On Thu, Aug 20, 2015 at 8:30 AM, Robert Krüger <[email protected]> wrote:
> > > On Wed, Aug 19, 2015 at 8:08 PM, Carl Eugen Hoyos <[email protected]> > wrote: > >> Robert Krüger <krueger <at> lesspain.de> writes: >> >> > encoder : Lavf56.23.100 >> >> How was the input file created? >> >> using the apple h.264 encoder for encoding and libavformat for muxing. I > presume something is at least unusual in the file and I am trying to deduce > what from analyzing the file but have not found anything, that's why I ask > about ffmpeg internals of h.264/mov metadata detection as that might give > me a clue. I thought I remembered that ffmpeg's policy was to give > container metadata higher priority than bitstream metadata but that would > contradict what I am observing. > I just enabled logging and could verify that the colr atom vaues are correctly parsed: [mov,mp4,m4a,3gp,3g2,mj2 @ 0x7fa82b802a00] nclc: pri 1 trc 1 matrix 1Input #0, mov,mp4,m4a,3gp,3g2,mj2, from '/Users/krueger/tmp/colortest/h264-pipeline-out.mov': so it seems to indicate that something else overrides this but I am unable to find in the code where that might happen. _______________________________________________ ffmpeg-user mailing list [email protected] http://ffmpeg.org/mailman/listinfo/ffmpeg-user
