Instead of doing it on every seek (libavcodec calls get_format on every
seek), reinitialize the decoder only if the video resolution changes.
Note that this may be relatively naive, since we e.g. (or: in
particular) don't check for profile changes. But it's not worse than the
state before the get_format change, and at least it paints over the
current vaapi breakage (issue #646).
This "sometimes" crashed when seeking. The fault apparently lies in
libavcodec: the decoder returns an unreferenced frame! This is
completely insane, but somehow I'm apparently still expected to
work this around. As a reaction, I will drop Libav 9 support in the
next commit. (While this commit will go into release/0.3.)
Apparently the "right" place to initialize the hardware decoder is in
the libavcodec get_format callback.
This doesn't change vda.c and vdpau_old.c, because I don't have OSX, and
vdpau_old.c is probably going to be removed soon (if Libav ever manages
to release Libav 10). So for now the init_decoder callback added with
this commit is optional.
This also means vdpau.c and vaapi.c don't have to manage and check the
image parameters anymore.
This change is probably needed for when libavcodec VDA supports gets a
new iteration of its API.
Like with the previous commit, this is probably not needed, but it's
unclear whether that really is the case. Most likely, it used to be
needed by some demuxer, and now the only demuxer left that could
_possibly_ trigger this is demux_mkv.c.
Note that mjpeg is the only decoder that reads the extra_huff option,
and nothing in libavformat actually sets the option. So maybe it's
fundamentally not needed anymore.
This case can't happen with the normal realvideo codepath in
demux_mkv.c, because the code would errors out if the extradata is too
small, and everything would be broken anyway in the case the vd_lavc.c
condition is actually triggered.
It still might happen with VfW-muxed realvideo in Matroska, though.
Basically, I'm hoping this doesn't matter anyway, and that the vd_lavc.c
code was for other old demuxers, like demux_avi or demux_rm. Following
the commit history, it's not really clear for what demuxer this code
was added.
Set the flag CODEC_FLAG_OUTPUT_CORRUPT by default. Note that there is
also CODEC_FLAG2_SHOW_ALL, which is older, but this seems to be ffmpeg
only.
Note that whether you want this enabled depends on the user. Some might
prefer that only good frames are output, while others want the decoder
to try as hard as possible to output _anything_. Since mplayer/mpv is
rather the kind of player that tries hard instead of being "clever", set
the new default to override libavcodec's default.
A nice way to test this is switching video tracks. Since mpv doesn't
wait for the next key frame, it'll start feeding the decoder with a
packet from the middle of the stream.
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.
The tmsg stuff was for the internal gettext() based translation system,
which nobody ever attempted to use and thus was removed. mp_gtext() and
set_osd_tmsg() were also for this.
mp_dbg was once enabled in debug mode only, but since we have log level
for enabling debug messages, it seems utterly useless.
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.
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.
If the timebase is set, it's used for converting the packet timestamps.
Otherwise, the previous method of reinterpret-casting the mpv style
double timestamps to libavcodec style int64_t timestamps is used.
Also replace the kind of awkward mp_get_av_frame_pkt_ts() function by
mp_pts_from_av(), which simply converts timestamps in a way the old
function did. (Plus it takes a timebase parameter, similar to the
addition to mp_set_av_packet().)
Note that this should not change anything yet. The code in ad_lavc.c and
vd_lavc.c passes NULL for the timebase parameters. We could set
AVCodecContext.pkt_timebase and use that if we want to give libavcodec
"proper" timestamps.
This could be important for ad_lavc.c: some codecs (opus, probably mp3
and aac too) have weird requirements about doing decoding preroll on the
container level, and thus require adjusting the audio start timestamps
in some cases. libavcodec doesn't tell us how much was skipped, so we
either get shifted timestamps (by the length of the skipped data), or we
give it proper timestamps. (Note: libavcodec interprets or changes
timestamps only if pkt_timebase is set, which by default it is not.)
This would require selecting a timebase though, so I feel uncomfortable
with the idea. At least this change paves the way, and will allow some
testing.
PIX_FMT_* -> AV_PIX_FMT_* (except some pixdesc constants)
enum PixelFormat -> enum AVPixelFormat
Losen some version checks in certain newer pixel formats.
av_pix_fmt_descriptors -> av_pix_fmt_desc_get
This removes support for FFmpeg 1.0.x, which is even older than
Libav 9.x. Support for it probably was already broken, and its
libswresample was rejected by our build system anyway because it's
broken.
