1

I'm confused, why aren't the scales in the plot (with the labels 6000, 4000, 3000, etc.) forming a right angle with the Planckian locus? In my understanding, for a color not on the locus, one chooses the perceptionally closest color on the locus and takes the color temperature from there. But if the metric of CIE 1931 is already based on perception (Euclidean distance ~ perceptional distance), shouldn't the scales at least locally (they don't have to be straight, since the locus isn't straight) form a right angle?

Also, just so I understand, at infinite temperature, the black body radiates the same amount of energy in all frequencies, i.e. has a flat spectral power distribution. Does that mean that the blend of all visible frequencies at same intensity appears bluish white to the human eye, rather than flat white?

4
  • this question should probably have been asked on a science stack such as cognitive psychology, physics or so. Its not really graphics design which isnt a science but an art of laying out images and text for presentation. Also note color theory is not really a science, ita about how to mix pleasing color combinations (So arts not science). The CIE graph is not a color theory thing but a color science one which falls under cognitive psychology.
    – joojaa
    Mar 23, 2023 at 21:25
  • Anyway, there is no such thing as absolute white, which is why you could choose different white points. Insead human visual system is adaptive and the white can be in different places depending on prevalent conditions. But yeah the human eye is not linear, or equally sensitive to all wavelengths so yes in precence of a reference it would be blueish. But then could arange prevalent conditions so that it would look white and your considered white pointva red leaning hue. See blue gold dress illusion.
    – joojaa
    Mar 23, 2023 at 21:31
  • Thanks for the comment, I won't post similar questions here again! Since I already received a nice answer, I think it would be unfair to delete the question now.
    – fweth
    Mar 24, 2023 at 8:46
  • Thats fine, the reason though is not that you cant ask as such but you wont get as broad viewership here
    – joojaa
    Mar 24, 2023 at 14:05

2 Answers 2

2

CIE 1931 XYZ is highly non-linear in how much the color seems to be changed when the coordinates are changed a certain amount in different visible color points. I'm afraid drawing from a point a straight perpendicular line to the blackbody curve doesn't give the visually closest blackbody radiation color, it gives only what's closest point of the curve in this diagram.

To get the mapping you obviously want, do this:

Present the color in the CIELAB system. You can find the conversion formulas in numerous sources. Get the CIELAB chromaticity diagram, search there with a straight perpendicular line what's the visually closest blackbody radiation color in the same luminosity level.

Another thing:

Your assumption "the blackbody radiation spectral distribution approaches uniform when the temprature goes towards infinity" is false. See this Wikipedia snippet of the famous Planck's blackbody radiation formula:

enter image description here

If one can read it he sees that the higher the temperature, the more the higher frequencies get weight. The 3rd power of the frequency v in the nominator shows it.

At about 6000K temperature the spectral distribution happens to be such that our sight nerves get the stimulus combination which is seen as white. In higher temperatures the weight of what makes the blue, increases. But that's said already in the older answer.

3
  • Thanks a lot for the thorough answer! I'm just confused because other sources say that at infinite temperature, the SPD distribution across a fixed, finite spectrum (like visible light) approaches the Rayleigh-Jeans law which has the 4th power w.r.t. wavelength?
    – fweth
    Mar 24, 2023 at 10:25
  • Ah, I understand now, it's both, \lambda^4 and \nu^2, like on the Wikipedia page...
    – fweth
    Mar 24, 2023 at 10:59
  • It's the same answer I get from SymPy, when I plug in Planck's law: i.imgur.com/j3eiPvT.png
    – fweth
    Mar 24, 2023 at 11:28
1

The CIE 1931 chromaticity chart has a centre, which is the "whitepoint", marking the chromaticity of white and greys.

Chromaticities which are in a straight line from this whitepoint to the edge have the same "hue", which is part of the technical definition of a colour. It looks to me that those lines have been drawn radiating out from the whitepoint so they are marking colours with the same hue as the blackbody radiation.

To answer your second question, blackbody radiation is not the same amount of energy of all frequencies at any temperature, it is always an uneven spread around a peak frequency. At around 6000 (the temperature of our sun) it is perceived as white, at higher temperatures it is bluish, and lower temperatures yellow and then red.

1
  • Thanks a lot for the answer! I was not talking about straight lines from the whitepoint to the edge, but how to map points to the Planckian locus, which do not directly sit on the locus. My understanding was that we choose the perceptionally closest color on the locus and take the temperature from there. But the graph indicates that we don't take the closest point in the CIE 1931 space, otherwise those small lines would be in a right angle with the locus.
    – fweth
    Mar 24, 2023 at 8:41

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.