Closed adworacz closed 1 year ago
Thanks!
Sure thing! I'm working on a few more updates, including cleaning up the AVX2 degrain code like I did for the SSE2 code, and rolling around some ideas of how to take advantage of X264's high bit depth support for SAD/SATD calculations. That last one might be require some minor hoop jumping, but I have an idea that hopefully makes it pretty clean.
Seems to be about the same speed (maybe slightly faster), and more importantly fixes all of the NASM build spam we were getting before.
Also opens up the door to potentially use AVX2 SAD functions, although I have yet to experiment with integrating these.
This is all literally copy-paste from the x264 project.