Mostly untested; it does compile with Libav 9.9.
Refactor the PTS handling code to make it cleaner, and to separate the
bits that use PTS sorting.
Add a heuristic to fall back to DTS if the PTS us non-monotonic. This
code is based on what FFmpeg/Libav use for ffplay/avplay and also
best_effort_timestamp (which is only in FFmpeg). Basically, this 1. just
uses the DTS if PTS is unset, and 2. ignores PTS entirely if PTS is non-
monotonic, but DTS is sorted.
The code is pretty much the same as in Libav [1]. I'm not sure if all of
it is really needed, or if it does more than what the paragraph above
mentions. But maybe it's fine to cargo-cult this.
This heuristic fixes playback of mpeg4 in ogm, which returns packets
with PTS==DTS, even though the PTS timestamps should follow codec
reordering. This is probably a libavformat demuxer bug, but good luck
trying to fix it.
The way vd_lavc.c returns the frame PTS and DTS to dec_video.c is a bit
inelegant, but maybe better than trying to mess the PTS back into the
decoder callback again.
[1] https://git.libav.org/?p=libav.git;a=blob;f=cmdutils.c;h=3f1c667075724c5cde69d840ed5ed7d992898334;hb=fa515c2088e1d082d45741bbd5c05e13b0500804#l1431
These used the suffix _resync_stream, which is a bit misleading. Nothing
gets "resynchronized", they really just reset state.
(Some audio decoders actually used to "resync" by reading packets for
resuming playback, but that's not the case anymore.)
Also move the function in dec_video.c to the top of the file.
Having the DTS directly can be useful for restoring PTS values.
The avi file format doesn't actually store PTS values, just DTS. An
older hack explicitly exported the DTS as PTS (ignoring the [I assume]
genpts generated non-sense PTS), which is not necessary anymore due to
this change.
Instead of passing the PTS as separate field, pass it as part of the
usual data structures. Basically, this removes strange artifacts from
the API. (It's not finished, though: the final decoded PTS goes through
strange paths, and filter_video() finally overwrites the decoded
mp_image's pts field with it.)
We also stop using libavcodec's reordered_opaque fields, and use
AVPacket.pts and AVFrame.pkt_pts. This is slightly unorthodox, because
these pts fields are not "really" opaque anymore, yet we treat them as
such. But the end result should be the same, and reordered_opaque is
marked as partially deprecated (it's not clear whether it's really
deprecated).
When mpv is started with some video filters set (--vf is used), and
hardware decoding is requested, and hardware decoding would be possible,
but is prevented due to video filters that accept software formats only,
the fallback didn't work properly sometimes.
This fallback works rather violently: it tries to initialize the filter
chain, and if it fails it throws away the frame decoded using the
hardware, and retries with software. The case that didn't work was when
decoding the current packet didn't immediately lead to a new frame. Then
the filter chain wouldn't be reinitialized, and the playloop would stop
playback as soon as it encounters the error flag.
Fix this by resetting the filter error flag (back to "uninitialized"),
which is a rather violent, but somewhat working solution.
The fallback in general should perhaps be cleaned up later.
Now the actual decoder doesn't need to care about this anymore, and it's
handled in generic code instead. This simplifies vd_lavc.c, and in
particular we don't need to detect format changes in the old way
anymore.
The only reason why these structs were dynamically allocated was to
avoid recursive includes in stheader.h, which is (or was) a very central
file included by almost all other files. (If a struct is referenced via
a pointer type only, it can be forward referenced, and the definition of
the struct is not needed.) Now that they're out of stheader.h, this
difference doesn't matter anymore, and the code can be simplified.
Also sneak in some sanity checks.
This used to be needed to access the generic stream header from the
specific headers, which in turn was needed because the decoders had
access only to the specific headers. This is not the case anymore, so
this can finally be removed again.
Also move the "format" field from the specific headers to sh_stream.
This is similar to the sh_audio commit.
This is mostly cosmetic in nature, except that it also adds automatical
freeing of the decoder driver's state struct (which was in
sh_video->context, now in dec_video->priv).
Also remove all the stheader.h fields that are not needed anymore.
The existing code tried to remove the "extra" profile flags for h264.
FF_PROFILE_H264_INTRA doesn't matter for us at all, because it's set
only for profiles the vdpau/vaapi APIs don't support.
