Encoding with “butteraugli distance =1” ensures that you compress optimally without the human eye being able to see the difference, it’s an awesome “fire and forget” setting for all your images besides lossless. You basically don’t have to guess or check your saved image if it looks good enough.
Unlimited size, multiple layers, better colors, more bit depth. Avif has severe limitations as an image format.
But for specific cases like comics and stronger compression at about 60% quality it’s currently better. I’ve heard jpegXL encoders still have some ways to improve though. For stronger compression avif can use “ssimulacra2” which is a great metric to avoid compression artifacts for encoders. So ideally jpegXL would support both butteraugly (high quality) and simulacra2 for low quality / high compression.
But for specific cases like comics and stronger compression at about 60% quality it’s currently better. I’ve heard jpegXL encoders still have some ways to improve though. For stronger compression avif can use “ssimulacra2” which is a great metric to avoid compression artifacts for encoders. So ideally jpegXL would support both butteraugly (high quality) and simulacra2 for low quality / high compression.