this.) It seems the right solution to queuing multiple VO frames is rendering on VO-backed framebuffers, like vo_vdpau.c does. This requires VO driver support, and is out of scope of this commit. As consequence, the VO has a queue size of 1. The existing video queue is just needed to compute frame duration, and will be moved out in the next commit.
2014-08-12 21:02:08 +00:00
double time_frame = MPMAX(mpctx->time_frame, -1);
int64_t pts = mp_time_us() + (int64_t)(time_frame * 1e6);
player: move video display code out of the playloop Basically move the code from playloop.c to video.c. The new function write_video() now contains the code that was part of run_playloop(). There are no functional changes, except handling "new_frame_shown" slightly differently. This is done so that we don't need new a new MPContext field or a return value for write_video() to signal this condition. Instead, it's handled indirectly.
2014-07-30 21:24:08 +00:00
video: better pipelining with vf_vapoursynth Most of this is explained in the code comments. This change should improve performance with vapoursynth, especially if concurrent requests are used. This should change nothing if vf_vapoursynth is not in the filter chain, since non-threaded filters obviously can not asynchronously finish filtering of frames.
2015-01-03 02:01:58 +00:00
// wait until VO wakes us up to get more frames
player: add display sync mode If this mode is enabled, the player tries to strictly synchronize video to display refresh. It will adjust playback speed to match the display, so if you play 23.976 fps video on a 24 Hz screen, playback speed is increased by approximately 1/1000. Audio wll be resampled to keep up with playback. This is different from the default sync mode, which will sync video to audio, with the consequence that video might skip or repeat a frame once in a while to make video keep up with audio. This is still unpolished. There are some major problems as well; in particular, mkv VFR files won't work well. The reason is that Matroska is terrible and rounds timestamps to milliseconds. This makes it rather hard to guess the framerate of a section of video that is playing. We could probably fix this by just accepting jittery timestamps (instead of explicitly disabling the sync code in this case), but I'm not ready to accept such a solution yet. Another issue is that we are extremely reliant on OS video and audio APIs working in an expected manner, which of course is not too often the case. Consequently, the new sync mode is a bit fragile.
2015-08-10 16:43:25 +00:00
// (NB: in theory, the 1st frame after display sync mode change uses the
// wrong waiting mode)
if (!vo_is_ready_for_frame(vo, mpctx->display_sync_active ? -1 : pts)) {
video: better pipelining with vf_vapoursynth Most of this is explained in the code comments. This change should improve performance with vapoursynth, especially if concurrent requests are used. This should change nothing if vf_vapoursynth is not in the filter chain, since non-threaded filters obviously can not asynchronously finish filtering of frames.
2015-01-03 02:01:58 +00:00
if (video_feed_async_filter(mpctx) < 0)
goto error;
return;
}
video: move display and timing to a separate thread The VO is run inside its own thread. It also does most of video timing. The playloop hands the image data and a realtime timestamp to the VO, and the VO does the rest. In particular, this allows the playloop to do other things, instead of blocking for video redraw. But if anything accesses the VO during video timing, it will block. This also fixes vo_sdl.c event handling; but that is only a side-effect, since reimplementing the broken way would require more effort. Also drop --softsleep. In theory, this option helps if the kernel's sleeping mechanism is too inaccurate for video timing. In practice, I haven't ever encountered a situation where it helps, and it just burns CPU cycles. On the other hand it's probably actively harmful, because it prevents the libavcodec decoder threads from doing real work. Side note: Originally, I intended that multiple frames can be queued to the VO. But this is not done, due to problems with OSD and other certain features. OSD in particular is simply designed in a way that it can be neither timed nor copied, so you do have to render it into the video frame before you can draw the next frame. (Subtitles have no such restriction. sd_lavc was even updated to fix this.) It seems the right solution to queuing multiple VO frames is rendering on VO-backed framebuffers, like vo_vdpau.c does. This requires VO driver support, and is out of scope of this commit. As consequence, the VO has a queue size of 1. The existing video queue is just needed to compute frame duration, and will be moved out in the next commit.
2014-08-12 21:02:08 +00:00
video: pass future frames to VO Now the VO can request a number of future frames with the last parameter of vo_set_queue_params(). This will be helpful to fix the interpolation code. Note that the first frame (after playback start or seeking) will usually not have any future frames (to make seeking fast). Near the end of the file, the number of future frames will become lower as well.
