mpv/player/playloop.c

978 lines
31 KiB
C
Raw Normal View History

/*
* This file is part of MPlayer.
*
* MPlayer is free software; you can redistribute it and/or modify
* it under the terms of the GNU General Public License as published by
* the Free Software Foundation; either version 2 of the License, or
* (at your option) any later version.
*
* MPlayer is distributed in the hope that it will be useful,
* but WITHOUT ANY WARRANTY; without even the implied warranty of
* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
* GNU General Public License for more details.
*
* You should have received a copy of the GNU General Public License along
* with MPlayer; if not, write to the Free Software Foundation, Inc.,
* 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA.
*/
#include <stddef.h>
#include <stdbool.h>
#include <inttypes.h>
#include <math.h>
#include <assert.h>
#include "config.h"
#include "talloc.h"
#include "common/msg.h"
#include "options/options.h"
#include "common/common.h"
#include "common/encode.h"
#include "options/m_property.h"
#include "common/playlist.h"
2013-12-17 00:23:09 +00:00
#include "input/input.h"
#include "misc/dispatch.h"
#include "osdep/terminal.h"
#include "osdep/timer.h"
#include "audio/mixer.h"
#include "audio/decode/dec_audio.h"
#include "audio/filter/af.h"
#include "audio/out/ao.h"
#include "demux/demux.h"
#include "stream/stream.h"
#include "sub/osd.h"
#include "video/filter/vf.h"
#include "video/decode/dec_video.h"
#include "video/out/vo.h"
#include "core.h"
#include "command.h"
void pause_player(struct MPContext *mpctx)
{
mpctx->opts->pause = 1;
if (mpctx->video_out)
vo_control(mpctx->video_out, VOCTRL_RESTORE_SCREENSAVER, NULL);
if (mpctx->paused)
goto end;
mpctx->paused = true;
mpctx->step_frames = 0;
mpctx->time_frame -= get_relative_time(mpctx);
mpctx->osd_function = 0;
mpctx->paused_for_cache = false;
if (mpctx->video_out && mpctx->d_video && mpctx->video_out->config_ok)
vo_control(mpctx->video_out, VOCTRL_PAUSE, NULL);
if (mpctx->ao && mpctx->d_audio)
ao_pause(mpctx->ao); // pause audio, keep data if possible
// Only print status if there's actually a file being played.
if (mpctx->num_sources)
print_status(mpctx);
end:
mp_notify(mpctx, mpctx->opts->pause ? MPV_EVENT_PAUSE : MPV_EVENT_UNPAUSE, 0);
}
void unpause_player(struct MPContext *mpctx)
{
mpctx->opts->pause = 0;
if (mpctx->video_out && mpctx->opts->stop_screensaver)
vo_control(mpctx->video_out, VOCTRL_KILL_SCREENSAVER, NULL);
if (!mpctx->paused)
goto end;
// Don't actually unpause while cache is loading.
if (mpctx->paused_for_cache)
goto end;
mpctx->paused = false;
mpctx->osd_function = 0;
if (mpctx->ao && mpctx->d_audio)
ao_resume(mpctx->ao);
if (mpctx->video_out && mpctx->d_video && mpctx->video_out->config_ok)
vo_control(mpctx->video_out, VOCTRL_RESUME, NULL); // resume video
(void)get_relative_time(mpctx); // ignore time that passed during pause
end:
mp_notify(mpctx, mpctx->opts->pause ? MPV_EVENT_PAUSE : MPV_EVENT_UNPAUSE, 0);
}
void add_step_frame(struct MPContext *mpctx, int dir)
{
if (!mpctx->d_video)
return;
if (dir > 0) {
mpctx->step_frames += 1;
unpause_player(mpctx);
} else if (dir < 0) {
if (!mpctx->backstep_active && !mpctx->hrseek_active) {
mpctx->backstep_active = true;
mpctx->backstep_start_seek_ts = mpctx->vo_pts_history_seek_ts;
pause_player(mpctx);
}
}
}
// Clear some playback-related fields on file loading or after seeks.
void reset_playback_state(struct MPContext *mpctx)
{
reset_video_state(mpctx);
reset_audio_state(mpctx);
reset_subtitle_state(mpctx);
mpctx->hrseek_active = false;
mpctx->hrseek_framedrop = false;
mpctx->playback_pts = MP_NOPTS_VALUE;
mpctx->last_seek_pts = MP_NOPTS_VALUE;
mpctx->restart_complete = false;
#if HAVE_ENCODING
encode_lavc_discontinuity(mpctx->encode_lavc_ctx);
#endif
}
// return -1 if seek failed (non-seekable stream?), 0 otherwise
static int mp_seek(MPContext *mpctx, struct seek_params seek,
bool timeline_fallthrough)
{
struct MPOpts *opts = mpctx->opts;
uint64_t prev_seek_ts = mpctx->vo_pts_history_seek_ts;
if (!mpctx->demuxer)
return -1;
if (!mpctx->demuxer->seekable) {
MP_ERR(mpctx, "Can't seek in this file.\n");
return -1;
}
if (mpctx->stop_play == AT_END_OF_FILE)
mpctx->stop_play = KEEP_PLAYING;
double hr_seek_offset = opts->hr_seek_demuxer_offset;
bool hr_seek_very_exact = seek.exact > 1;
// Always try to compensate for possibly bad demuxers in "special"
// situations where we need more robustness from the hr-seek code, even
// if the user doesn't use --hr-seek-demuxer-offset.
