Comment by MadameMinty
8 days ago
Abysmal encoding time, poor lossless performance, poor lossy performance at perceptually indistinguishable quality compared to JPEG XL.
So far I haven't performed a benchmark where AVIF has a decisive advantage in any real scenario; none of the scenarios I tested accept the quality loss or compute necessary.
Have you published this benchmark anywhere? It seems to contradict those publicly available, so it would be interesting to examine. To be clear: I doubt your claim and your competency to make it.
Benchmarking code or BTFO. JXL propaganda is getting too irritating at this point.
For the rest of us, this resource debunks all stupid jxl propaganda.
https://jakearchibald.com/2026/jpeg-xl-vs-avif/
This source doesn't debunk anything, it's literally AVIF propaganda that cherry picks parameters and images. It demands JXL match AVIF size before comparing encode time at abysmal q50 factor. Would you encode anything worth showing at q50?
Who cares if AVIF is slightly smaller at some specific scenario when Mercury is in retrograde, when it takes a century to encode?
My entire point is that some 3% size gain is acceptable for shortening the encode time by 97%, which is what JPEG XL does.