2015-07-01 17:22:40 +00:00
assert(mpctx->num_next_frames >= 1);
player: refactor display-sync frame duration calculations Get rid of get_past_frame_durations(), which was a bit too messy. Add a past_frames array, which contains the same information in a more reasonable way. This also means that we can get the exact current and past frame durations without going through awful stuff. (The main problem is that vo_pts_history contains future frames as well, which is needed for frame backstepping etc., but gets in the way here.) Also disable the automatic disabling of display-sync if the frame duration changes, and extend the frame durations allowed for display sync. To allow arbitrarily high durations, vo.c needs to be changed to pause and potentially redraw OSD while showing a single frame, so they're still limited. In an attempt to deal with VFR, calculate the overall speed using the average FPS. The frame scheduling itself does not use the average FPS, but the duration of the current frame. This does not work too well, but provides a good base for further improvements. Where this commit actually helps a lot is dealing with rounded timestamps, e.g. if the container framerate is wrong or unknown, or if the muxer wrote incorrectly rounded timestamps. While the rounding errors apparently can't be get rid of completely in the general case, this is still much better than e.g. disabling display-sync completely just because some frame durations go out of bounds.
2015-11-13 21:45:40 +00:00
if (mpctx->num_past_frames >= MAX_NUM_VO_PTS)
mpctx->num_past_frames--;
MP_TARRAY_INSERT_AT(mpctx, mpctx->past_frames, mpctx->num_past_frames, 0,
(struct frame_info){0});
command: add vsync-ratio property This is very "illustrative", unlike the video-speed-correction property, and thus useful. It can also be used to observe scheduling errors, which are not detected by the core. (These happen due to rounding errors; possibly not evne our fault, but coming from files with rounded timestamps and so on.)
2015-11-13 21:48:32 +00:00
mpctx->past_frames[0] = (struct frame_info){
.pts = mpctx->next_frames[0]->pts,
.num_vsyncs = -1,
};
player: refactor display-sync frame duration calculations Get rid of get_past_frame_durations(), which was a bit too messy. Add a past_frames array, which contains the same information in a more reasonable way. This also means that we can get the exact current and past frame durations without going through awful stuff. (The main problem is that vo_pts_history contains future frames as well, which is needed for frame backstepping etc., but gets in the way here.) Also disable the automatic disabling of display-sync if the frame duration changes, and extend the frame durations allowed for display sync. To allow arbitrarily high durations, vo.c needs to be changed to pause and potentially redraw OSD while showing a single frame, so they're still limited. In an attempt to deal with VFR, calculate the overall speed using the average FPS. The frame scheduling itself does not use the average FPS, but the duration of the current frame. This does not work too well, but provides a good base for further improvements. Where this commit actually helps a lot is dealing with rounded timestamps, e.g. if the container framerate is wrong or unknown, or if the muxer wrote incorrectly rounded timestamps. While the rounding errors apparently can't be get rid of completely in the general case, this is still much better than e.g. disabling display-sync completely just because some frame durations go out of bounds.
2015-11-13 21:45:40 +00:00
calculate_frame_duration(mpctx);
video: limit number of frames sent to VO to the VO requested amount vo_frame can have more than 1 frame - the extra frames are future references, which are sometimes useful for filtering (vo_opengl interpolation). There's no harm in reducing the number of frames sent to the VO requested amount of future frames, so do that. Doesn't actually reduce the number of concurrently in use frames in practice.
2016-07-07 14:10:13 +00:00
int req = vo_get_num_req_frames(mpctx->video_out);
assert(req >= 1 && req <= VO_MAX_REQ_FRAMES);
video: move up vo_frame setup
2015-07-28 21:54:39 +00:00
struct vo_frame dummy = {
.pts = pts,
.duration = -1,
video: disable interpolation during framestepping It just causes annoying artifacts. Interestingly, this means keeping down the frame stepping key (".") will play video with interpolation disabled.
2015-08-25 19:48:11 +00:00
.still = mpctx->step_frames > 0,
video: limit number of frames sent to VO to the VO requested amount vo_frame can have more than 1 frame - the extra frames are future references, which are sometimes useful for filtering (vo_opengl interpolation). There's no harm in reducing the number of frames sent to the VO requested amount of future frames, so do that. Doesn't actually reduce the number of concurrently in use frames in practice.
2016-07-07 14:10:13 +00:00
.num_frames = MPMIN(mpctx->num_next_frames, req),
player: add display sync mode If this mode is enabled, the player tries to strictly synchronize video to display refresh. It will adjust playback speed to match the display, so if you play 23.976 fps video on a 24 Hz screen, playback speed is increased by approximately 1/1000. Audio wll be resampled to keep up with playback. This is different from the default sync mode, which will sync video to audio, with the consequence that video might skip or repeat a frame once in a while to make video keep up with audio. This is still unpolished. There are some major problems as well; in particular, mkv VFR files won't work well. The reason is that Matroska is terrible and rounds timestamps to milliseconds. This makes it rather hard to guess the framerate of a section of video that is playing. We could probably fix this by just accepting jittery timestamps (instead of explicitly disabling the sync code in this case), but I'm not ready to accept such a solution yet. Another issue is that we are extremely reliant on OS video and audio APIs working in an expected manner, which of course is not too often the case. Consequently, the new sync mode is a bit fragile.