// The value is arbitrary, but should be "good enough" in most situations.
if (hr_seek_very_exact)
hr_seek_offset = MPMAX(hr_seek_offset, 0.5); // arbitrary
bool hr_seek = opts->correct_pts && seek.exact >= 0;
hr_seek &= (opts->hr_seek == 0 && seek.type == MPSEEK_ABSOLUTE) ||
opts->hr_seek > 0 || seek.exact > 0;
if (seek.type == MPSEEK_FACTOR || seek.amount < 0 ||
(seek.type == MPSEEK_ABSOLUTE && seek.amount < mpctx->last_chapter_pts))
mpctx->last_chapter_seek = -2;
if (seek.type == MPSEEK_FACTOR) {
double len = get_time_length(mpctx);
if (len > 0 && !mpctx->demuxer->ts_resets_possible) {
seek.amount = seek.amount * len + get_start_time(mpctx);
seek.type = MPSEEK_ABSOLUTE;
}
}
2014-07-28 23:00:54 +00:00
int direction = 0;
if (seek.type == MPSEEK_RELATIVE) {
seek.type = MPSEEK_ABSOLUTE;
2014-07-28 23:00:54 +00:00
direction = seek.amount > 0 ? 1 : -1;
seek.amount += get_current_time(mpctx);
}
hr_seek &= seek.type == MPSEEK_ABSOLUTE; // otherwise, no target PTS known
double demuxer_amount = seek.amount;
if (mpctx->timeline) {
bool need_reset = false;
demuxer_amount = timeline_set_from_time(mpctx, seek.amount,
&need_reset);
if (demuxer_amount == -1) {
assert(!need_reset);
mpctx->stop_play = AT_END_OF_FILE;
if (mpctx->d_audio && !timeline_fallthrough) {
// Seek outside of the file -> clear audio from current position
clear_audio_decode_buffers(mpctx);
clear_audio_output_buffers(mpctx);
}
return -1;
}
if (need_reset) {
reinit_video_chain(mpctx);
reinit_audio_chain(mpctx);
reinit_subs(mpctx, 0);
reinit_subs(mpctx, 1);
}
}
int demuxer_style = 0;
switch (seek.type) {
case MPSEEK_FACTOR:
demuxer_style |= SEEK_ABSOLUTE | SEEK_FACTOR;
break;
case MPSEEK_ABSOLUTE:
demuxer_style |= SEEK_ABSOLUTE;
break;
}
2014-07-28 23:00:54 +00:00
if (hr_seek || direction < 0)
demuxer_style |= SEEK_BACKWARD;
2014-07-28 23:00:54 +00:00
else if (direction > 0)
demuxer_style |= SEEK_FORWARD;
if (hr_seek || opts->mkv_subtitle_preroll)
demuxer_style |= SEEK_SUBPREROLL;
if (hr_seek)
demuxer_amount -= hr_seek_offset;
demux_seek(mpctx->demuxer, demuxer_amount, demuxer_style);
// Seek external, extra files too:
for (int t = 0; t < mpctx->num_tracks; t++) {
struct track *track = mpctx->tracks[t];
if (track->selected && track->is_external && track->demuxer) {
double main_new_pos;
if (seek.type == MPSEEK_ABSOLUTE) {
main_new_pos = seek.amount - mpctx->video_offset;
} else {
main_new_pos = get_main_demux_pts(mpctx);
}
demux_seek(track->demuxer, main_new_pos, SEEK_ABSOLUTE);
}
}
if (!timeline_fallthrough)
clear_audio_output_buffers(mpctx);
reset_playback_state(mpctx);
if (timeline_fallthrough) {
// Important if video reinit happens.
mpctx->vo_pts_history_seek_ts = prev_seek_ts;
} else {
mpctx->vo_pts_history_seek_ts++;
mpctx->backstep_active = false;
}
/* Use the target time as "current position" for further relative
* seeks etc until a new video frame has been decoded */
if (seek.type == MPSEEK_ABSOLUTE) {
mpctx->video_pts = seek.amount;
mpctx->last_seek_pts = seek.amount;
}
// The hr_seek==false case is for skipping frames with PTS before the
// current timeline chapter start. It's not really known where the demuxer
// level seek will end up, so the hrseek mechanism is abused to skip all
// frames before chapter start by setting hrseek_pts to the chapter start.
// It does nothing when the seek is inside of the current chapter, and
// seeking past the chapter is handled elsewhere.
if (hr_seek || mpctx->timeline) {
mpctx->hrseek_active = true;
mpctx->hrseek_framedrop = !hr_seek_very_exact;
mpctx->hrseek_pts = hr_seek ? seek.amount
: mpctx->timeline[mpctx->timeline_part].start;
}
mpctx->start_timestamp = mp_time_sec();
2014-07-21 17:31:25 +00:00
mpctx->sleeptime = 0;
mp_notify(mpctx, MPV_EVENT_SEEK, NULL);
mp_notify(mpctx, MPV_EVENT_TICK, NULL);
return 0;
}
player: handle seek delays differently The code removed from handle_input_and_seek_coalesce() did two things: 1. If there's a queued seek, stop accepting non-seek commands, and delay them to the next playloop iteration. 2. If a seek is executing (i.e. the seek was unqueued, and now it's trying to decode and display the first video frame), stop accepting seek commands (and in fact all commands that were queued after the first seek command). This logic is disabled if seeking started longer than 300ms ago. (To avoid starvation.) I'm not sure why 1. would be needed. It's still possible that a command immediately executed after a seek command sees a "seeking in progress" state, because it affects queued seeks only, and not seeks in progress. Drop this code, since it can easily lead to input starvation, and I'm not aware of any disadvantages. The logic in 2. is good to make seeking behave much better, as it guarantees that the video display is updated frequently. Keep the core idea, but implement it differently. Now this logic is applied to seeks only. Commands after the seek can execute freely, and like with 1., I don't see a reason why they couldn't. However, in some cases, seeks are supposed to be executed instantly, so queue_seek() needs an additional parameter to signal the need for immediate update. One nice thing is that commands like sub_seek automatically profit from the seek delay logic. On the other hand, hitting chapter seek multiple times still does not update the video on chapter boundaries (as it should be). Note that the main goal of this commit is actually simplification of the input processing logic and to allow all commands to be executed immediately.
