4 ms·

RGB is a colour model. It is misinformation. The reason that a given HSL / HSV implementation may not accurately deliver a given chromaticity vector has more

RGB is a colour model. It is misinformation. The reason that a given HSL / HSV implementation may not accurately deliver a given chromaticity vector has more to do with the nature of colour spaces than the model's implementation implicitly. So I disagree, it is misinformation, and the reasons for choosing a given color model / colour encoding scheme are varied. An incorrect and simple answer for a complex series of problems.

jacobolus 11y ago
Sorry to sound negative here, I’m not trying to discourage you. You seem to have a bunch of misconceptions about color and color science. My primary recommendation is to try reading a couple books about the subject, which should help you get a better grip on the parts you’re currently missing. > RGB is a colour model Without some further qualification, “color model” and “color space” are effectively synonyms. > It is misinformation. There’s really nothing particularly wrong with the linked article. I don’t like the author’s proposed HUSL space, but his explanation of why HSL and HSV are bad tools is straightforward and reasonable. > ... a given HSL / HSV implementation may not accurately deliver a given chromaticity vector ... In a technical sense, “chromaticity vector” is not very meaningful as a term when discussing gamma-corrected RGB coordinates, and is completely meaningless w/r/t HSL or HSV. Chromaticity is a precise technical term, referring to the projection of trichromatic matching functions onto a two-dimensional plane perpendicular to luminance. When people talk about “chromaticity” they usually either mean xy coordinates in xyY space, or sometimes u', v' coordinates, or very occasionally some other chromaticity plane, such as more recent spaces based on more explicit measurements of human cone responses. The Cr and Cb coordinates of a Y'CrCb model are sometimes referred to as “chromaticity” coordinates, but in a technical sense that is a wrong and misleading term to use. I’m not sure what “given chromaticity vector” you are talking about, but if you want to get chromaticity from R'G'B' values, you should first linearize them, then transform from RGB -> XYZ, and then divide by Y to find x and y. > ... has more to do with the nature of colour spaces than the model's implementation implicitly This is not a comprehensible string of words. I’ve spent more than a decade studying color science, color reproduction, etc., and I don’t have the foggiest idea what you’re getting at.
- kelsolaar 11y ago
  Thanks for that, terminology is important in any domain but I found it to be decisive when you talk about colour science. A lot of people are using incorrect terms in incorrect contexts which leads to even more confusion.
kelsolaar 11y ago
Your colour terms usage is really sloppy to the point it makes no sense in some instances (jacobolus comment is really spot on). Here is the terms reference when it comes to colour science: http://eilv.cie.co.at/ http://eilv.cie.co.at/

jacobolus 11y ago

Sorry to sound negative here, I’m not trying to discourage you. You seem to have a bunch of misconceptions about color and color science. My primary recommendation is to try reading a couple books about the subject, which should help you get a better grip on the parts you’re currently missing. > RGB is a colour model Without some further qualification, “color model” and “color space” are effectively synonyms. > It is misinformation. There’s really nothing particularly wrong with the linked article. I don’t like the author’s proposed HUSL space, but his explanation of why HSL and HSV are bad tools is straightforward and reasonable. > ... a given HSL / HSV implementation may not accurately deliver a given chromaticity vector ... In a technical sense, “chromaticity vector” is not very meaningful as a term when discussing gamma-corrected RGB coordinates, and is completely meaningless w/r/t HSL or HSV. Chromaticity is a precise technical term, referring to the projection of trichromatic matching functions onto a two-dimensional plane perpendicular to luminance. When people talk about “chromaticity” they usually either mean xy coordinates in xyY space, or sometimes u', v' coordinates, or very occasionally some other chromaticity plane, such as more recent spaces based on more explicit measurements of human cone responses. The Cr and Cb coordinates of a Y'CrCb model are sometimes referred to as “chromaticity” coordinates, but in a technical sense that is a wrong and misleading term to use. I’m not sure what “given chromaticity vector” you are talking about, but if you want to get chromaticity from R'G'B' values, you should first linearize them, then transform from RGB -> XYZ, and then divide by Y to find x and y. > ... has more to do with the nature of colour spaces than the model's implementation implicitly This is not a comprehensible string of words. I’ve spent more than a decade studying color science, color reproduction, etc., and I don’t have the foggiest idea what you’re getting at.

kelsolaar 11y ago
Thanks for that, terminology is important in any domain but I found it to be decisive when you talk about colour science. A lot of people are using incorrect terms in incorrect contexts which leads to even more confusion.

kelsolaar 11y ago

Your colour terms usage is really sloppy to the point it makes no sense in some instances (jacobolus comment is really spot on). Here is the terms reference when it comes to colour science: http://eilv.cie.co.at/ http://eilv.cie.co.at/