2015-08-10 16:43:25 +00:00
.num_vsyncs = 1,
video: move up vo_frame setup
2015-07-28 21:54:39 +00:00
};
for (int n = 0; n < dummy.num_frames; n++)
dummy.frames[n] = mpctx->next_frames[n];
struct vo_frame *frame = vo_frame_ref(&dummy);
command: add vsync-ratio property This is very "illustrative", unlike the video-speed-correction property, and thus useful. It can also be used to observe scheduling errors, which are not detected by the core. (These happen due to rounding errors; possibly not evne our fault, but coming from files with rounded timestamps and so on.)
2015-11-13 21:48:32 +00:00
double diff = mpctx->past_frames[0].approx_duration;
player: refactor display-sync frame duration calculations Get rid of get_past_frame_durations(), which was a bit too messy. Add a past_frames array, which contains the same information in a more reasonable way. This also means that we can get the exact current and past frame durations without going through awful stuff. (The main problem is that vo_pts_history contains future frames as well, which is needed for frame backstepping etc., but gets in the way here.) Also disable the automatic disabling of display-sync if the frame duration changes, and extend the frame durations allowed for display sync. To allow arbitrarily high durations, vo.c needs to be changed to pause and potentially redraw OSD while showing a single frame, so they're still limited. In an attempt to deal with VFR, calculate the overall speed using the average FPS. The frame scheduling itself does not use the average FPS, but the duration of the current frame. This does not work too well, but provides a good base for further improvements. Where this commit actually helps a lot is dealing with rounded timestamps, e.g. if the container framerate is wrong or unknown, or if the muxer wrote incorrectly rounded timestamps. While the rounding errors apparently can't be get rid of completely in the general case, this is still much better than e.g. disabling display-sync completely just because some frame durations go out of bounds.
2015-11-13 21:45:40 +00:00
if (opts->untimed || vo->driver->untimed)
diff = -1; // disable frame dropping and aspects of frame timing
video: fix and simplify video format changes and last frame display The previous commit broke these things, and fixing them is separate in this commit in order to reduce the volume of changes. Move the image queue from the VO to the playback core. The image queue is a remnant of the old way how vdpau was implemented, and increasingly became more and more an artifact. In the end, it did only one thing: computing the duration of the current frame. This was done by taking the PTS difference between the current and the future frame. We keep this, but by moving it out of the VO, we don't have to special-case format changes anymore. This simplifies the code a lot. Since we need the queue to compute the duration only, a queue size larger than 2 makes no sense, and we can hardcode that. Also change how the last frame is handled. The last frame is a bit of a problem, because video timing works by showing one frame after another, which makes it a special case. Make the VO provide a function to notify us when the frame is done, instead. The frame duration is used for that. This is not perfect. For example, changing playback speed during the last frame doesn't update the end time. Pausing will not stop the clock that times the last frame. But I don't think this matters for such a corner case.
2014-08-12 21:17:35 +00:00
if (diff >= 0) {
video: move display and timing to a separate thread The VO is run inside its own thread. It also does most of video timing. The playloop hands the image data and a realtime timestamp to the VO, and the VO does the rest. In particular, this allows the playloop to do other things, instead of blocking for video redraw. But if anything accesses the VO during video timing, it will block. This also fixes vo_sdl.c event handling; but that is only a side-effect, since reimplementing the broken way would require more effort. Also drop --softsleep. In theory, this option helps if the kernel's sleeping mechanism is too inaccurate for video timing. In practice, I haven't ever encountered a situation where it helps, and it just burns CPU cycles. On the other hand it's probably actively harmful, because it prevents the libavcodec decoder threads from doing real work. Side note: Originally, I intended that multiple frames can be queued to the VO. But this is not done, due to problems with OSD and other certain features. OSD in particular is simply designed in a way that it can be neither timed nor copied, so you do have to render it into the video frame before you can draw the next frame. (Subtitles have no such restriction. sd_lavc was even updated to fix this.) It seems the right solution to queuing multiple VO frames is rendering on VO-backed framebuffers, like vo_vdpau.c does. This requires VO driver support, and is out of scope of this commit. As consequence, the VO has a queue size of 1. The existing video queue is just needed to compute frame duration, and will be moved out in the next commit.