2014-02-07 21:29:50 +00:00
// This combines consecutive seek requests.
void queue_seek(struct MPContext *mpctx, enum seek_type type, double amount,
player: handle seek delays differently The code removed from handle_input_and_seek_coalesce() did two things: 1. If there's a queued seek, stop accepting non-seek commands, and delay them to the next playloop iteration. 2. If a seek is executing (i.e. the seek was unqueued, and now it's trying to decode and display the first video frame), stop accepting seek commands (and in fact all commands that were queued after the first seek command). This logic is disabled if seeking started longer than 300ms ago. (To avoid starvation.) I'm not sure why 1. would be needed. It's still possible that a command immediately executed after a seek command sees a "seeking in progress" state, because it affects queued seeks only, and not seeks in progress. Drop this code, since it can easily lead to input starvation, and I'm not aware of any disadvantages. The logic in 2. is good to make seeking behave much better, as it guarantees that the video display is updated frequently. Keep the core idea, but implement it differently. Now this logic is applied to seeks only. Commands after the seek can execute freely, and like with 1., I don't see a reason why they couldn't. However, in some cases, seeks are supposed to be executed instantly, so queue_seek() needs an additional parameter to signal the need for immediate update. One nice thing is that commands like sub_seek automatically profit from the seek delay logic. On the other hand, hitting chapter seek multiple times still does not update the video on chapter boundaries (as it should be). Note that the main goal of this commit is actually simplification of the input processing logic and to allow all commands to be executed immediately.
2014-02-07 21:29:50 +00:00
int exact, bool immediate)
{
struct seek_params *seek = &mpctx->seek;
switch (type) {
case MPSEEK_RELATIVE:
player: handle seek delays differently The code removed from handle_input_and_seek_coalesce() did two things: 1. If there's a queued seek, stop accepting non-seek commands, and delay them to the next playloop iteration. 2. If a seek is executing (i.e. the seek was unqueued, and now it's trying to decode and display the first video frame), stop accepting seek commands (and in fact all commands that were queued after the first seek command). This logic is disabled if seeking started longer than 300ms ago. (To avoid starvation.) I'm not sure why 1. would be needed. It's still possible that a command immediately executed after a seek command sees a "seeking in progress" state, because it affects queued seeks only, and not seeks in progress. Drop this code, since it can easily lead to input starvation, and I'm not aware of any disadvantages. The logic in 2. is good to make seeking behave much better, as it guarantees that the video display is updated frequently. Keep the core idea, but implement it differently. Now this logic is applied to seeks only. Commands after the seek can execute freely, and like with 1., I don't see a reason why they couldn't. However, in some cases, seeks are supposed to be executed instantly, so queue_seek() needs an additional parameter to signal the need for immediate update. One nice thing is that commands like sub_seek automatically profit from the seek delay logic. On the other hand, hitting chapter seek multiple times still does not update the video on chapter boundaries (as it should be). Note that the main goal of this commit is actually simplification of the input processing logic and to allow all commands to be executed immediately.
2014-02-07 21:29:50 +00:00
seek->immediate |= immediate;
if (seek->type == MPSEEK_FACTOR)
return; // Well... not common enough to bother doing better
seek->amount += amount;
seek->exact = MPMAX(seek->exact, exact);
if (seek->type == MPSEEK_NONE)
seek->exact = exact;
if (seek->type == MPSEEK_ABSOLUTE)
return;
if (seek->amount == 0) {
*seek = (struct seek_params){ 0 };
return;
}
seek->type = MPSEEK_RELATIVE;
return;
case MPSEEK_ABSOLUTE:
case MPSEEK_FACTOR:
*seek = (struct seek_params) {
.type = type,
.amount = amount,
.exact = exact,
player: handle seek delays differently The code removed from handle_input_and_seek_coalesce() did two things: 1. If there's a queued seek, stop accepting non-seek commands, and delay them to the next playloop iteration. 2. If a seek is executing (i.e. the seek was unqueued, and now it's trying to decode and display the first video frame), stop accepting seek commands (and in fact all commands that were queued after the first seek command). This logic is disabled if seeking started longer than 300ms ago. (To avoid starvation.) I'm not sure why 1. would be needed. It's still possible that a command immediately executed after a seek command sees a "seeking in progress" state, because it affects queued seeks only, and not seeks in progress. Drop this code, since it can easily lead to input starvation, and I'm not aware of any disadvantages. The logic in 2. is good to make seeking behave much better, as it guarantees that the video display is updated frequently. Keep the core idea, but implement it differently. Now this logic is applied to seeks only. Commands after the seek can execute freely, and like with 1., I don't see a reason why they couldn't. However, in some cases, seeks are supposed to be executed instantly, so queue_seek() needs an additional parameter to signal the need for immediate update. One nice thing is that commands like sub_seek automatically profit from the seek delay logic. On the other hand, hitting chapter seek multiple times still does not update the video on chapter boundaries (as it should be). Note that the main goal of this commit is actually simplification of the input processing logic and to allow all commands to be executed immediately.
2014-02-07 21:29:50 +00:00
.immediate = immediate,
};
return;
case MPSEEK_NONE:
*seek = (struct seek_params){ 0 };
return;
}
abort();
}
void execute_queued_seek(struct MPContext *mpctx)
{
if (mpctx->seek.type) {
// Let explicitly imprecise seeks cancel precise seeks:
if (mpctx->hrseek_active && mpctx->seek.exact < 0)
mpctx->start_timestamp = -1e9;
player: handle seek delays differently The code removed from handle_input_and_seek_coalesce() did two things: 1. If there's a queued seek, stop accepting non-seek commands, and delay them to the next playloop iteration. 2. If a seek is executing (i.e. the seek was unqueued, and now it's trying to decode and display the first video frame), stop accepting seek commands (and in fact all commands that were queued after the first seek command). This logic is disabled if seeking started longer than 300ms ago. (To avoid starvation.) I'm not sure why 1. would be needed. It's still possible that a command immediately executed after a seek command sees a "seeking in progress" state, because it affects queued seeks only, and not seeks in progress. Drop this code, since it can easily lead to input starvation, and I'm not aware of any disadvantages. The logic in 2. is good to make seeking behave much better, as it guarantees that the video display is updated frequently. Keep the core idea, but implement it differently. Now this logic is applied to seeks only. Commands after the seek can execute freely, and like with 1., I don't see a reason why they couldn't. However, in some cases, seeks are supposed to be executed instantly, so queue_seek() needs an additional parameter to signal the need for immediate update. One nice thing is that commands like sub_seek automatically profit from the seek delay logic. On the other hand, hitting chapter seek multiple times still does not update the video on chapter boundaries (as it should be). Note that the main goal of this commit is actually simplification of the input processing logic and to allow all commands to be executed immediately.
