[FFmpeg-devel] Bug report: decoder errors post-revision 13613

[Andy Parkins]

Hello,

With latest SVN ffmpeg I get video corruption when playing from an
RTSP MPEG4 source.  Like this:

$ ./ffplay -stats rtsp://1.2.3.4/encoder1
FFplay version git-dbda489, Copyright (c) 2003-2008 Fabrice Bellard, et al.
  configuration: --enable-static --disable-shared --enable-pthreads --disable-vhook --enable-extra-warnings --disable-ffmpeg --disable-ffserver --disable-encoders --disable-decoders --disable-muxers --disable-demuxers --disable-parsers --disable-filters --disable-bsfs --disable-devices --enable-decoder=mpeg4 --enable-encoder=mpeg4 --enable-decoder=zlib --enable-muxer=m4v --enable-demuxer=m4v --enable-demuxer=mpegvideo --enable-demuxer=mpegps --enable-demuxer=rm --enable-muxer=mpegts --enable-demuxer=mpegts --enable-demuxer=mpegtsraw --enable-muxer=rtp --enable-demuxer=rtp --enable-demuxer=rtsp --enable-demuxer=sdp --enable-parser=mpeg4video
  libavutil     49.11. 0 / 49.11. 0
  libavcodec    52. 0. 0 / 52. 0. 0
  libavformat   52.22. 1 / 52.22. 1
  libavdevice   52. 1. 0 / 52. 1. 0
  built on Oct 13 2008 11:18:36, gcc: 4.2.4 (Debian 4.2.4-3)
Input #0, rtsp, from 'rtsp://1.2.3.4/encoder1':   0B
  Duration: N/A, start: -2.-40000, bitrate: N/A
    Stream #0.0: Video: mpeg4, yuv420p, 720x576 [PAR 1:1 DAR 5:4],  8.33 tb(r)
Title: RTP session
[mpeg4 @ 0xa192730]warning: first frame is no keyframe
[mpeg4 @ 0xa192730]concealing 215 DC, 215 AC, 215 MV errors
[mpeg4 @ 0xa192730]ac-tex damaged at 39 29
[mpeg4 @ 0xa192730]Error at MB: 1373
[mpeg4 @ 0xa192730]ac-tex damaged at 23 30
[mpeg4 @ 0xa192730]Error at MB: 1403
[mpeg4 @ 0xa192730]ac-tex damaged at 20 30
[mpeg4 @ 0xa192730]Error at MB: 1400
[mpeg4 @ 0xa192730]1. marker bit missing in 3. esc
[mpeg4 @ 0xa192730]Error at MB: 1404
[mpeg4 @ 0xa192730]ac-tex damaged at 32 30
... etc ...

I've git-bisected my way to revision 13613 as being the first revision
where the corruption appears.

The negative start time shown in stream stats started appearing from
revision 14045, but I believe that that is unrelated to the corruption.

I'm sorry I can't offer more information about what is occurring.  I
can probably create a recording from the source and upload it
somewhere if that would help.



Andy
-- 
Dr Andy Parkins, M Eng (hons), MIET
andyparkins at gmail.com