- From: Isaac Muse via GitHub <sysbot+gh@w3.org>
- Date: Wed, 06 Mar 2024 21:05:06 +0000
- To: public-css-archive@w3.org
> I've checked my work many times. We cannot make progress about how `oklch(90% 90% 0deg)` should appear on a given screen if we do not agree on its physical meaning. > Here's an example that should feel less controversial. Consider the color `color(display-p3 1 0.3 0.2)`. Suppose you have a display with sRGB primaries, but they can go above 1 (say, to 1.5). That display can produce this color, because it's equivalent to `color(srgb 1.087, 0.204, 0.106)`, which is in that display's capabilities. That color will look identical on this slightly-extended-sRGB display as it does on a true P3 display. I agree that all parties must come to an understanding or there will be a deadlock, and what I'm saying is that the perception that you are somehow getting more SDR colors by raising the white point is not correct. RGB spaces are designed to represent their gamut, yes they can be allowed to extend, but the meaning of 0 and 1 doesn't mean what it used to at that point, you are only now representing wider colors in a system they were not designed for, but it will allow round trip conversions. Some GMAs may potentially angle the color reduction through lightness and chroma to capture a closer color in some circumstances (which is what I thought you were doing initially). This "brilliant magenta" in your example _appears_ to be produced due to clipping a color outside of the current display profile. I have a mac with a Display P3 monitor, albeit not an XDR monitor, but I can produce brilliant magentas by simply clipping `oklch(90% 90% 0deg)` as well. This makes me think that your display is clipping out-of-gamut colors due to them not being in your current profile. <img width="1351" alt="Screenshot 2024-03-06 at 1 37 56 PM" src="https://github.com/w3c/csswg-drafts/assets/1055125/a3745905-5495-4683-b2ee-ba013b2a4e14"> If the argument is that somehow Rec. 2020 lightness is brighter than Display P3 and sRGB because of gamut size, then what you are suggesting completely uproots everything about CSS colors as they are currently defined. They all (sRGB, Display-p3, Rec. 2020) have D65 white points (not including Pro Photo, etc.), they don't scale up higher because they are wider (and that term wider gamut should be an indication that they are wider, and not taller), and none of these spaces are designed to provide HDR headroom for highlights. Anyway, my intention is not to cause contention, but only to point out that I think there is a fundamental flaw with the argument here. Additionally, I don't think authors want results akin to clipping, which it seems you are arguing for, they want consistent lightness as that is a huge deal when we are talking about contrast, a most sought-after solution in the CSS world. -- GitHub Notification of comment by facelessuser Please view or discuss this issue at https://github.com/w3c/csswg-drafts/issues/9449#issuecomment-1981784401 using your GitHub account -- Sent via github-notify-ml as configured in https://github.com/w3c/github-notify-ml-config
Received on Wednesday, 6 March 2024 21:05:07 UTC