[FFmpeg-trac] #7379(avformat:new): Matroska regression after ff_isom_write_avcc() changes

FFmpeg trac at avcodec.org
Fri Aug 24 01:39:14 EEST 2018


#7379: Matroska regression after ff_isom_write_avcc() changes
-------------------------------------+-------------------------------------
             Reporter:  gontandre    |                     Type:  defect
               Status:  new          |                 Priority:  normal
            Component:  avformat     |                  Version:
             Keywords:  mkv, h264,   |  unspecified
  ff_isom_write_avcc                 |               Blocked By:
             Blocking:               |  Reproduced by developer:  0
Analyzed by developer:  0            |
-------------------------------------+-------------------------------------
 Summary of the bug:

 I have found the issue after moving from ffmpeg version 3.4.1 to 4.0.2.
 I use ffmpeg to grab video from USB camera.
 I use codec copy flag to just avoid video transcoding.
 Saving stream into matroska container worked fine on ffmpeg 3.4.1, but on
 4.0.2 it is failed with following error - "Could not write header for
 output file #0 (incorrect codec parameters ?): Invalid data found when
 processing input".
 I have saved the stream into flv container. Attempt to save this flv to
 mkv is failed with the same error.
 I have made my own investigation and figured out that the issue it related
 to changes in ff_isom_write_avcc() call. In 4.0.2 it returns error if data
 size is zero, while on ffmpeg 3.4.1 it returns 0. So I have created the
 patch to just ignore that error for now.
 Anyway I should understand that is regression, or maybe my camera
 generates some unsupported/inconsistent stream format?

 Thanks a lot for your support.

 How to reproduce:
 {{{
 $ ./ffmpeg.exe -v trace -i matroska_and_ff_isom_write_avcc.flv -vcodec
 copy test.mkv > matroska_and_ff_isom_write_avcc.log 2>&1
 }}}

--
Ticket URL: <https://trac.ffmpeg.org/ticket/7379>
FFmpeg <https://ffmpeg.org>
FFmpeg issue tracker


More information about the FFmpeg-trac mailing list