Thanks, I had extracted the algorithm from the source code of the previous page already. I am now trying to study it and interpret it because I see it is not a continuous function but has IF statements. Not as easy as I thought at first.
For now I understand RGB well but I cannot intuitively translate to HSL and back. I’ll have to keep at it.
Actually, sailor, the conversion is continuous. However, it is not smooth, as I mentioned before. This is part of the appeals of both Luv and Lab, as the XYZ/Luv and XYZ/Lab conversions are smooth. (XYZ/RGB is a simple linear transform and thus infinitely smooth.) Both Luv and Lab can be converted to cylindrical coordinates, and when this is done you get a parameterization of colorspace called Lch[sub]ab[/sub] that resembles HSV but is far more perceptually uniform.
See http://www.brucelindbloom.com/index.html?Equations.html for a great many useful equations in color space conversions.
Not quite. No frequency ever “tapers out” as you increase temperature. A 20,000 K blackbody will produce more red light (and more infrared, and more of every other frequency) than a 10,000 K blackbody. What happens is that you’ll get a greater increase on the blue end than on the red end, so the average color will be more bluish.
And humans only have two different kinds of cones. If we had three, then we would be able to distinguish a color space with four dimensions (since we also have rods which are sensitive to total luminosity).
No, we (usually! Re. above) have 3 types of cones. I think you may be getting confused by the fact that we have two types of photoreceptors, rods and cones. (Rods come in one flavour, and are specialised for low light vision - they do not contribute significantly to daytime vision in any respect).
In relation to sailor’s ‘discontinuous’ functions, some colour equations can be different depending circumstances. For example, the equation for CIE Lightness (L) is different depending on whether you are working with low or high light levels.
Chronos, your blackbody statement would be true if you assumed equal radiative surfaces. However, most color temperature (and other color work) is based on an assumption of equal energy, not equal area. Obviously a 25000K black body will emit more radiation per square centimeter than a 5000K black body.
Also, please document your assertion that humans have only two types of cones. I have read countless treatises that insist that there are at least three different photoreceptors (commonly referred to as L, M, and S). Rods play little role in color vision, in part because there are no rods in the fovea, where the bulk of color vision takes place (see, for example, this page at the Color Vision Research Laboratory).
…plus, the photopigment in rods will become entirely bleached in daylight, rendering rods (temporarily) useless!