5 ms·
As far a I understand this benchmark JXL was using 8 CPU cores, while QOI naturally only used one. If you were to plot the graph with compute used (watts?) inst
by phoboslab 3y ago
As far a I understand this benchmark JXL was using 8 CPU cores, while QOI naturally only used one. If you were to plot the graph with compute used (watts?) instead of Mpx/s, QOI would compare much better.
Also, curious that they only benchmarked QOI for "non-photographic images (manga)", where QOI fares quite badly because it doesn't have palleted mode. QOI does much better with photos.
- lonjil 3y agoActually, they did try QOI for the photographic images: > Not shown on the chart is QOI, which clocked in at 154 Mpx/s to achieve 17 bpp, which may be “quite OK” but is quite far from Pareto-optimal, considering the lowest effort setting of libjxl compresses down to 11.5 bpp at 427 Mpx/s (so it is 2.7 times as fast and the result is 32.5% smaller). 17 bpp is way outside the area shown in the graph. All the other results would've gotten squished and been harder to read, had QOI been shown.
- phoboslab 3y agoThanks, I missed that. I just ran qoibench on the photos they used[1] and QOI does indeed fair pretty badly with a compression ratio of 71.1% vs. 49.3% for PNG. The photos in the QOI benchmark suite[2] somehow compress a lot better (e.g. photo_kodak/, photo_tecnick/ and photo_wikipedia/). I guess it's the film grain with the high resolution photos used in [1]. [1] https://imagecompression.info/test_images/ https://imagecompression.info/test_images/ [2] https://qoiformat.org/benchmark/ https://qoiformat.org/benchmark/