Oklab could be used to lower the saturation like in gamut clipping or gamut compression techniques without modify the perceptual hue
1 Like
I donât know if it is super useful in my regular workflow. I primarily did this as a personal experiment to exercise my ability to write a module and learn about the oklab color space.
A couple of places I can see this module being used in vkdt is
- fine tuning some color mappings that lie just past the rec2020 gamut.
- [with some changes to the implementation,] clipping from rec2020 to something smaller like sRGB with some additional knobs to control how that compression is accomplished.
I can let @hanatos correct me, but I think the included colour module pretty much already does all this when performing gamut mapping. Itâs implementation is based on the abney effect.
2 Likes
right⌠i mean⌠the saturation slider is kinda gamut aware, but only in this world. values outside spectral locus pose a problem. i havenât been very motivated to try and come up with something that works beyond too⌠but in general it would of course be much nicer to have something that âjust worksâ even when fed rubbish.
it follows the abney effect by using a lut constructed from spectra that have a main peak around the same wavelength. the observation/assumption is that surface reflectances keep their main lobe even under heavy indirect self-bounces. i.e. very rough surfaces still appear to have the same colour even though youâre looking at the reflectance spectrum multiplied to itself a couple of times. with that, it can walk the gamut in a couple of ways while keeping perceived hue (under this assumption) constant. unfortunately this kind of reasoning requires actual reflectance spectra (no negative energies).