So what's the final verdict here, can this be pushed or not?Timo - did you manage to test it over last weekend?
I haven't found the time, sorry.I'm generally not opposed to this. It does not disrupt normal use, and spinning up nvenc does have a surprisingly hefty overhead, so it makes sense to have features like that.
One request I'd have for the patch though, if one of the new max_ parameters is set, and dyn res changing is not supported, there should be an error.
Likewise if the res does change and max_width/height were not set. Silently not doing anything seems bad.Also, someone needs to ack the non-nvenc changes. Not sure if that explanation is really needed, since this is an nvenc specific edge case.
smime.p7s
Description: S/MIME Cryptographic Signature
_______________________________________________ ffmpeg-devel mailing list ffmpeg-devel@ffmpeg.org https://ffmpeg.org/mailman/listinfo/ffmpeg-devel
