Zak

IIRC, the biggest reason that rav1e is slower than aomenc is because aomenc does a massive amount of search space pruning at the higher speed preset, particularly when it comes...

There's also another very important factor to take into account when comparing aomenc and SVT-AV1 against rav1e: unless I missed something while parsing the code, rav1e never voluntarily denoises the...

So basically, one of the 1st steps we should do to improve quality is implement the full set of CDEF search strengths. The current method, which is picking CDEF strength...

Oh, I completely forgot there are multiple format specific crates. Thanks for the tip. And yeah, a PR to enable flate2 utilization replacing the current deflate crate seems like a...

It might be a bug with djxl 0.6.1 since I just tried manually decoding it both with djxl 0.6.1 and djxl 0.8.0 Git from 2022-09-06, and only djxl 0.6.1 returned...

@lavilao @FarisR99 You people do realize that ffmpeg does have glsl shader compatibility? libplacebo has been a thing in ffmpeg for a while now.

Yes, and you just gave me an idea as well for further speedups and threading benefits. If we could do it noise estimation at a transparent tile level(no actual tiles...

I see this needs to be refactored since I'm seriously restarting my encodes.

Everything seems good on my end, as all flags seem to work(except --ip-qidx-ratio as I haven't tested it yet). I'd wait until tomorrow to merge it since there might be...

I should have been clearer: I haven't tested it yet :P