2014-02-07 21:29:50 +00:00
/* If the user seeks continuously (keeps arrow key down)
* try to finish showing a frame from one location before doing
* another seek (which could lead to unchanging display). */
if (!mpctx->seek.immediate && mpctx->video_status < STATUS_READY &&
player: handle seek delays differently The code removed from handle_input_and_seek_coalesce() did two things: 1. If there's a queued seek, stop accepting non-seek commands, and delay them to the next playloop iteration. 2. If a seek is executing (i.e. the seek was unqueued, and now it's trying to decode and display the first video frame), stop accepting seek commands (and in fact all commands that were queued after the first seek command). This logic is disabled if seeking started longer than 300ms ago. (To avoid starvation.) I'm not sure why 1. would be needed. It's still possible that a command immediately executed after a seek command sees a "seeking in progress" state, because it affects queued seeks only, and not seeks in progress. Drop this code, since it can easily lead to input starvation, and I'm not aware of any disadvantages. The logic in 2. is good to make seeking behave much better, as it guarantees that the video display is updated frequently. Keep the core idea, but implement it differently. Now this logic is applied to seeks only. Commands after the seek can execute freely, and like with 1., I don't see a reason why they couldn't. However, in some cases, seeks are supposed to be executed instantly, so queue_seek() needs an additional parameter to signal the need for immediate update. One nice thing is that commands like sub_seek automatically profit from the seek delay logic. On the other hand, hitting chapter seek multiple times still does not update the video on chapter boundaries (as it should be). Note that the main goal of this commit is actually simplification of the input processing logic and to allow all commands to be executed immediately.
2014-02-07 21:29:50 +00:00
mp_time_sec() - mpctx->start_timestamp < 0.3)
return;
mp_seek(mpctx, mpctx->seek, false);
mpctx->seek = (struct seek_params){0};
}
}
double get_time_length(struct MPContext *mpctx)
{
struct demuxer *demuxer = mpctx->demuxer;
if (!demuxer)
return 0;
if (mpctx->timeline)
return mpctx->timeline[mpctx->num_timeline_parts].start;
double len = demuxer_get_time_length(demuxer);
if (len >= 0)
return len;
// Unknown
return 0;
}
/* If there are timestamps from stream level then use those (for example
* DVDs can have consistent times there while the MPEG-level timestamps
* reset). */
double get_current_time(struct MPContext *mpctx)
{
struct demuxer *demuxer = mpctx->demuxer;
if (!demuxer)
return 0;
if (mpctx->playback_pts != MP_NOPTS_VALUE)
return mpctx->playback_pts;
if (mpctx->last_seek_pts != MP_NOPTS_VALUE)
return mpctx->last_seek_pts;
return 0;
}
double get_playback_time(struct MPContext *mpctx)
{
double cur = get_current_time(mpctx);
double start = get_start_time(mpctx);
return cur >= start ? cur - start : cur;
}
// Return playback position in 0.0-1.0 ratio, or -1 if unknown.
double get_current_pos_ratio(struct MPContext *mpctx, bool use_range)
{
struct demuxer *demuxer = mpctx->demuxer;
if (!demuxer)
return -1;
double ans = -1;
double start = get_start_time(mpctx);
double len = get_time_length(mpctx);
if (use_range) {
double startpos = rel_time_to_abs(mpctx, mpctx->opts->play_start);
double endpos = get_play_end_pts(mpctx);
if (endpos == MP_NOPTS_VALUE || endpos > start + len)
endpos = start + len;
if (startpos == MP_NOPTS_VALUE || startpos < start)
startpos = start;
if (endpos < startpos)
endpos = startpos;
start = startpos;
len = endpos - startpos;
}
double pos = get_current_time(mpctx);
if (len > 0 && !demuxer->ts_resets_possible) {
ans = MPCLAMP((pos - start) / len, 0, 1);
} else {
int64_t size;
if (demux_stream_control(demuxer, STREAM_CTRL_GET_SIZE, &size) > 0) {
if (size > 0 && demuxer->filepos >= 0)
ans = MPCLAMP(demuxer->filepos / (double)size, 0, 1);
}
}
if (use_range) {
if (mpctx->opts->play_frames > 0)
ans = MPMAX(ans, 1.0 -
mpctx->max_frames / (double) mpctx->opts->play_frames);
}
return ans;
}
int get_percent_pos(struct MPContext *mpctx)
{
int pos = get_current_pos_ratio(mpctx, false) * 100;
return MPCLAMP(pos, 0, 100);
}
// -2 is no chapters, -1 is before first chapter
int get_current_chapter(struct MPContext *mpctx)
{
if (!mpctx->num_chapters)
return -2;
double current_pts = get_current_time(mpctx);
int i;
for (i = 0; i < mpctx->num_chapters; i++)
if (current_pts < mpctx->chapters[i].start)
break;
return MPMAX(mpctx->last_chapter_seek, i - 1);
}
char *chapter_display_name(struct MPContext *mpctx, int chapter)
{
char *name = chapter_name(mpctx, chapter);
char *dname = name;
if (name) {
dname = talloc_asprintf(NULL, "(%d) %s", chapter + 1, name);
} else if (chapter < -1) {
dname = talloc_strdup(NULL, "(unavailable)");
} else {
int chapter_count = get_chapter_count(mpctx);
if (chapter_count <= 0)
dname = talloc_asprintf(NULL, "(%d)", chapter + 1);
else
dname = talloc_asprintf(NULL, "(%d) of %d", chapter + 1,
chapter_count);
}
if (dname != name)
talloc_free(name);
return dname;
}
// returns NULL if chapter name unavailable
char *chapter_name(struct MPContext *mpctx, int chapter)
{
if (chapter < 0 || chapter >= mpctx->num_chapters)
return NULL;
return talloc_strdup(NULL, mpctx->chapters[chapter].name);
}
// returns the start of the chapter in seconds (-1 if unavailable)
double chapter_start_time(struct MPContext *mpctx, int chapter)
{
if (chapter == -1)
return get_start_time(mpctx);
if (chapter >= 0 && chapter < mpctx->num_chapters)
return mpctx->chapters[chapter].start;
return MP_NOPTS_VALUE;
}
int get_chapter_count(struct MPContext *mpctx)
{
return mpctx->num_chapters;
}
// Seek to a given chapter. Queues the seek.