2014-08-12 21:02:08 +00:00
// expected A/V sync correction is ignored
player: separate controls for user and video controlled speed For video sync, we want separate playback speed controls for user- requested speed and the "correction" speed for video timing. Further, we use this separation to make sure only a resampler is inserted if playback speed is only changed for video sync correction. As of this commit, this is basically inactive code. It's just preparation for the video sync code (the following commit).
2015-08-10 16:40:16 +00:00
diff /= mpctx->video_speed;
video: move display and timing to a separate thread The VO is run inside its own thread. It also does most of video timing. The playloop hands the image data and a realtime timestamp to the VO, and the VO does the rest. In particular, this allows the playloop to do other things, instead of blocking for video redraw. But if anything accesses the VO during video timing, it will block. This also fixes vo_sdl.c event handling; but that is only a side-effect, since reimplementing the broken way would require more effort. Also drop --softsleep. In theory, this option helps if the kernel's sleeping mechanism is too inaccurate for video timing. In practice, I haven't ever encountered a situation where it helps, and it just burns CPU cycles. On the other hand it's probably actively harmful, because it prevents the libavcodec decoder threads from doing real work. Side note: Originally, I intended that multiple frames can be queued to the VO. But this is not done, due to problems with OSD and other certain features. OSD in particular is simply designed in a way that it can be neither timed nor copied, so you do have to render it into the video frame before you can draw the next frame. (Subtitles have no such restriction. sd_lavc was even updated to fix this.) It seems the right solution to queuing multiple VO frames is rendering on VO-backed framebuffers, like vo_vdpau.c does. This requires VO driver support, and is out of scope of this commit. As consequence, the VO has a queue size of 1. The existing video queue is just needed to compute frame duration, and will be moved out in the next commit.
2014-08-12 21:02:08 +00:00
if (mpctx->time_frame < 0)
diff += mpctx->time_frame;
video: move up vo_frame setup
2015-07-28 21:54:39 +00:00
frame->duration = MPCLAMP(diff, 0, 10) * 1e6;
video: move display and timing to a separate thread The VO is run inside its own thread. It also does most of video timing. The playloop hands the image data and a realtime timestamp to the VO, and the VO does the rest. In particular, this allows the playloop to do other things, instead of blocking for video redraw. But if anything accesses the VO during video timing, it will block. This also fixes vo_sdl.c event handling; but that is only a side-effect, since reimplementing the broken way would require more effort. Also drop --softsleep. In theory, this option helps if the kernel's sleeping mechanism is too inaccurate for video timing. In practice, I haven't ever encountered a situation where it helps, and it just burns CPU cycles. On the other hand it's probably actively harmful, because it prevents the libavcodec decoder threads from doing real work. Side note: Originally, I intended that multiple frames can be queued to the VO. But this is not done, due to problems with OSD and other certain features. OSD in particular is simply designed in a way that it can be neither timed nor copied, so you do have to render it into the video frame before you can draw the next frame. (Subtitles have no such restriction. sd_lavc was even updated to fix this.) It seems the right solution to queuing multiple VO frames is rendering on VO-backed framebuffers, like vo_vdpau.c does. This requires VO driver support, and is out of scope of this commit. As consequence, the VO has a queue size of 1. The existing video queue is just needed to compute frame duration, and will be moved out in the next commit.
2014-08-12 21:02:08 +00:00
}
player: move video display code out of the playloop Basically move the code from playloop.c to video.c. The new function write_video() now contains the code that was part of run_playloop(). There are no functional changes, except handling "new_frame_shown" slightly differently. This is done so that we don't need new a new MPContext field or a return value for write_video() to signal this condition. Instead, it's handled indirectly.
2014-07-30 21:24:08 +00:00
video: pass future frames to VO Now the VO can request a number of future frames with the last parameter of vo_set_queue_params(). This will be helpful to fix the interpolation code. Note that the first frame (after playback start or seeking) will usually not have any future frames (to make seeking fast). Near the end of the file, the number of future frames will become lower as well.
2015-07-01 17:22:40 +00:00
mpctx->video_pts = mpctx->next_frames[0]->pts;
player: move video display code out of the playloop Basically move the code from playloop.c to video.c. The new function write_video() now contains the code that was part of run_playloop(). There are no functional changes, except handling "new_frame_shown" slightly differently. This is done so that we don't need new a new MPContext field or a return value for write_video() to signal this condition. Instead, it's handled indirectly.
2014-07-30 21:24:08 +00:00
mpctx->last_vo_pts = mpctx->video_pts;
video: use demuxer-signaled duration for last video frame Helps with gif, probably does unwanted things with other formats. This doesn't handle --end quite correctly, but this could be added later. Fixes #3924.
