OKLab2 / consistent implementation of Euclidean deltaE's

[allefeld]

I suggest that all deltaE functions which are defined as a Euclidean distance in some (modified) color space have that modified color space explicitly implemented, such that they themselves can be implemented via distance. One example is deltaEOK2 which is the Euclidean distance in OKLab modified such that a and b are multiplied by 2 – there could be a corresponding color space "OKLab2". Or "OKL2ab"?

Related to that a question: Is there any reason not to use such an "OKLab2" instead of OKLab throughout?

[svgeesus]

there could be a corresponding color space "OKLab2". Or "OKL2ab"?

There could be, but there is not.

Related to that a question: Is there any reason not to use such an "OKLab2" instead of OKLab throughout?

deltaEOK2 is there presently for testing, while not disturbing existing users of deltaEOK.

If, as suspected, it is generally better then it will become the recommended distance metric for Oklab and Oklch.