Intel's Open-Source SVT-AV1 Video Encoder Ends May With Another Performance Boost

[phoronix]

Phoronix: Intel's Open-Source SVT-AV1 Video Encoder Ends May With Another Performance Boost

It's been very fascinating to watch the speed improvements of Intel's SVT-AV1 open-source AV1 video encoder since in February when being made aware of Intel's new SVT video projects. The SVT-AV1 project is ending out May with another step-up in performance for what is already one of the fastest CPU-based AV1 video encoders...

http://www.phoronix.com/scan.php?page=news_item&px=SVT-AV1-Ending-May-Faster

[tildearrow]

Intel's Open-Source SVT-AV1 Video Encoder Ends May With Another Quality Reduction

I wish there was a quality benchmark besides a speed one.

[curaga]

Write one for PTS https://github.com/kornelski/dssim/tree/dssim1-c is one good algorithm for comparing compressed image quality. Build that, output source and dest videos to png, run it and sum the numbers. Won't take long to script.

[milkylainen]

Intel's Open-Source SVT-AV1 Video Encoder Ends May With Another Quality Reduction

I wish there was a quality benchmark besides a speed one.

Second that. Plus I wish the SVT-suite would add H264.

[karasu]

Intel's Open-Source SVT-AV1 Video Encoder Ends May With Another Quality Reduction

I wish there was a quality benchmark besides a speed one.

Quality is now better, but you have to choose between speed and quality.

https://www.reddit.com/r/AV1/comment...han_x264_x265/

[Michael]

Write one for PTS https://github.com/kornelski/dssim/tree/dssim1-c is one good algorithm for comparing compressed image quality. Build that, output source and dest videos to png, run it and sum the numbers. Won't take long to script.

Actually PTS has built-in support for a decade now of doing image quality comparisons pixel by pixel to highlight the changes, just isn't used much - https://www.phoronix.com/scan.php?pa...nvik_iqc&num=1

[andreano]

Cool that PTS can do that, but is the comparison metric selectable?

There are many metrics that are used to apprximate human vision – we should use one of them. Preferrably VMAF – it goes for being the best.

[Michael]

Cool that PTS can do that, but is the comparison metric selectable?

There are many metrics that are used to apprximate human vision - we should use one of them. But please use VMAF, it goes for being the best.

There is just one implemented: my own basically comparing the RGB values of each pixel directly... I haven't touched any of the code though in a decade due to time / interest.