bool mp_seek_chapter(struct MPContext *mpctx, int chapter)
{
int num = get_chapter_count(mpctx);
if (num == 0)
return false;
if (chapter < -1 || chapter >= num)
return false;
mpctx->last_chapter_seek = -2;
double pts = chapter_start_time(mpctx, chapter);
if (pts == MP_NOPTS_VALUE)
return false;
player: handle seek delays differently The code removed from handle_input_and_seek_coalesce() did two things: 1. If there's a queued seek, stop accepting non-seek commands, and delay them to the next playloop iteration. 2. If a seek is executing (i.e. the seek was unqueued, and now it's trying to decode and display the first video frame), stop accepting seek commands (and in fact all commands that were queued after the first seek command). This logic is disabled if seeking started longer than 300ms ago. (To avoid starvation.) I'm not sure why 1. would be needed. It's still possible that a command immediately executed after a seek command sees a "seeking in progress" state, because it affects queued seeks only, and not seeks in progress. Drop this code, since it can easily lead to input starvation, and I'm not aware of any disadvantages. The logic in 2. is good to make seeking behave much better, as it guarantees that the video display is updated frequently. Keep the core idea, but implement it differently. Now this logic is applied to seeks only. Commands after the seek can execute freely, and like with 1., I don't see a reason why they couldn't. However, in some cases, seeks are supposed to be executed instantly, so queue_seek() needs an additional parameter to signal the need for immediate update. One nice thing is that commands like sub_seek automatically profit from the seek delay logic. On the other hand, hitting chapter seek multiple times still does not update the video on chapter boundaries (as it should be). Note that the main goal of this commit is actually simplification of the input processing logic and to allow all commands to be executed immediately.
2014-02-07 21:29:50 +00:00
queue_seek(mpctx, MPSEEK_ABSOLUTE, pts, 0, true);
mpctx->last_chapter_seek = chapter;
mpctx->last_chapter_pts = pts;
return true;
}
static void handle_osd_redraw(struct MPContext *mpctx)
{
if (!mpctx->video_out || !mpctx->video_out->config_ok)
return;
// If we're playing normally, let OSD be redrawn naturally as part of
// video display.
if (!mpctx->paused) {
if (mpctx->sleeptime < 0.1 && mpctx->video_status == STATUS_PLAYING)
return;
}
// Don't redraw immediately during a seek (makes it significantly slower).
if (mp_time_sec() - mpctx->start_timestamp < 0.1)
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
bool want_redraw = osd_query_and_reset_want_redraw(mpctx->osd);
if (!want_redraw)
return;
vo_redraw(mpctx->video_out);
mpctx->sleeptime = 0;
}
static void handle_pause_on_low_cache(struct MPContext *mpctx)
{
struct MPOpts *opts = mpctx->opts;
if (!mpctx->demuxer)
return;
int64_t fill = -1;
demux_stream_control(mpctx->demuxer, STREAM_CTRL_GET_CACHE_FILL, &fill);
int cache_kb = fill > 0 ? (fill + 1023) / 1024 : -1;
bool idle = mp_get_cache_idle(mpctx);
if (mpctx->paused && mpctx->paused_for_cache) {
if (cache_kb < 0 || cache_kb >= opts->stream_cache_unpause || idle) {
mpctx->paused_for_cache = false;
if (!opts->pause)
unpause_player(mpctx);
}
mpctx->sleeptime = MPMIN(mpctx->sleeptime, 0.2);
} else {
if (cache_kb >= 0 && cache_kb <= opts->stream_cache_pause && !idle &&
opts->stream_cache_pause < opts->stream_cache_unpause)
{
bool prev_paused_user = opts->pause;
pause_player(mpctx);
mpctx->paused_for_cache = true;
opts->pause = prev_paused_user;
}
}
// Also update cache properties.
