[FFmpeg-devel] [PATCH v2] lavf/movenc: suggest video_track_timescale for invalid timescale

[Sven C. Dack]

On 12/10/16 13:44, Mark Thompson wrote:
> How about the approach in the test program below?  This makes the whole timebase
> fraction; the suggested timescale would then be the denominator of that.  (It is
> currently just a program linked with lavu, I could make it into a patch if
> someone else thinks this might actually be a good idea.  It would need better
> error handling and some thoughts about overflow, too.)
>
> It shows that we can do significantly better in the 417083/10000000 case than
> any of the currently-suggested approaches (divide by 1000, guess 1/24 or
> 1001/24000):

It may be too much. We only know of 3 video files, which don't play on the Apple 
player. We don't know how they got produced, what produced them, what caused 
them to have this deviation, if it was done intentionally or by accident, nor do 
we know why Apple chooses to deviate in their playback from other players, why 
exactly it fails and where else it fails apart from these 3 files. We really 
only know about 3 files, which could have already been fixed with the help of 
ffmpeg and without applying any patches to it. So I wouldn't take it this too 
far. Patches are supposed to be simple and one should try to make then even 
simpler, so say the rules of the project.