The FF_PROFILE_H264_CONSTRAINED flag on the other hand is added to
H264_BASELINE, except that it makes the file a real subset of H264_MAIN
and H264_HIGH. Removing that flag would select the BASELINE profile,
which appears to be rarely supported by hardware decoders. This means we
accidentally rejected perfectly hardware decodable files. Use MAIN for
it instead.
(vaapi has explicit support for CONSTRAINED_BASELINE, but it seems to be
a new thing, and is not reported as supported where I tried. So don't
bother to check it, and do the same as on vdpau.)
See github issue #204.
This removes "--hwdec=crystalhd".
I doubt anyone even tried to use this. But even if someone wants to
use it, the decoders can still be explicitly invoked with e.g.:
--vd=lavc:h264_crystalhd
The only advantage our special code provided was fallback to
software decoding. (But I'm not sure how the ffmpeg crystalhd
pseudo-decoder actually behaves.)
Removing this will allow some simplifications as soon as we don't need
vdpau_old.c anymore.
Most hardware decoding APIs provide some OpenGL interop. This allows
using vo_opengl, without having to read the video data back from GPU.
This requires adding a backend for each hardware decoding API. (Each
backend is an entry in gl_hwdec_vaglx[].) The backends expose video data
as a set of OpenGL textures.
Add infrastructure to support this. The next commit will add support for
VA-API.
The configure followed 5 different convetions of defines because the next guy
always wanted to introduce a new better way to uniform it[1]. For an
hypothetic feature 'hurr' you could have had:
* #define HAVE_HURR 1 / #undef HAVE_DURR
* #define HAVE_HURR / #undef HAVE_DURR
* #define CONFIG_HURR 1 / #undef CONFIG_DURR
* #define HAVE_HURR 1 / #define HAVE_DURR 0
* #define CONFIG_HURR 1 / #define CONFIG_DURR 0
All is now uniform and uses:
* #define HAVE_HURR 1
* #define HAVE_DURR 0
We like definining to 0 as opposed to `undef` bcause it can help spot typos
and is very helpful when doing big reorganizations in the code.
[1]: http://xkcd.com/927/ related
There are some Microsoft Windows symbols which are traditionally used by
the mplayer core, because it used to be convenient (avi was the big
format, using binary windows decoders made sense...). So these symbols
have the exact same definition as the Windows one, and if mplayer is
compiled on Windows, the symbols from windows.h are used.
This broke recently just because some files were shuffled around, and
the symbols defined in ms_hdr.h collided with windows.h ones. Since we
don't have windows binary decoders anymore, there's not the slightest
reason our symbols should have the same names. Rename them to reduce the
risk for collision, and to fix the recent regression.
Drop WAVEFORMATEXTENSIBLE, because it's mostly unused. ao_dsound defines
its own version if the windows headers don't define it, and ao_wasapi is
not available on systems where this symbol is missing.
Also reindent ms_hdr.h.
We had some code for checking profiles earlier, which was removed in
commits 2508f38 and adfb71b. These commits mentioned that (working) hw
decoding was sometimes prevented due to profile checking, but I can't
find the samples anymore that showed this behavior. Also, I changed my
opinion, and I think checking the profiles is something that should be
done for better fallback to software decoding behavior.
The checks roughly follow VLC's vdpau profile checks, although we do
not check codec levels. (VLC's profile checks aren't necessarily
completely correct, but they're a welcome help anyway.)
Add a --vd-lavc-check-hw-profile option, which skips the profile check.
We mixed the "old" AVFrame management functions (avcodec_alloc_frame,
avcodec_free_frame) with reference counting. This doesn't work
correctly; you must use av_frame_alloc and av_frame_free. Of course
ffmpeg doesn't warn us about the bad usage, but will just mess up
things silently. (Thanks a lot...)
While the alloc function seems to be 100% compatible, the free function
will do bad things, such as freeing memory that might still be
referenced by another frame. I didn't experience any actual bugs, but
maybe that was pure luck.
This commit adds the --force-window option, which will cause mpv always
to create a window when started. This can be useful when pretending that
mpv is a GUI application (which it isn't, but users pretend anyway), and
playing audio files would run mpv in the background without giving a
window to control it.
