Hacker News Reader

Comment by dehrmann | Hacker News Reader

The more and more I learn about color, the more I learn what we're taught as children is wrong.

Primary colors are arbitrary, the only reason there's three is possibly because we have three types of cones, but really to cover most of the visible spectrum.

So yeah, color spaces are a thing.

Cone sensitivity doesn't line up with primary colors, and some are more sensitive than others.

Tetrachromacy exists in humans, so some people can see more colors.

Magenta and violet are very different. Magenta isn't a real color, two colors we interpret as one.

The sky is actually blue and purple. If you look hard at it you can see both, kinda like seeing RGB on a white LCD.

And I'm sure there's more.

View on HN

>Magenta and violet are very different. Magenta isn't a real color, two colors we interpret as one.

It's more accurate to say magenta is a color that can't be produced from a monochromatic light source. It's a mix of ~450nm and ~650nm light. It's still a color, though; the term "color" is more akin to the final output of the brain, as opposed to the light entering the eyes.

My high school art teacher made us make color wheels using red, yellow, and blue paint. She showed us books that use these "primary colors" and said that this makes them more pleasing to children.

I thought about it for a moment. If these are the primary colors, why aren't they what printers use? Printers use cyan, magenta, and yellow. And there's a great symmetry there, since those are all the secondary colors to light's primary red, green, and blue, which are the colors used in computer monitors. And the only thing that makes these colors primary to us is that they're the colors that the cones in our eyes perceive.

So I decided to experiment. When I made my color wheel, I substituted blue for cyan and red for magenta. The color wheel I produced was much more vibrant and beautiful.

There are a lot of things my school teachers tried to teach me that show up on wikipedia's common misconceptions list. For example, the "equal-transit-time explanation of aerofoil lift". I got pretty jaded about this stuff. Now I don't trust anyone's explanations unless I can understand them on a deeper level.

> equal-transit-time explanation of aerofoil lift

Seems like flying an airplane upside down disproves that explanation pretty quickly.

I remember this one distinctly because in school they just gave me a handout and asked me to come up with the explanation myself. I thought about it. I didn't know. It seemed like if you had an angle of attack then the air would hit against the bottom -- like a knife would want to continue to travel straight through a fluid. I didn't know what the significance of the curve over the top of the wing was. They tried to guide me toward the equal-transit-time explanation and I was like, uh, ok, is that really true? I remained suspicious of this for a long time.

> If these are the primary colors, why aren't they what printers use? Printers use cyan, magenta, and yellow.

There really is no such thing as "the" primary colors. The school primary colors are still primary as those used by a printer, and are based on historically widely used pigments. CMY (and usually K) allow for a wider gamut. But even this isn't perfect. There are printing processes with more primaries to get a wider gamut.

> And the only thing that makes these colors primary to us is that they're the colors that the cones in our eyes perceive.

This isn't quite right. For one, our cones are not monochromatic receptors, and moreover, they overlap! There isn't really just one true red, green, blue used in computer monitors either.

Because of the way our brain perceives colors (metamerism), you can create a wide gamut of colors with "alternative" primaries.

What sort of alternative primaries? Just slightly-adjusted colors that still trigger each of the three cones individually?

You can't trigger cones individually. The cones are responsive to a wide-spectrum and they overlap, especially in the case of the L and M receptors. The peak wavelength of the L receptor (the "reddest") is about 580nm -- that is not red, that's yellow-green.

Color vision and stimulus is not a straightforward mapping of primaries triggering cones. If it were that simple you could trivially render all perceivable colors with 3 chosen primary colors. This is impossible to do. You can mathematically define 3 primaries that cover the entire visible spectrum but they cannot physically exist (complete, but imaginary). Any chosen set of 3 primaries is a compromise. For subtractive materials it is trickier, which is why photo inkjet printers will use up to 8 primaries.

You might find this informative: https://web.archive.org/web/20080717034228/http://www.handpr.... The section starting with Maxwell and the "3 artist's misconceptions" especially.

But the only general definition of primary is basically just any set of colorants that can be mixed to get a useful gamut. In subtractive materials, this is why you won't see a painter messing around with mixing cyan, magenta, and yellow (better explained in the link).

Printers use cyan, magenta and yellow because printing is a subtractive process. It's the complement of the additive RGB that monitors use. For example, cyan acts as a filter for red, so a combination of magenta and yellow filters out everything but red and thus appears red. Ultimately, the difference is in whether something emits light (a monitor) or has to reflect light (a white piece of paper).

You're right in that what we're taught is often incomplete or misguided. Teachers are fallible. But as a child you assume their authority implies them being correct. I reckon seeing through that illusion is an important part of growing up. And to me, part of us growing up as humanity must involve not having to rely on the authority of governing bodies.

The additive/subtractive colors thing is has a really neat sort of symmetry to it. That's how I knew it was the correct explanation. There was no such beautiful internal logic to the Red Blue Yellow system and I couldn't figure out how people came up with it.

I never believed my teachers. As early as 4th grade they were treating me like a troublemaker for not following rules like their three-paragraph essay format.

I don't like the idea of having authorities on knowledge. I much prefer Montessori or Socratic teaching methods, or explorations. They're harder to do, but they produce a better understanding of the material and they allow the student to teach the teacher as well.

Yeah, unfortunately color perception is one of those things that is a bit too complex for a self-discovery method. While the "symmetry" explanation is satisfying it really isn't correct at all. Color perception and color matching within art (where it was useful) and more recently as a science is something that is complicated and took many years of the best scientific minds to figure out. Sometimes you need authorities on knowledge, and stand on the shoulders of giants as it were.

https://web.archive.org/web/20080717034228/http://www.handpr...

The teachers aren't completely "wrong", they were just conveying a simplification of history of pigments (also touched on in that article). It is after all true you can mix those colors and get a wide-range of colors (including a blacker black then you would with CMY). But any pedagogy that says there is such a thing as "primary" colors that make all colors is necessarily going to be wrong, even if its CMY.

I'm red-green colorblind.

So blue/green and green/yellow seem like more plausible categories than blue and green. I can see red, but green pretty much looks like bluish yellow.

That’s interesting. I am also red/green colourblind and ok with red. For me though it’s more that greens and browns can be confusing.

I have problems with anything involving green. So yes, greens and browns. Many US green signal lights look dull blue. Also with red, but less so. I can't tell some pinks from grays. And in any mixture, reddish vs greenish.

Reply on news.ycombinator.com