media-driver VS rav1e

You might need to install the non-free intel media driver: https://github.com/intel/media-driver

Intel had VP9 decode since Kaby Lake (7th gen); original Skylake (6th gen) could not handle it (see here).

I've heard intel-media-driver was the way to go for supported hardware, but I also heard that it could decode slower than libva-intel-driver (like #925), and that Intel doesn't officially support some kind of decode feature in intel-media-driver, but it's implemented in libva-intel-driver.

I reproduced the issue on my end and opened a ticket to let intel devs know. https://github.com/intel/media-driver/issues/1628

I can see 22.6.6 was released in December ( Releases · intel/media-driver (github.com) ), but I'm having troubles getting it installed.

You should be able to choose FFmpeg VAAPI encoder, though it is suggested that kernel 6.2 is needed for this to actually work. I can confirm that it's broken with my A770, while I can choose the device, encoder is failing. No such problem with my Rapter Lake iGPU.

On GitHub there are similar problems reported: https://github.com/intel/media-driver/issues/1282

Source: https://github.com/intel/media-driver/blob/master/docs/media_features.md

Sure. You'll see it very often in codec implementations. From rav1e, a fast AV1 encoder mostly written in Rust: https://github.com/xiph/rav1e/tree/master/src/x86

Large portions of the algorithm have been translated into assembly for ARM and x86. Shaving even a couple percent off something like motion compensation search will add up to meaningful gains.

Or the current reference implementation of JPEG: https://github.com/libjpeg-turbo/libjpeg-turbo/tree/main/sim...

Again, if anything AOM would be the one restricting licenses to AV1 (if they chose to) except AOM has stated and also published AV1 in a way to allow license free access to development (which allows people to make forks of the official build like it's open source) and usage. (1)(2) I don't see why they would suddenly change this.

You can speed up the sampling process with --vmaf n_subsample=5, which in my experience works more accurately than either 2 or 4, possibly due to this bug/feature present in multiple encoders. You might also need to manually set the number of threads used for VMAF calculation with --vmaf n_threads=16, but YMMV.

That remember me of https://github.com/xiph/rav1e which is an AV1 encoder

I put in the request for ffmpeg passthrough mastering-display data a few years ago and haven't heard of any support yet.

It's getting there.. Initial steps for FFmepg: https://patchwork.ffmpeg.org/project/ffmpeg/list/?series=8444 rav1e: https://github.com/xiph/rav1e/pull/3000

It's entirely possible to re-use bits of other HW encoders for the first pass (motion estimation, etc).