2016-12-21 17:18:24 +00:00
mpctx->last_frame_duration =
mpctx->next_frames[0]->pkt_duration / mpctx->video_speed;
player: move video display code out of the playloop Basically move the code from playloop.c to video.c. The new function write_video() now contains the code that was part of run_playloop(). There are no functional changes, except handling "new_frame_shown" slightly differently. This is done so that we don't need new a new MPContext field or a return value for write_video() to signal this condition. Instead, it's handled indirectly.
2014-07-30 21:24:08 +00:00
player: refactor display-sync frame duration calculations Get rid of get_past_frame_durations(), which was a bit too messy. Add a past_frames array, which contains the same information in a more reasonable way. This also means that we can get the exact current and past frame durations without going through awful stuff. (The main problem is that vo_pts_history contains future frames as well, which is needed for frame backstepping etc., but gets in the way here.) Also disable the automatic disabling of display-sync if the frame duration changes, and extend the frame durations allowed for display sync. To allow arbitrarily high durations, vo.c needs to be changed to pause and potentially redraw OSD while showing a single frame, so they're still limited. In an attempt to deal with VFR, calculate the overall speed using the average FPS. The frame scheduling itself does not use the average FPS, but the duration of the current frame. This does not work too well, but provides a good base for further improvements. Where this commit actually helps a lot is dealing with rounded timestamps, e.g. if the container framerate is wrong or unknown, or if the muxer wrote incorrectly rounded timestamps. While the rounding errors apparently can't be get rid of completely in the general case, this is still much better than e.g. disabling display-sync completely just because some frame durations go out of bounds.
2015-11-13 21:45:40 +00:00
shift_frames(mpctx);
player: minor refactor for A/V diff computation Calculate the A/V difference directly in the display sync code, instead of the awkward current way, which reuses the fields for audio sync. We still set time_frame, because it makes falling back to audio sync somewhat smoother.
2015-10-28 22:50:54 +00:00
schedule_frame(mpctx, frame);
player: rearrange some A/V-sync related code Move the update_avsync_before_frame() call further down. Moving it closer to where the time_frame value is used (and which the function updates) should make the code more readable. With this change, there's no need anymore to reset the time_frame value on the video reconfig path. Move the update_avsync_after_frame() up. Now no meaningful amount of time passes since the previous get_relative_time() call anymore, and the second one can be removed.
2015-01-30 20:18:30 +00:00
player: rate-limit OSD text update There's no need to update OSD messages and the terminal status if nobody is going to see it. Since the player doesn't block on video display anymore, this update happens to often and probably burns slightly more CPU than necessary. (OSD redrawing is handled separately, so it's just mostly useless text processing and such.) Change it so that it's updated only on every video frame or all 50ms (whatever comes first). For VO OSD, we could in theory try to lock to the OSD redraw heuristic or the display refresh rate, but that's more complicated and doesn't work for the terminal status.
2014-09-25 18:25:24 +00:00
mpctx->osd_force_update = true;
player: move video display code out of the playloop Basically move the code from playloop.c to video.c. The new function write_video() now contains the code that was part of run_playloop(). There are no functional changes, except handling "new_frame_shown" slightly differently. This is done so that we don't need new a new MPContext field or a return value for write_video() to signal this condition. Instead, it's handled indirectly.
2014-07-30 21:24:08 +00:00
update_osd_msg(mpctx);
video: move up vo_frame setup
2015-07-28 21:54:39 +00:00
vo_queue_frame(vo, frame);
vo: change internal API for drawing frames draw_image_timed is renamed to draw_frame. struct frame_timing is renamed to vo_frame. flip_page_timed is merged into draw_frame (the additional parameters are part of struct vo_frame). draw_frame also deprecates VOCTRL_REDRAW_FRAME, and replaces it with a method that works for both VOs which can cache the current frame, and VOs which need to redraw it anyway. This is preparation to making the interpolation and (work in progress) display sync code saner. Lots of other refactoring, and also some simplifications.
2015-07-01 17:24:28 +00:00
video: pass future frames to VO Now the VO can request a number of future frames with the last parameter of vo_set_queue_params(). This will be helpful to fix the interpolation code. Note that the first frame (after playback start or seeking) will usually not have any future frames (to make seeking fast). Near the end of the file, the number of future frames will become lower as well.
2015-07-01 17:22:40 +00:00
// The frames were shifted down; "initialize" the new first entry.
if (mpctx->num_next_frames >= 1)
handle_new_frame(mpctx);
video: refactor aspects of queue and EOF handling Instead of touching the 2-entry queue in mpctx->next_frame directly, move some of it to functions.
