XYB: is the Chroma from Luma adjustment correction really in the JXL implementation?

[facelessuser]

XYB: is the Chroma from Luma adjustment correction really in the JXL implementation?

[facelessuser]

lol, well, I guess Twitter and Youtube are the references. I did see code in the reference related to chroma-from-luma about calculating the best configuration for the channels. And I agree it is difficult to follow. It seems it can be toggled, so I guess that is why they mention the "default".

I admit that the adjustment is a bit more lab-like and more intuitive. Thanks for your insight. I wanted to make sure I wasn't crazy.

[danburzo]

I couldn't really follow the libjxl repo, so I can't attest to what the code is doing. The Culori implementation follows the footnote in the whitepaper at page 6, which indeed does not capture the (default?) chroma-from-luma transform. The adjustment aims to bring the achromatic axis at X = B = 0, and the author Jon Sneyers calls it the actual XYB space according to the final JPEG XL specification. Unfortunately I don't have a non-social-media reference for it.