if (cache_kb > 0 || mpctx->next_cache_update > 0) {
double now = mp_time_sec();
if (mpctx->next_cache_update <= now) {
mpctx->next_cache_update = cache_kb > 0 ? now + 0.25 : 0;
mp_notify(mpctx, MP_EVENT_CACHE_UPDATE, NULL);
}
if (mpctx->next_cache_update > 0) {
mpctx->sleeptime =
MPMIN(mpctx->sleeptime, mpctx->next_cache_update - now);
}
}
}
static void handle_heartbeat_cmd(struct MPContext *mpctx)
{
struct MPOpts *opts = mpctx->opts;
if (opts->heartbeat_cmd && !mpctx->paused && mpctx->video_out) {
double now = mp_time_sec();
if (mpctx->next_heartbeat <= now) {
mpctx->next_heartbeat = now + opts->heartbeat_interval;
system(opts->heartbeat_cmd);
}
mpctx->sleeptime = MPMIN(mpctx->sleeptime, mpctx->next_heartbeat - now);
}
}
static void handle_cursor_autohide(struct MPContext *mpctx)
{
struct MPOpts *opts = mpctx->opts;
struct vo *vo = mpctx->video_out;
if (!vo)
return;
bool mouse_cursor_visible = mpctx->mouse_cursor_visible;
double now = mp_time_sec();
unsigned mouse_event_ts = mp_input_get_mouse_event_counter(mpctx->input);
if (mpctx->mouse_event_ts != mouse_event_ts) {
mpctx->mouse_event_ts = mouse_event_ts;
mpctx->mouse_timer = now + opts->cursor_autohide_delay / 1000.0;
mouse_cursor_visible = true;
}
if (mpctx->mouse_timer > now) {
mpctx->sleeptime = MPMIN(mpctx->sleeptime, mpctx->mouse_timer - now);
} else {
mouse_cursor_visible = false;
}
if (opts->cursor_autohide_delay == -1)
mouse_cursor_visible = true;
if (opts->cursor_autohide_delay == -2)
mouse_cursor_visible = false;
if (opts->cursor_autohide_fs && !opts->vo.fullscreen)
mouse_cursor_visible = true;
if (mouse_cursor_visible != mpctx->mouse_cursor_visible)
vo_control(vo, VOCTRL_SET_CURSOR_VISIBILITY, &mouse_cursor_visible);
mpctx->mouse_cursor_visible = mouse_cursor_visible;
}
static void handle_input_and_seek_coalesce(struct MPContext *mpctx)
{
mp_cmd_t *cmd;
while ((cmd = mp_input_get_cmd(mpctx->input, 0, 1)) != NULL) {
mp_dispatch_queue_process(mpctx->dispatch, 0);
cmd = mp_input_get_cmd(mpctx->input, 0, 0);
run_command(mpctx, cmd);
mp_cmd_free(cmd);
if (mpctx->stop_play)
break;
}
mp_dispatch_queue_process(mpctx->dispatch, 0);
}
void add_frame_pts(struct MPContext *mpctx, double pts)
{
if (pts == MP_NOPTS_VALUE || mpctx->hrseek_framedrop) {
mpctx->vo_pts_history_seek_ts++; // mark discontinuity
return;
}
if (mpctx->vo_pts_history_pts[0] == pts) // may be called multiple times
return;
for (int n = MAX_NUM_VO_PTS - 1; n >= 1; n--) {
mpctx->vo_pts_history_seek[n] = mpctx->vo_pts_history_seek[n - 1];
mpctx->vo_pts_history_pts[n] = mpctx->vo_pts_history_pts[n - 1];
}
mpctx->vo_pts_history_seek[0] = mpctx->vo_pts_history_seek_ts;
mpctx->vo_pts_history_pts[0] = pts;
}
static double find_previous_pts(struct MPContext *mpctx, double pts)
{
for (int n = 0; n < MAX_NUM_VO_PTS - 1; n++) {
if (pts == mpctx->vo_pts_history_pts[n] &&
mpctx->vo_pts_history_seek[n] != 0 &&
mpctx->vo_pts_history_seek[n] == mpctx->vo_pts_history_seek[n + 1])
{
return mpctx->vo_pts_history_pts[n + 1];
}
}
return MP_NOPTS_VALUE;
}
static double get_last_frame_pts(struct MPContext *mpctx)
{
if (mpctx->vo_pts_history_seek[0] == mpctx->vo_pts_history_seek_ts)
return mpctx->vo_pts_history_pts[0];
return MP_NOPTS_VALUE;
}
static void handle_backstep(struct MPContext *mpctx)
{
if (!mpctx->backstep_active)
return;
double current_pts = mpctx->last_vo_pts;
mpctx->backstep_active = false;
if (mpctx->d_video && current_pts != MP_NOPTS_VALUE) {
double seek_pts = find_previous_pts(mpctx, current_pts);
if (seek_pts != MP_NOPTS_VALUE) {
player: handle seek delays differently The code removed from handle_input_and_seek_coalesce() did two things: 1. If there's a queued seek, stop accepting non-seek commands, and delay them to the next playloop iteration. 2. If a seek is executing (i.e. the seek was unqueued, and now it's trying to decode and display the first video frame), stop accepting seek commands (and in fact all commands that were queued after the first seek command). This logic is disabled if seeking started longer than 300ms ago. (To avoid starvation.) I'm not sure why 1. would be needed. It's still possible that a command immediately executed after a seek command sees a "seeking in progress" state, because it affects queued seeks only, and not seeks in progress. Drop this code, since it can easily lead to input starvation, and I'm not aware of any disadvantages. The logic in 2. is good to make seeking behave much better, as it guarantees that the video display is updated frequently. Keep the core idea, but implement it differently. Now this logic is applied to seeks only. Commands after the seek can execute freely, and like with 1., I don't see a reason why they couldn't. However, in some cases, seeks are supposed to be executed instantly, so queue_seek() needs an additional parameter to signal the need for immediate update. One nice thing is that commands like sub_seek automatically profit from the seek delay logic. On the other hand, hitting chapter seek multiple times still does not update the video on chapter boundaries (as it should be). Note that the main goal of this commit is actually simplification of the input processing logic and to allow all commands to be executed immediately.
2014-02-07 21:29:50 +00:00
queue_seek(mpctx, MPSEEK_ABSOLUTE, seek_pts, 2, true);
} else {
double last = get_last_frame_pts(mpctx);
if (last != MP_NOPTS_VALUE && last >= current_pts &&
mpctx->backstep_start_seek_ts != mpctx->vo_pts_history_seek_ts)
{
MP_ERR(mpctx, "Backstep failed.\n");
player: handle seek delays differently The code removed from handle_input_and_seek_coalesce() did two things: 1. If there's a queued seek, stop accepting non-seek commands, and delay them to the next playloop iteration. 2. If a seek is executing (i.e. the seek was unqueued, and now it's trying to decode and display the first video frame), stop accepting seek commands (and in fact all commands that were queued after the first seek command). This logic is disabled if seeking started longer than 300ms ago. (To avoid starvation.) I'm not sure why 1. would be needed. It's still possible that a command immediately executed after a seek command sees a "seeking in progress" state, because it affects queued seeks only, and not seeks in progress. Drop this code, since it can easily lead to input starvation, and I'm not aware of any disadvantages. The logic in 2. is good to make seeking behave much better, as it guarantees that the video display is updated frequently. Keep the core idea, but implement it differently. Now this logic is applied to seeks only. Commands after the seek can execute freely, and like with 1., I don't see a reason why they couldn't. However, in some cases, seeks are supposed to be executed instantly, so queue_seek() needs an additional parameter to signal the need for immediate update. One nice thing is that commands like sub_seek automatically profit from the seek delay logic. On the other hand, hitting chapter seek multiple times still does not update the video on chapter boundaries (as it should be). Note that the main goal of this commit is actually simplification of the input processing logic and to allow all commands to be executed immediately.