This doesn't actually create the window immediately: it only does so
only after initializing playback and when it is clear that there won't
be any actual video. This could be a problem when starting slow or
completely stuck network streams (mpv would remain frozen in the
background), or if video initialization somehow is stuck forever in
an in-between state (like when the decoder doesn't output a video
frame, but doesn't return an error either). Well, we can pretend only
so much that mpv is a GUI application.
Now writing -1 to the 'aspect' property resets the video to the auto
aspect ratio. Returning the aspect from the property becomes a bit more
complicated, because we still try to return the container aspect ratio
if no frame has been decoded yet.
This code is actually quite inefficient: it reuses the (slow, simple)
screenshot code. It uses an inefficient method to read the image
(vaGetImage() instead of vaDeriveImage()), allocates new memory for
each frame that is read, and it tries all image formats again each
time.
Also, in my tests it always picked NV12 as image format, which is not
ideal if you actually want to filter the video, and vo_xv can't handle
this format without conversion either.
However, a user confirmed that it worked for him, so everything is fine.
This will allow GPU read-back with process_image.
We have to restructure how init_vo() works. Instead of initializing the
VO before process_image is called, rename init_vo() to
update_image_params(), and let it update the params only. Then we really
initialize the VO after process_image.
As a consequence of these changes, already decoded hw frames are
correctly unreferenced if creation of the filter chain fails. This
could trigger assertions on VO uninitialization, because it's not
allowed to reference hw frames past VO lifetime.
In init_vo(), if sh->aspect is 0 or last_sample_aspect_ratio is set,
sh->aspect is overwritten. With software decoding fallback behaviour,
this makes the aspect ratio from container ignored since
last_sample_aspect_ratio is already set in first try with hardware
decoding.
This affects VOs which just reuse the mp_image from draw_image() to
return screenshots. The aspect of these images is never different
from the aspect the screenshots should be, so there's no reason to
adjust the aspect in these cases.
Other VOs still need it in order to restore the original image
attributes.
This requires some changes to the video filter code to make sure that
the aspect in the passed mp_images is consistent.
The changes in mplayer.c and vd_lavc.c are (probably) not strictly
needed for this commit, but contribute to consistency.
Decoding H264 using Video Decode Acceleration used the custom 'vda_h264_dec'
decoder in FFmpeg.
The Good: This new implementation has some advantages over the previous one:
- It works with Libav: vda_h264_dec never got into Libav since they prefer
client applications to use the hwaccel API.
- It is way more efficient: in my tests this implementation yields a
reduction of CPU usage of roughly ~50% compared to using `vda_h264_dec` and
~65-75% compared to h264 software decoding. This is mainly because
`vo_corevideo` was adapted to perform direct rendering of the
`CVPixelBufferRefs` created by the Video Decode Acceleration API Framework.
The Bad:
- `vo_corevideo` is required to use VDA decoding acceleration.
- only works with versions of ffmpeg/libav new enough (needs reference
refcounting). That is FFmpeg 2.0+ and Libav's git master currently.
The Ugly: VDA was hardcoded to use UYVY (2vuy) for the uploaded video texture.
One one end this makes the code simple since Apple's OpenGL implementation
actually supports this out of the box. It would be nice to support other
output image formats and choose the best format depending on the input, or at
least making it configurable. My tests indicate that CPU usage actually
increases with a 420p IMGFMT output which is not what I would have expected.
NOTE: There is a small memory leak with old versions of FFmpeg and with Libav
since the CVPixelBufferRef is not automatically released when the AVFrame is
deallocated. This can cause leaks inside libavcodec for decoded frames that
are discarded before mpv wraps them inside a refcounted mp_image (this only
happens on seeks).
For frames that enter mpv's refcounting facilities, this is not a problem
since we rewrap the CVPixelBufferRef in our mp_image that properly forwards
CVPixelBufferRetain/CvPixelBufferRelease calls to the underying
CVPixelBufferRef.
So, for FFmpeg use something more recent than `b3d63995` for Libav the patch
was posted to the dev ML in July and in review since, apparently, the proposed
fix is rather hacky.
About this issue, it would be better if the surfaces could be allocated
with the real size, and the vdpau video mixer could be created with that
size as well. That would be a bit hard, because the real surface size
had to be communicated to vdpau. So I'm going with this solution. vaapi
seems to be fine with either surface size, so there's hopefully no
problem.
Instead of passing AVFrame. This also moves the mysterious logic about
the size of the allocated image to common code, instead of duplicating
it everywhere.
wm4
Stefano Pigozzi
xylosper