2015-03-25 12:33:34 +00:00
player: move video display code out of the playloop Basically move the code from playloop.c to video.c. The new function write_video() now contains the code that was part of run_playloop(). There are no functional changes, except handling "new_frame_shown" slightly differently. This is done so that we don't need new a new MPContext field or a return value for write_video() to signal this condition. Instead, it's handled indirectly.
2014-07-30 21:24:08 +00:00
mpctx->shown_vframes++;
video: move display and timing to a separate thread The VO is run inside its own thread. It also does most of video timing. The playloop hands the image data and a realtime timestamp to the VO, and the VO does the rest. In particular, this allows the playloop to do other things, instead of blocking for video redraw. But if anything accesses the VO during video timing, it will block. This also fixes vo_sdl.c event handling; but that is only a side-effect, since reimplementing the broken way would require more effort. Also drop --softsleep. In theory, this option helps if the kernel's sleeping mechanism is too inaccurate for video timing. In practice, I haven't ever encountered a situation where it helps, and it just burns CPU cycles. On the other hand it's probably actively harmful, because it prevents the libavcodec decoder threads from doing real work. Side note: Originally, I intended that multiple frames can be queued to the VO. But this is not done, due to problems with OSD and other certain features. OSD in particular is simply designed in a way that it can be neither timed nor copied, so you do have to render it into the video frame before you can draw the next frame. (Subtitles have no such restriction. sd_lavc was even updated to fix this.) It seems the right solution to queuing multiple VO frames is rendering on VO-backed framebuffers, like vo_vdpau.c does. This requires VO driver support, and is out of scope of this commit. As consequence, the VO has a queue size of 1. The existing video queue is just needed to compute frame duration, and will be moved out in the next commit.
2014-08-12 21:02:08 +00:00
if (mpctx->video_status < STATUS_PLAYING) {
player: move video display code out of the playloop Basically move the code from playloop.c to video.c. The new function write_video() now contains the code that was part of run_playloop(). There are no functional changes, except handling "new_frame_shown" slightly differently. This is done so that we don't need new a new MPContext field or a return value for write_video() to signal this condition. Instead, it's handled indirectly.
2014-07-30 21:24:08 +00:00
mpctx->video_status = STATUS_READY;
video: move display and timing to a separate thread The VO is run inside its own thread. It also does most of video timing. The playloop hands the image data and a realtime timestamp to the VO, and the VO does the rest. In particular, this allows the playloop to do other things, instead of blocking for video redraw. But if anything accesses the VO during video timing, it will block. This also fixes vo_sdl.c event handling; but that is only a side-effect, since reimplementing the broken way would require more effort. Also drop --softsleep. In theory, this option helps if the kernel's sleeping mechanism is too inaccurate for video timing. In practice, I haven't ever encountered a situation where it helps, and it just burns CPU cycles. On the other hand it's probably actively harmful, because it prevents the libavcodec decoder threads from doing real work. Side note: Originally, I intended that multiple frames can be queued to the VO. But this is not done, due to problems with OSD and other certain features. OSD in particular is simply designed in a way that it can be neither timed nor copied, so you do have to render it into the video frame before you can draw the next frame. (Subtitles have no such restriction. sd_lavc was even updated to fix this.) It seems the right solution to queuing multiple VO frames is rendering on VO-backed framebuffers, like vo_vdpau.c does. This requires VO driver support, and is out of scope of this commit. As consequence, the VO has a queue size of 1. The existing video queue is just needed to compute frame duration, and will be moved out in the next commit.
2014-08-12 21:02:08 +00:00
// After a seek, make sure to wait until the first frame is visible.
vo_wait_frame(vo);
player: add some debug output for seeking
2015-06-18 20:31:55 +00:00
MP_VERBOSE(mpctx, "first video frame after restart shown\n");
video: move display and timing to a separate thread The VO is run inside its own thread. It also does most of video timing. The playloop hands the image data and a realtime timestamp to the VO, and the VO does the rest. In particular, this allows the playloop to do other things, instead of blocking for video redraw. But if anything accesses the VO during video timing, it will block. This also fixes vo_sdl.c event handling; but that is only a side-effect, since reimplementing the broken way would require more effort. Also drop --softsleep. In theory, this option helps if the kernel's sleeping mechanism is too inaccurate for video timing. In practice, I haven't ever encountered a situation where it helps, and it just burns CPU cycles. On the other hand it's probably actively harmful, because it prevents the libavcodec decoder threads from doing real work. Side note: Originally, I intended that multiple frames can be queued to the VO. But this is not done, due to problems with OSD and other certain features. OSD in particular is simply designed in a way that it can be neither timed nor copied, so you do have to render it into the video frame before you can draw the next frame. (Subtitles have no such restriction. sd_lavc was even updated to fix this.) It seems the right solution to queuing multiple VO frames is rendering on VO-backed framebuffers, like vo_vdpau.c does. This requires VO driver support, and is out of scope of this commit. As consequence, the VO has a queue size of 1. The existing video queue is just needed to compute frame duration, and will be moved out in the next commit.