2014-02-07 21:29:50 +00:00
queue_seek(mpctx, MPSEEK_ABSOLUTE, current_pts, 2, true);
} else if (!mpctx->hrseek_active) {
MP_VERBOSE(mpctx, "Start backstep indexing.\n");
// Force it to index the video up until current_pts.
// The whole point is getting frames _before_ that PTS,
// so apply an arbitrary offset. (In theory the offset
// has to be large enough to reach the previous frame.)
mp_seek(mpctx, (struct seek_params){
.type = MPSEEK_ABSOLUTE,
.amount = current_pts - 1.0,
}, false);
// Don't leave hr-seek mode. If all goes right, hr-seek
// mode is cancelled as soon as the frame before
// current_pts is found during hr-seeking.
// Note that current_pts should be part of the index,
// otherwise we can't find the previous frame, so set the
// seek target an arbitrary amount of time after it.
if (mpctx->hrseek_active) {
mpctx->hrseek_pts = current_pts + 10.0;
mpctx->hrseek_framedrop = false;
mpctx->backstep_active = true;
}
} else {
mpctx->backstep_active = true;
}
}
}
}
static void handle_sstep(struct MPContext *mpctx)
{
struct MPOpts *opts = mpctx->opts;
if (mpctx->stop_play || !mpctx->restart_complete)
return;
if (opts->step_sec > 0 && !mpctx->paused) {
set_osd_function(mpctx, OSD_FFW);
player: handle seek delays differently The code removed from handle_input_and_seek_coalesce() did two things: 1. If there's a queued seek, stop accepting non-seek commands, and delay them to the next playloop iteration. 2. If a seek is executing (i.e. the seek was unqueued, and now it's trying to decode and display the first video frame), stop accepting seek commands (and in fact all commands that were queued after the first seek command). This logic is disabled if seeking started longer than 300ms ago. (To avoid starvation.) I'm not sure why 1. would be needed. It's still possible that a command immediately executed after a seek command sees a "seeking in progress" state, because it affects queued seeks only, and not seeks in progress. Drop this code, since it can easily lead to input starvation, and I'm not aware of any disadvantages. The logic in 2. is good to make seeking behave much better, as it guarantees that the video display is updated frequently. Keep the core idea, but implement it differently. Now this logic is applied to seeks only. Commands after the seek can execute freely, and like with 1., I don't see a reason why they couldn't. However, in some cases, seeks are supposed to be executed instantly, so queue_seek() needs an additional parameter to signal the need for immediate update. One nice thing is that commands like sub_seek automatically profit from the seek delay logic. On the other hand, hitting chapter seek multiple times still does not update the video on chapter boundaries (as it should be). Note that the main goal of this commit is actually simplification of the input processing logic and to allow all commands to be executed immediately.
2014-02-07 21:29:50 +00:00
queue_seek(mpctx, MPSEEK_RELATIVE, opts->step_sec, 0, true);
}
if (mpctx->video_status >= STATUS_DRAINING) {
if (mpctx->max_frames >= 0)
mpctx->stop_play = PT_NEXT_ENTRY;
if (mpctx->step_frames > 0 && !mpctx->paused)
pause_player(mpctx);
}
}
static void handle_loop_file(struct MPContext *mpctx)
{
struct MPOpts *opts = mpctx->opts;
if (opts->loop_file && mpctx->stop_play == AT_END_OF_FILE) {
set_osd_function(mpctx, OSD_FFW);
queue_seek(mpctx, MPSEEK_ABSOLUTE, get_start_time(mpctx), 0, true);
}
}
static void handle_keep_open(struct MPContext *mpctx)
{
struct MPOpts *opts = mpctx->opts;
if (opts->keep_open && mpctx->stop_play == AT_END_OF_FILE) {
mpctx->stop_play = KEEP_PLAYING;
mpctx->playback_pts = mpctx->last_vo_pts;
if (!mpctx->opts->pause)
pause_player(mpctx);
}
}
static void handle_chapter_change(struct MPContext *mpctx)
{
int chapter = get_current_chapter(mpctx);
if (chapter != mpctx->last_chapter) {
mpctx->last_chapter = chapter;
mp_notify(mpctx, MPV_EVENT_CHAPTER_CHANGE, NULL);
}
}
// Execute a forceful refresh of the VO window, if it hasn't had a valid frame
// for a while. The problem is that a VO with no valid frame (vo->hasframe==0)
// doesn't redraw video and doesn't OSD interaction. So screw it, hard.
void handle_force_window(struct MPContext *mpctx, bool reconfig)
{
// Don't interfere with real video playback
if (mpctx->d_video)
return;
struct vo *vo = mpctx->video_out;
if (!vo)
return;
if (!vo->config_ok || reconfig) {
MP_INFO(mpctx, "Creating non-video VO window.\n");
// Pick whatever works
int config_format = 0;
for (int fmt = IMGFMT_START; fmt < IMGFMT_END; fmt++) {
if (vo->driver->query_format(vo, fmt)) {
config_format = fmt;
break;
}
}
int w = 960;
int h = 480;
struct mp_image_params p = {
.imgfmt = config_format,
.w = w, .h = h,
.d_w = w, .d_h = h,
};
vo_reconfig(vo, &p, 0);
vo_redraw(vo);
mp_notify(mpctx, MPV_EVENT_VIDEO_RECONFIG, NULL);
}
}
// Potentially needed by some Lua scripts, which assume TICK always comes.
