xine-lib/src, branch master xine-lib git mirror ff_video_decoder: refcount AVFrames. 2015-08-08T14:36:25+00:00 Torsten Jager t.jager@gmx.de 2015-08-08T14:36:25+00:00 e6a81be036e9170fffe519c6b87c9d7f0facb672 And do flush via a separate AVFrame. Maybe this fixes some rare race. BTW. libavcodec 56 left an ambigous impression to me. It does now respect flush requests even in multithread mode. Good. On the flipside, that "too many DR1 frames in MPEG-TS" issue returned. The very same stream needed between 7 and 17 frames, depending on the start time of playback :-/ Also, the future of MPEG4 style postprocessing seems uncertain again. 
And do flush via a separate AVFrame.
Maybe this fixes some rare race.

BTW. libavcodec 56 left an ambigous impression to me.

It does now respect flush requests even in multithread mode. Good.

On the flipside, that "too many DR1 frames in MPEG-TS" issue returned.
The very same stream needed between 7 and 17 frames, depending on the
start time of playback :-/
Also, the future of MPEG4 style postprocessing seems uncertain again.
video_out_opengl2: allow more frames for DR1 safety. 2015-08-08T14:36:20+00:00 Torsten Jager t.jager@gmx.de 2015-08-08T14:36:20+00:00 6ea419dc98ba8407d4a979623fa776d655d45751 






video_out: log frame usage.
2015-08-08T14:36:15+00:00

Torsten Jager
t.jager@gmx.de

2015-08-08T14:36:15+00:00

3df13323cf6c6ae991b726ed7748b701ece3e33a

This is not 100% exact, for the sake of speed.
However, it helps when setting the number of vo frames.





This is not 100% exact, for the sake of speed.
However, it helps when setting the number of vo frames.







ff_video_decoder: conditionalize AVCodecContext.stream_codec_tag.
2015-08-08T14:36:10+00:00

Torsten Jager
t.jager@gmx.de

2015-08-08T14:36:10+00:00

fc7d6ba9f6fe2e7cc8abd20430b59e1ff579d2b7












Copyright year update by hg log.
2015-07-24T12:58:05+00:00

Torsten Jager
t.jager@gmx.de

2015-07-24T12:58:05+00:00

88e923fbc5f6750480a4e9425e79607db59a656d












Get year from ID3v1 tag.
2015-07-24T12:58:00+00:00

Torsten Jager
t.jager@gmx.de

2015-07-24T12:58:00+00:00

ff9b615545c2af47829ba1086442024bee0e6098












libbluray 0.8.0 can play .iso images
2015-04-22T09:05:20+00:00

Petri Hintukainen
phintuka@users.sourceforge.net

2015-04-22T09:05:20+00:00

666337a70b33a1ba0e4df418125a2625f6a85579












Reduce logging
2015-04-22T08:52:15+00:00

Petri Hintukainen
phintuka@users.sourceforge.net

2015-04-22T08:52:15+00:00

a201bc8a5d3dc248f5dd35ee2f40d542c91e7c86












demux_ts: Another safety against freeze on heavy seeking.
2015-04-02T15:50:00+00:00

Torsten Jager
t.jager@gmx.de

2015-04-02T15:50:00+00:00

ade430c70fc0e00443fcdb635722c85cc6691bfb

VDPAU seems not to need this nasty workaround (and it seems
not to be harmed by it either).

Doing this inside ff_video_decoder would affect all demuxers,
and only for MPEG-TS I am halfway sure that there will follow
another sequence header after seek.

Alternatively, demux_ts could seek to keyframes. This is not
as simple as it sounds. Many DVB streams I have seen only
show "Auxillary data" and "Sequence extended info" NAL units
openly. All the rest hides behind escapes...





VDPAU seems not to need this nasty workaround (and it seems
not to be harmed by it either).

Doing this inside ff_video_decoder would affect all demuxers,
and only for MPEG-TS I am halfway sure that there will follow
another sequence header after seek.

Alternatively, demux_ts could seek to keyframes. This is not
as simple as it sounds. Many DVB streams I have seen only
show "Auxillary data" and "Sequence extended info" NAL units
openly. All the rest hides behind escapes...







More fast unaligned integer hacks.
2015-03-06T18:48:50+00:00

Torsten Jager
t.jager@gmx.de

2015-03-06T18:48:50+00:00

965c0ed960d6cf2acf4bdfed6eb9389be101eab3

Taken from my private "TJtools" project.





Taken from my private "TJtools" project.