Re: [csswg-drafts] [css-color-4] Channel clipping breaks author expectations, especially when using 'perceptually uniform' spaces (#9449)

@jamesnw Thanks for working on refining this! My concern is that while mapping 0% and 100% lightness to black and white respectively seems to match author expectations, it actually seems to bring you *farther* from the actual color specified (at least wrt ΔE). And yes, the discontinuity when going from 99% to 100%.

> 2. Looking at  @tabatkins Step 6 [above](https://github.com/w3c/csswg-drafts/issues/9449#issuecomment-1944235895), the second scale step seems to be necessary, even if the color is already gamut. I played around with it a bit, and it seemed like the results improved with a second scale.

That’s very interesting. I would expect Scale to be a no-op for an in-gamut color 🤔 

> 4. My observation is that the DeltaE's are often incredibly close between Scale LH and CSS, even when different channels are vastly different- see [deploy-preview-438--colorjs.netlify.app/apps/gamut-mapping/?color=oklch(60%25+47%25+90)](https://deploy-preview-438--colorjs.netlify.app/apps/gamut-mapping/?color=oklch(60%25+47%25+90))

Is that a neutral observation or are you meaning to say that ΔE2000 is a poor measure of color difference? If the latter, do you think a different ΔE might work better? I could even make it configurable.

-- 
GitHub Notification of comment by LeaVerou
Please view or discuss this issue at https://github.com/w3c/csswg-drafts/issues/9449#issuecomment-1944286819 using your GitHub account


-- 
Sent via github-notify-ml as configured in https://github.com/w3c/github-notify-ml-config

Received on Wednesday, 14 February 2024 17:29:29 UTC