Artefacts in MPEG-4 ASP elementary stream

[Carl Eugen Hoyos]

I will upload an ASP elementary streams that shows accumulating artefacts with ffmpeg, plays fine with mplayer -vc xvid -demuxer lavf (-demuxer mpeg4es produces different artefacts).

$ ffmpeg -i tears.m4v -ss 9.65 out.png
ffmpeg version N-48881-g7980cca Copyright (c) 2000-2013 the FFmpeg developers
  built on Jan 14 2013 17:11:14 with gcc 4.7 (SUSE Linux)
  configuration: --enable-gpl --disable-indev=jack
  libavutil      52. 14.100 / 52. 14.100
  libavcodec     54. 89.100 / 54. 89.100
  libavformat    54. 59.107 / 54. 59.107
  libavdevice    54.  3.102 / 54.  3.102
  libavfilter     3. 32.100 /  3. 32.100
  libswscale      2.  1.103 /  2.  1.103
  libswresample   0. 17.102 /  0. 17.102
  libpostproc    52.  2.100 / 52.  2.100
[m4v @ 0x2bcbde0] Estimating duration from bitrate, this may be inaccurate
Input #0, m4v, from 'tears.m4v':
  Duration: N/A, start: 0.000000, bitrate: N/A
    Stream #0:0: Video: mpeg4, yuv420p, 640x480 [SAR 1:1 DAR 4:3], 29.97 tbr, 1200k tbn, 30k tbc
Output #0, image2, to 'out.png':
  Metadata:
    encoder         : Lavf54.59.107
    Stream #0:0: Video: png, rgb24, 640x480 [SAR 1:1 DAR 4:3], q=2-31, 200 kb/s, 90k tbn, 29.97 tbc
Stream mapping:
  Stream #0:0 -> #0:0 (mpeg4 -> png)
Press [q] to stop, [?] for help
[mpeg4 @ 0x2bbacc0] looks like this file was encoded with (divx4/(old)xvid/opendivx) -> forcing low_delay flag
frame=    1 fps=0.0 q=0.0 Lsize=N/A time=00:00:00.03 bitrate=N/A
video:474kB audio:0kB subtitle:0 global headers:0kB muxing overhead -100.004531%

out.png shows artefacts that can not be seen with 00000290.png:
$ mplayer tears.m4v -demuxer lavf -vc xvid -vo png

[Michael Niedermayer]

you can use -idct 14
is there any player that can autodetect this ?
if not this ticket probably should be closed,

[Carl Eugen Hoyos]

The original file was ​http://samples.mplayerhq.hu/V-codecs/RMP4/greenlines.rmp4.p.avi produced with Sigma Designs' "REALmagic" encoder and the original question should have been if the FourCC is needed to identify the idct (since xvid apparently doesn't need it).

Which (kind of) sample will fail to decode correctly with the xvid decoder?

[Michael Niedermayer]

Replying to cehoyos:

Which (kind of) sample will fail to decode correctly with the xvid decoder?

If xvid uses this idct by default (instead of somehow detecting itself as encoder) then more or less any sample with large gop encoded by a different encoder with different idct could potentially fail. most samples would of course not fail