AV1

494 readers

1 users here now

Community for AV1 video codec

founded 1 year ago

MODERATORS

[SVT-AV1 git] Introduce the variance boost feature (proper psy based variance spatio-temporal adaptive quantization) (gitlab.com)

submitted 4 months ago by [email protected] to c/[email protected]

4 comments fedilink hide all child comments

Yes, it's finally been released after cooking for so long in the svt-av1-psy fork!

you are viewing a single comment's thread
view the rest of the comments

[–] [email protected] 2 points 4 months ago (1 children)

I know everyone is really excited about this. If I’m using ffmpeg, how do I use it (once ffmpeg merges this update)?

[–] [email protected] 1 points 4 months ago (1 children)

Compile FFmpeg with the SVT-AV1-PSY fork

[–] [email protected] 1 points 4 months ago (1 children)

That’s it? I don’t need any arguments or options when I do the transcode?

[–] [email protected] 3 points 4 months ago* (last edited 3 months ago) (1 children)

This adds several new options to the SVT-AV1 standalone encoder (--enable-variance-boost, --variance-boost-strength). I notice that, by default, variance boost appears to be turned off.

To use it with ffmpeg, you'd likely need to explicitly enable variance boost in the source code where ffmpeg applies the config options to the svtav1 encoder.

For testing, it's probably easier to run the standalone SvtAv1EncApp with aforementioned parameters and pipe ffmpeg's rawvideo output to it. Either way, I assume this functionality will eventually become an option for ffmpeg's -svtav1-params, now that it's in the SVT-AV1 main tree.

Disclaimer: this is just what I concluded from a quick check of the patch and ffmpeg code.