2014-08-12 21:02:08 +00:00
}
player: move video display code out of the playloop Basically move the code from playloop.c to video.c. The new function write_video() now contains the code that was part of run_playloop(). There are no functional changes, except handling "new_frame_shown" slightly differently. This is done so that we don't need new a new MPContext field or a return value for write_video() to signal this condition. Instead, it's handled indirectly.
2014-07-30 21:24:08 +00:00
screenshot_flip(mpctx);
mp_notify(mpctx, MPV_EVENT_TICK, NULL);
player: use different variable to indicate coverart Slightly better.
2016-02-01 21:14:32 +00:00
if (mpctx->vo_chain->is_coverart)
player: move video display code out of the playloop Basically move the code from playloop.c to video.c. The new function write_video() now contains the code that was part of run_playloop(). There are no functional changes, except handling "new_frame_shown" slightly differently. This is done so that we don't need new a new MPContext field or a return value for write_video() to signal this condition. Instead, it's handled indirectly.
2014-07-30 21:24:08 +00:00
mpctx->video_status = STATUS_EOF;
player: some further playloop cleanups Handle --term-playing-msg at a better place. Move MPV_EVENT_TICK hack into a separate function. Also add some words to the client API that you shouldn't use it. (But better leave breaking it for later.) Handle --frames and frame_step differently. Remove the mess from the playloop, and do it after frame display. Give up on the weird semantics for audio-only mode (they didn't make sense anyway), and adjust the manpage accordingly.
2014-08-03 18:25:03 +00:00
if (mpctx->video_status != STATUS_EOF) {
if (mpctx->step_frames > 0) {
mpctx->step_frames--;
player: unmess pause state handling Merge the pause_player() and unpause_player() functions. Make sure the pause events are emitted properly. We can now set the internal pause state based on a predicate, instead of e.g. handle_pause_on_low_cache() making a mess to trigger the internal pause state as wanted. Preparation for some more changes.
2017-04-14 16:22:45 +00:00
if (!mpctx->step_frames)
set_pause_state(mpctx, true);
player: some further playloop cleanups Handle --term-playing-msg at a better place. Move MPV_EVENT_TICK hack into a separate function. Also add some words to the client API that you shouldn't use it. (But better leave breaking it for later.) Handle --frames and frame_step differently. Remove the mess from the playloop, and do it after frame display. Give up on the weird semantics for audio-only mode (they didn't make sense anyway), and adjust the manpage accordingly.
2014-08-03 18:25:03 +00:00
}
player: never overwrite stop_play field This is a real pain: if a quit command is received, it's set to PT_QUIT. And then other code could overwrite it, making it not quit. The annoying bit is that stop_play is written and read in many places. Just not overwriting it unconditionally seems to be the best course of action.
2015-07-08 19:31:31 +00:00
if (mpctx->max_frames == 0 && !mpctx->stop_play)
player: signal EOF when using --frames
2014-10-10 13:14:11 +00:00
mpctx->stop_play = AT_END_OF_FILE;
player: some further playloop cleanups Handle --term-playing-msg at a better place. Move MPV_EVENT_TICK hack into a separate function. Also add some words to the client API that you shouldn't use it. (But better leave breaking it for later.) Handle --frames and frame_step differently. Remove the mess from the playloop, and do it after frame display. Give up on the weird semantics for audio-only mode (they didn't make sense anyway), and adjust the manpage accordingly.
2014-08-03 18:25:03 +00:00
if (mpctx->max_frames > 0)
mpctx->max_frames--;
}
video: fix and simplify video format changes and last frame display The previous commit broke these things, and fixing them is separate in this commit in order to reduce the volume of changes. Move the image queue from the VO to the playback core. The image queue is a remnant of the old way how vdpau was implemented, and increasingly became more and more an artifact. In the end, it did only one thing: computing the duration of the current frame. This was done by taking the PTS difference between the current and the future frame. We keep this, but by moving it out of the VO, we don't have to special-case format changes anymore. This simplifies the code a lot. Since we need the queue to compute the duration only, a queue size larger than 2 makes no sense, and we can hardcode that. Also change how the last frame is handled. The last frame is a bit of a problem, because video timing works by showing one frame after another, which makes it a special case. Make the VO provide a function to notify us when the frame is done, instead. The frame duration is used for that. This is not perfect. For example, changing playback speed during the last frame doesn't update the end time. Pausing will not stop the clock that times the last frame. But I don't think this matters for such a corner case.
