Hi,

> Could we somehow get the AVFrame pointer in your custom callback data
> or the VmafMetadata data? When all of the expected metadata is written
> then you'll know you're ready to forward the frame through the
> filtergraph. Doing it this way hopefully avoids a lot of the code and
> traversing through linked lists looking for AVFrame pointers as you've
> done.

Thanks for the suggestion! 

I think main difficulty of the implementation is sending non-monotical vmaf 
frames in order. Tracking the written metadata is not enough in my opinion. We 
need some kind of structure that need to put frames in order before they sent 
or it should need to send in order somehow. I have been working on kinda 
similar solution for couple of days that I did with linked list but with using 
av_fifo instead of my implementation. It looks working fine. Tests are welcome.

https://github.com/yigithanyigit/FFmpeg/pull/3/files

NOTE: This implementation uses small vmaf patch that I mentioned earlier in the 
thread.

Thanks,
Yigithan

_______________________________________________
ffmpeg-devel mailing list
ffmpeg-devel@ffmpeg.org
https://ffmpeg.org/mailman/listinfo/ffmpeg-devel

To unsubscribe, visit link above, or email
ffmpeg-devel-requ...@ffmpeg.org with subject "unsubscribe".

Reply via email to