Today I got the chance to ask our computer graphics professor a question that I asked myself quite some time ago: what the fuck is the color pink?
Like, the color right before infrared is red, the color right before ultraviolet is violet. And every other color is some wavelength between those two colors. EXCEPT FOR PINK!?
On a hue color wheel pink is between red and violet, so it's wavelength has to be somewhere around there, right??
Well, turns out pink is the color humans perceive when red and violet are mixed (duh), meaning the red and blue cones are stimulated. Since both cones respond to wavelengths on the opposite ends of the visible spectrum, there is no monochromatic wavelength that would trigger both, hence there is no wavelength that looks pink.
That's also the reason pink does not appear in a rainbow, because there white sunlight light, a mixture of (almost, hi Astro-fedi) all monochromatic wavelengths, is refracted based on wavelength, so no mixed colors occur in it.
Sven222 mag das.
teilten dies erneut
Marika
Als Antwort auf Marika • • •second mind blow: due to the curved, convex shape of the visible light perception, it is impossible to find three light sources that when addictively mixed together will be able to represent every visible color. Mixtures between the colors will always form a triangle between the three colors in the above graph, and there is no triangle that covers the whole space with its vertices corresponding to a color that actually exists.
So there's no way to build a perfect display, at least not by using only three colors.
William D. Jones
Als Antwort auf Marika • • •Marika
Als Antwort auf William D. Jones • • •not quite, the outer curve represents all visible light to the human eye. The inner triangle is some color space used to define a subset of those colors on a computer (and possibly display on a screen). If you read my follow up, you'll understand why the outer colors cannot possibly be shown by a display: yuustan.space/notes/a1clx7iyx1d81fk2
But of course you're correct that the colors you see when viewing this picture on your screen are not the colors they are meant to be. You would have to paint the graph in real life (and view under direct sunlight) to have the accurate colors (a color printer won't do either... They usually don't form a triangle but a hexagon: CYMK is what printers use: www.ttamayo.com/wp-content/uploads/2018/01/comparison-of-some-rgb-and-cmyk-colour-gamuts-on-a-cie-1931-xy-chromaticity-diagram.png)
Marika
2024-12-03 22:40:38
:neocat_floof_happy: lexi
Als Antwort auf Marika • • •Marika
Als Antwort auf :neocat_floof_happy: lexi • • •Marika
Als Antwort auf Marika • • •Samuel
Als Antwort auf Marika • • •windwitch🍃tragus
Als Antwort auf Marika • • •Marika
Als Antwort auf windwitch🍃tragus • • •thanks uwuw
(Although I knew some parts of this explanation before, it was only today that I finally got the whole picture. It's kinda funny to me how in daily life I don't think much about my perception of the world, because that is just what I assume and expect it to be. It gets really interesting when talking to colorblind people or reading about Tetrachromacy.)
Paul Bone
Als Antwort auf Marika • • •Marika
Als Antwort auf Paul Bone • • •Marika
Als Antwort auf Marika • • •At least I can be confident in my explanation now ><
- YouTube
www.youtube.comivy
Als Antwort auf Marika • • •Marika mag das.
bkim
Als Antwort auf Marika • • •if there's no perfect display with 3 colors, how can I see some variation on the most curved parts of that path when viewing through my phone?
I concede that the upper green part looks very homogenous, which makes sense, but I should be able to perceive the vertices of my display, right?
Marika
Als Antwort auf bkim • • •well, already the colors encoded in the image are not the colors they are supposed to be. Most pictures (such as the PNG image I posted above) by default use the sRGB color space, which can itself just represent a triangle of colors. So even if you screen could represent all colors, they were not stored in the picture to begin with.
It is possible though to use a color space that can store all the existing colors (and even non-existing colors ^^). Some image formats allow to specify the color space they are intended to be interpreted with in the image metadata.
sjolsen (Home Ultimate edition / Girl Inside®)
Als Antwort auf Marika • • •Sensitiver Inhalt
you've got the right idea except for this part: almost all colors are non-spectral; pink isn't special in that regard. only the curved boundary ("spectral locus") of the chromaticity diagram corresponds to monochromatic light.
(also there absolutely are wavelengths that stimulate all three types of cone cells; they're all fairly broadband and for this reason it isn't really correct to call them "red" or "blue" cones)
Marika
Als Antwort auf sjolsen (Home Ultimate edition / Girl Inside®) • • •Sensitiver Inhalt
thanks! The more I read the more I realize I only scratched the surface of the topic. And I tried to say something without really knowing definitions of words (and then it got shared a lot, which is kinda overwhelming). When I was saying "color" I actually meant those completely saturated colors, whereas lighter / desaturated colors are obviously mixtures of those. So to me the surprising thing was that pink is one of those mixture colors, although it looks like a saturated color?
Like, for other mixed colors besides pink (and white?) there exists a monochromatic color that has the same hue as the mixed color.
(yes, it was stupid of me to say this, especially because I looked at the wavelength / cone cell stimulation diagram right beforehand. so the better way to say it would have been that there is no monochromatic wavelength which can trigger the red and blue cones in a way that the result looks pink. Which I guess is, because that wavelength right where the red and blue curves intersect would also trigger the green cone a lot, whereas two monochromatic lights at the outer sides of the red and blue curves can trigger both without also triggering the green cone)