2014-08-12 21:17:35 +00:00
player: litter code with explicit wakeup calls This does 3 kinds of changes: - change sleeptime=x to mp_set_timeout() - change sleeptime=0 to mp_wakeup_core() calls (to be more explicit) - change commands etc. to call mp_wakeup_core() if they do changes that require the playloop to be rerun This is preparation for the following changes. The goal is to process client API requests without having to rerun the playloop every time. As of this commit, the changes should not change behavior. In particular, the playloop is still implicitly woken up on every command.
2016-09-16 12:24:15 +00:00
mp_wakeup_core(mpctx);
video: fix and simplify video format changes and last frame display The previous commit broke these things, and fixing them is separate in this commit in order to reduce the volume of changes. Move the image queue from the VO to the playback core. The image queue is a remnant of the old way how vdpau was implemented, and increasingly became more and more an artifact. In the end, it did only one thing: computing the duration of the current frame. This was done by taking the PTS difference between the current and the future frame. We keep this, but by moving it out of the VO, we don't have to special-case format changes anymore. This simplifies the code a lot. Since we need the queue to compute the duration only, a queue size larger than 2 makes no sense, and we can hardcode that. Also change how the last frame is handled. The last frame is a bit of a problem, because video timing works by showing one frame after another, which makes it a special case. Make the VO provide a function to notify us when the frame is done, instead. The frame duration is used for that. This is not perfect. For example, changing playback speed during the last frame doesn't update the end time. Pausing will not stop the clock that times the last frame. But I don't think this matters for such a corner case.
2014-08-12 21:17:35 +00:00
return;
error:
MP_FATAL(mpctx, "Could not initialize video chain.\n");
player: remove central uninit_player() function and flags mess Each subsystem (or similar thing) had an INITIALIZED_ flag assigned. The main use of this was that you could pass a bitmask of these flags to uninit_player(). Except in some situations where you wanted to uninitialize nearly everything, this wasn't really useful. Moreover, it was quite annoying that subsystems had most of the code in a specific file, but the uninit code in loadfile.c (because that's where uninit_player() was implemented). Simplify all this. Remove the flags; e.g. instead of testing for the INITIALIZED_AO flag, test whether mpctx->ao is set. Move uninit code to separate functions, e.g. uninit_audio_out().
2014-10-03 17:57:49 +00:00
uninit_video_chain(mpctx);
player: remove some further current_track dependencies Now it's used for initialization only for audio and video.
2016-02-05 22:41:44 +00:00
error_on_track(mpctx, track);
video: fix and simplify video format changes and last frame display The previous commit broke these things, and fixing them is separate in this commit in order to reduce the volume of changes. Move the image queue from the VO to the playback core. The image queue is a remnant of the old way how vdpau was implemented, and increasingly became more and more an artifact. In the end, it did only one thing: computing the duration of the current frame. This was done by taking the PTS difference between the current and the future frame. We keep this, but by moving it out of the VO, we don't have to special-case format changes anymore. This simplifies the code a lot. Since we need the queue to compute the duration only, a queue size larger than 2 makes no sense, and we can hardcode that. Also change how the last frame is handled. The last frame is a bit of a problem, because video timing works by showing one frame after another, which makes it a special case. Make the VO provide a function to notify us when the frame is done, instead. The frame duration is used for that. This is not perfect. For example, changing playback speed during the last frame doesn't update the end time. Pausing will not stop the clock that times the last frame. But I don't think this matters for such a corner case.
2014-08-12 21:17:35 +00:00
handle_force_window(mpctx, true);
player: litter code with explicit wakeup calls This does 3 kinds of changes: - change sleeptime=x to mp_set_timeout() - change sleeptime=0 to mp_wakeup_core() calls (to be more explicit) - change commands etc. to call mp_wakeup_core() if they do changes that require the playloop to be rerun This is preparation for the following changes. The goal is to process client API requests without having to rerun the playloop every time. As of this commit, the changes should not change behavior. In particular, the playloop is still implicitly woken up on every command.
2016-09-16 12:24:15 +00:00
mp_wakeup_core(mpctx);
player: move video display code out of the playloop Basically move the code from playloop.c to video.c. The new function write_video() now contains the code that was part of run_playloop(). There are no functional changes, except handling "new_frame_shown" slightly differently. This is done so that we don't need new a new MPContext field or a return value for write_video() to signal this condition. Instead, it's handled indirectly.
2014-07-30 21:24:08 +00:00
}