[FFmpeg-trac] #3900(avformat:new): ff_check_h264_startcode does not support 3-byte startcodes

FFmpeg trac at avcodec.org
Thu Oct 16 12:19:55 CEST 2014


#3900: ff_check_h264_startcode does not support 3-byte startcodes
------------------------------------+------------------------------------
             Reporter:  Lastique    |                    Owner:
                 Type:  defect      |                   Status:  new
             Priority:  normal      |                Component:  avformat
              Version:  git-master  |               Resolution:
             Keywords:  h264 avi    |               Blocked By:
             Blocking:              |  Reproduced by developer:  0
Analyzed by developer:  0           |
------------------------------------+------------------------------------

Comment (by Lastique):

 Replying to [comment:14 cehoyos]:
 > Don't you agree that the file you attached indicates that your patch is
 not correct as-is since it allows to write invalid transport streams?

 No, I don't. You didn't specify why exactly the patch is not correct
 except that "player X doesn't play such files". This is not a
 justification for me. What I'd like to see as a justification is a
 reference to a standard paper or a formal documentation of some sort
 stating that 3-byte start codes are not acceptable in this container or in
 these circumstances. Further, if such justification is shown, it should
 also be decided whether ffmpeg should perform the startcode conversion
 internally or not. I.e. if the limitation is container-specific, ffmpeg
 should handle this internally as a part of writing media file, IMHO.

 Besides:

 > That transport stream lacks SPS/PPS so that's why it doesn't play.

 Yes, that's probably the case. I don't really understand why they didn't
 get written though, the code is the same as it is for avi or other
 containers, and SPS/PPS were placed in the extradata.

--
Ticket URL: <https://trac.ffmpeg.org/ticket/3900#comment:17>
FFmpeg <https://ffmpeg.org>
FFmpeg issue tracker


More information about the FFmpeg-trac mailing list