static void handle_dummy_ticks(struct MPContext *mpctx)
{
if (mpctx->video_status == STATUS_EOF || mpctx->paused) {
if (mp_time_sec() - mpctx->last_idle_tick > 0.5) {
mpctx->last_idle_tick = mp_time_sec();
mp_notify(mpctx, MPV_EVENT_TICK, NULL);
}
}
}
void run_playloop(struct MPContext *mpctx)
{
struct MPOpts *opts = mpctx->opts;
double endpts = get_play_end_pts(mpctx);
bool end_is_new_segment = false;
#if HAVE_ENCODING
if (encode_lavc_didfail(mpctx->encode_lavc_ctx)) {
mpctx->stop_play = PT_QUIT;
return;
}
#endif
update_demuxer_properties(mpctx);
if (mpctx->timeline) {
double end = mpctx->timeline[mpctx->timeline_part + 1].start;
if (endpts == MP_NOPTS_VALUE || end < endpts) {
end_is_new_segment = true;
endpts = end;
}
}
if (opts->chapterrange[1] > 0) {
double end = chapter_start_time(mpctx, opts->chapterrange[1]);
if (end != MP_NOPTS_VALUE && (endpts == MP_NOPTS_VALUE || end < endpts))
endpts = end;
}
handle_cursor_autohide(mpctx);
handle_heartbeat_cmd(mpctx);
fill_audio_out_buffers(mpctx, endpts);
write_video(mpctx, endpts);
// We always make sure audio and video buffers are filled before actually
// starting playback. This code handles starting them at the same time.
if (mpctx->audio_status >= STATUS_READY &&
mpctx->video_status >= STATUS_READY)
{
if (mpctx->video_status == STATUS_READY) {
mpctx->video_status = STATUS_PLAYING;
get_relative_time(mpctx);
mpctx->sleeptime = 0;
}
if (mpctx->audio_status == STATUS_READY)
fill_audio_out_buffers(mpctx, endpts); // actually play prepared buffer
if (!mpctx->restart_complete) {
mpctx->hrseek_active = false;
mp_notify(mpctx, MPV_EVENT_PLAYBACK_RESTART, NULL);
mpctx->restart_complete = true;
if (opts->playing_msg && !mpctx->playing_msg_shown) {
mpctx->playing_msg_shown = true;
char *msg =
mp_property_expand_escaped_string(mpctx, opts->playing_msg);
MP_INFO(mpctx, "%s\n", msg);
talloc_free(msg);
}
}
}
if (mpctx->video_status == STATUS_EOF &&
mpctx->audio_status >= STATUS_PLAYING)
{
double a_pos = 0;
if (mpctx->d_audio)
a_pos = playing_audio_pts(mpctx);
mpctx->playback_pts = a_pos;
}
handle_dummy_ticks(mpctx);
update_osd_msg(mpctx);
update_subtitles(mpctx);
/* If we're paused, don't end playback yet. But if video is enabled, is EOF,
* and we don't have a video frame, then the user probably seeked outside
* of the video, and we want to quit. */
bool prevent_eof = mpctx->paused;
if (mpctx->d_video && mpctx->video_status == STATUS_EOF)
prevent_eof &= mpctx->video_out && vo_has_frame(mpctx->video_out);
/* Handles terminating on end of playback (or switching to next segment).
*
* It's possible for the user to simultaneously switch both audio
* and video streams to "disabled" at runtime. Handle this by waiting
* rather than immediately stopping playback due to EOF.
*/
if ((mpctx->d_audio || mpctx->d_video) && !prevent_eof &&
mpctx->audio_status == STATUS_EOF &&
mpctx->video_status == STATUS_EOF)
{
if (end_is_new_segment) {
mp_seek(mpctx, (struct seek_params){
.type = MPSEEK_ABSOLUTE,
.amount = mpctx->timeline[mpctx->timeline_part+1].start
}, true);
} else
mpctx->stop_play = AT_END_OF_FILE;
}
mp_handle_nav(mpctx);
handle_keep_open(mpctx);
handle_sstep(mpctx);
handle_loop_file(mpctx);
if (mpctx->stop_play)
return;
handle_osd_redraw(mpctx);
if (mpctx->sleeptime > 0) {
MP_STATS(mpctx, "start sleep");
mp_input_get_cmd(mpctx->input, mpctx->sleeptime * 1000, true);
MP_STATS(mpctx, "end sleep");
}
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
mpctx->sleeptime = 100.0; // infinite for all practical purposes
handle_pause_on_low_cache(mpctx);
handle_input_and_seek_coalesce(mpctx);
handle_backstep(mpctx);
handle_chapter_change(mpctx);
handle_force_window(mpctx, false);
execute_queued_seek(mpctx);
if (mpctx->opts->use_terminal)
getch2_poll();
}
// Waiting for the slave master to send us a new file to play.
void idle_loop(struct MPContext *mpctx)
{
// ================= idle loop (STOP state) =========================
bool need_reinit = true;
while (mpctx->opts->player_idle_mode && !mpctx->playlist->current
&& mpctx->stop_play != PT_QUIT)
{
mpctx->video_status = STATUS_EOF;
mpctx->audio_status = STATUS_EOF;
if (need_reinit) {
mp_notify(mpctx, MPV_EVENT_IDLE, NULL);
handle_force_window(mpctx, true);
}
need_reinit = false;
handle_dummy_ticks(mpctx);
int uninit = INITIALIZED_AO;
if (!mpctx->opts->force_vo)
uninit |= INITIALIZED_VO;
uninit_player(mpctx, uninit);
handle_force_window(mpctx, false);
update_osd_msg(mpctx);
handle_osd_redraw(mpctx);
2014-07-22 17:33:24 +00:00
mp_cmd_t *cmd = mp_input_get_cmd(mpctx->input, mpctx->sleeptime * 1000,
false);
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
mpctx->sleeptime = 100.0;
if (cmd)
run_command(mpctx, cmd);
mp_cmd_free(cmd);
mp_dispatch_queue_process(mpctx->dispatch, 0);
if (mpctx->opts->use_terminal)
getch2_poll();
}
}