Re:JPEG XL vs. AVIF for web use-cases
by JyrkiAlakuijala on Friday October 16, 2026 @04:56PM
(#50124567)
I believe that:
ssimulacra2 85 is rather close to median quality in the internet
ssimulacra2 90-95 is rather close to high quality in the internet
ssimulacra2 98+ is close to camera quality
Re:JPEG XL vs. AVIF for web use-cases
by jaffathecake on Friday October 16, 2026 @05:08PM
(#50124748)
'high quality' and 'very high quality' are different things.
There’s also data that isn’t needed in the final delivery format but is needed for editing. 8bit vs 16bit color for example. 8 bit is visually all you need for an image, but if you edit it to say being up the shadows, 8 bit will show banding because you’ve stretched it to the point the bands are visible. While 16bit has so much extra data it still looks perfect after being stretched.
JPEG XL is the ideal high end photography format. AVIF seems better for a space efficient web format.
Also the JXL decoder in Firefox and Chrome only returns 8 bits, unfortunately.
Re:JPEG XL vs. AVIF for web use-cases
by jaffathecake on Saturday October 17, 2026 @02:10AM
(#50128431)
Are you sure about that? Wouldn't that make the HDR example look bad in Chrome? The source image is 16 bit. https://jakearchibald.com/2026/jpeg-xl-vs-avif/#lossy---hdr
Re:JPEG XL vs. AVIF for web use-cases
by zbobet2012 on Saturday October 17, 2026 @01:59AM
(#50128381)
Human eyes can quite reliably see greater than 8 bit color. While the eye is only able to differentiate about 10 million colors at any given brightness, because its sensitive to brightness in a logarithmic fashion the eye needs about 24 bits of per channel depth to see no banding.
Presumably this is for HDR. In the SDR space I don’t think you’d see the difference between 8 and 16bit. And the way it’s encoded is logarithmic so you can express the visible range with less bits where your eyes are less sensitive.
And the article supports what I have said.
Quotation from the article:
> But JPEG XL catches up at very high quality. Around SSIMULACRA2 85, libjxl's default matches AVIF's default, and at maximum efforts, jixel 'extraslow' is ~6% smaller than AVIF s0. jixel 'slow' is 10–13% smaller than AVIF's default from SSIMULACRA2 80 upwards.