Just wanted to clarify that this feature is not just academic but
something that is quite important for broadcasters where we often
receive/send content of mixed alpha type.

Users often don't know the difference between the alpha type and send
material in whatever mode their editing software happens to use.
During playout premultiplied alpha has become much more common.
Especially since often output from e.g. Chrome is used during
composition (which is premultiplied) and newer video switchers seem to
move towards premultiplied. While more traditional workflows use
straight alpha. Hence we often end up with a mix we need to somehow
handle.

In our use case we try to guess the alpha mode with the "colordetect"
filter and/or let the users explicitly set the alpha mode and then we
forward that using setparams. Without this feature it is difficult
and/or inefficient to get composition using e.g. vf_overlay or
vf_libplacebo working well.

--
Robert
_______________________________________________
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