Is the oklch colour notation in CSS considered to be accessibility supported?

[TestPartners]

I had never heard of oklch or the Oklab color space until today when I came across it in a CSS file on a public sector website. MDN says it has full browser support, albeit only in the last two years or so.

Does anyone know if there are any issue with any assistive technologies? Do any of them use their own rendering engine that might not support oklch? And what about testing tools? Do we know if any of them do or don't support oklch?

FWIW, I cannot see any plausible reason for using this on most websites, certainly not on the one I was testing. As far as I can tell, it's intended for 3D rendering in games and animations.

[patrickhlauke]

I cannot see any plausible reason for using this on most websites, certainly not on the one I was testing. As far as I can tell, it's intended for 3D rendering in games and animations.

Supposedly, easier to read/understand by humans. And can be used to encode wide gamut colours that can't be expressed by regular RGB/HSL/hex.

https://evilmartians.com/chronicles/oklch-in-css-why-quit-rgb-hsl

And what about testing tools? Do we know if any of them do or don't support oklch?

The contrast calculations from WCAG 2.x don't, if i recall correctly, handle wide gamut and different colour spaces at all. So this will be a problem going forward...though if I recall correctly, a while ago we added a clarification to the contrast stuff that anchors it to sRGB colour space, so at least it should the maths should work after an initial conversion/downsampling to sRGB to at least achieve consistent-enough results.

[TestPartners]

Thanks, that's interesting. It never occurred to me that 16 million colours wasn't enough, but I'm not a graphic designer. This suggests that some oklch colours don't even have a hex or rgb equivalent. I guess the tool vendors will have to come up with a conversion algorithm, so I hope they all use the same one.

Manual testing will be interesting. Chrome displays colour swatches in the dev tools, but we would be using a hex or rgb colour picker, which would sometimes (always?) give the wrong values. I would hope this would only be relevant in borderline pass/fail situations, but I have no idea how to verify that.

It seems that there are problems with both the spec and browser implementations for OkLab and OkLCH color spaces. I was made aware of the following in a separate discussion:

w3c/csswg-drafts#9449
https://bugzilla.mozilla.org/show_bug.cgi?id=1847421
w3c/csswg-drafts#9715
https://bugs.webkit.org/show_bug.cgi?id=255939

This is a comment from the W3C Technical Director:
w3c/csswg-drafts#9449 (comment)

Until these issues are resolved, I will be advising our clients not to use oklch unless there is absolutely no alternative. It was entirely unnecessary on the website I was testing today.

And we haven't even touched on the limited hardware support for oklch, which just makes matters worse.