Why do computer screens etc use red, GREEN & blue?

February 16, 2005 12:50 PM

The primary colours are red,yellow & blue. All other colours can be made from them. Why then do computer screens etc use red, GREEN & blue?
posted by TiredStarling to computers & internet (23 comments total)

Quoth the Wiki:

"The human eye contains photoreceptor cells called cones which normally respond most to yellowish-green, green, and blue light (wavelengths of 564nm, 534nm, and 420nm respectively). The color yellow, for example, is perceived when the yellow-green receptor is stimulated slightly more than the green receptor, and the color red is perceived when the yellow-green receptor is stimulated significantly more than the green receptor.

Although the peak responsitivities of the cones do not occur at the red, green and blue wavelengths, those three colors are described as primary because they can be used relatively independently to stimulate the three kinds of cones."
posted by falconred at 12:57 PM on February 16, 2005

Red, yellow, and blue are the primary subtractive colors: in pigmets, red pigmet only reflects "red" light. While red, green, and blue are the primary additive colors. If you want to emit light, you have to use the additive method as you're adding light. If you're using a reflective method (paper, metal, plastic) you have to use the subtractive system.

Compare Additive color to Subtractive color.
posted by skynxnex at 1:05 PM on February 16, 2005

Red, yellow, and blue are not the primary colors in any meaningful sense.

Red, green and blue are the primary colors of light for the aforementioned reasons - because your eyes are most sensitive to them and it's easiest to use them.

Magenta, cyan, and yellow are the true primary pigment colors. That's why printer ink uses those three colors instead of straight-up red or blue. You'll notice, if you try to make purple out of red and blue, that you only get a murky grey-purple. Any kid with fingerpaints knows that red, yellow, and blue don't make every color you want.

You're taught red yellow and blue because they're CLOSE to magenta, cyan, and yellow, and they're some of the first color names you learn.
posted by u.n. owen at 1:05 PM on February 16, 2005

I've been thinking about color theory a lot recently as well (& even picked up a copy of the book "yellow & blue don't make green" which essentially teaches you how to mix pigments to create colors, arguing that there are in reality no pure primaries, but every blue is either a greenish blue or a purplish blue, etc).

Anyway, I have a piggyback question to this: what wavelength does reddish-purple fit into? How do we perceive a color that is somehow shorter than the shortest and longer than the longest visible wavelengths?
posted by mdn at 1:11 PM on February 16, 2005

I used to think as you do, Starling, until one day in high school my technical theater teacher shone 3 lights on a white cyc, gelled red, green, and blue, and intersecting in a Venn-diagram sort of way.

Where the red and green lights intersected, the color was brilliant yellow.

Skynknex's links go into more depth on the non-intuitive truths here.
posted by ikkyu2 at 1:11 PM on February 16, 2005

mdn, I think the color you're seeing is not a single wavelength. It's like hearing a touch-tone telephone, where each "tone" is actuall two distinct frequencies on top of each other. You hear (or in your case, see) the product of adding the two waves together.
posted by knave at 1:14 PM on February 16, 2005

knave: the purple color you see may in fact be a pure frequency.

The reason frequencies higher then blue seem sort of redish (making purple) is that the frequency response of the 'red' cones diminishes more slowly then those of the blue and green cones. So Violet light still triggers the red receptor, and it looks the same as if you'd mixed blue and red light together.
posted by delmoi at 1:28 PM on February 16, 2005

The primary colours are red,yellow & blue. All other colours can be made from them. Why then do computer screens etc use red, GREEN & blue?

Short answer, you were lied to about red yellow and blue.

One of the reasons that they might have lied is that Purple and Blue are part of the 'natural' color set of black, white, red, blue, green, orange, purple and pink which people find easy to learn and remember the names of.
posted by delmoi at 1:33 PM on February 16, 2005

knave, it's a lot more complicated than that: what you perceive as color is actually the dominant wavelength of an arbitrary EM spectrum.
posted by Eamon at 1:35 PM on February 16, 2005

sidetrack: has anyone ever figured out why certan musical chords have diffrent 'emotional' qualities? I'm thinking it could have something to do with the response of receptors in our ears. like a 4th or 5th dosn't cause as much interferance or something.
posted by delmoi at 1:36 PM on February 16, 2005

How do we perceive a color that is somehow shorter than the shortest and longer than the longest visible wavelengths?

The range of visible colors spans less than one octave. That is, blue light is about twice the frequency of red light. This is unlike sound. A piano, for example, spans seven octaves.

The question is what would light be like if it spanned more than an octave? Would we percieve it as we do sound? This question is mostly metaphysics but its an interesting aesthetic question.
posted by vacapinta at 1:51 PM on February 16, 2005

delmoi - That would make a good AskMe question if you're interested. I know a little bit about the topic, and there's likely other MeFites that know more. It's waaay of topic here though, so I'm reluctant to open that can of worms.
posted by raedyn at 2:00 PM on February 16, 2005

On second thought, Wikipedia's entry on Consonance might answer your question, delmoi - it's a start, anyway.
posted by raedyn at 2:13 PM on February 16, 2005

A picture is worth a thousand words
posted by spock at 2:39 PM on February 16, 2005

mdn: see this page. The colors along the long curved edge (colors shown are only approximations, your monitor does not display arbitrary spectra, yada yada yada) are the colors that monochromatic light appears as. Any color in the interior, or along the straight lower edge, requires light consisting of at least two different wavelengths. Wikipedia's CIE article is good too.

Eamon: what you perceive as color is actually the dominant wavelength of an arbitrary EM spectrum.

No, as the Wikipedia article you link to makes clear, monochromatic light of the dominant wavelength appears to be the same hue as the arbitrary spectrum itself, but that's not the same as being the same color--pink is not the same color as red. And even then, it's important to note that when we say it has the same hue, we mean hue as color scientists mean it, which is not quite the same as what artists mean by it.

Where color perception gets really fascinating is the next stage of processing after the photoreceptors. Why is it that red light + green light = yellow, which is seemingly a completely different color than red or green, rather than reddish-green or greenish-red? This is where you get into opponent processing, and the red/green/blue become red-vs-green/blue-vs-yellow/black-vs-white.
posted by DevilsAdvocate at 4:28 PM on February 16, 2005

Yeah, I've often wondered how the world would look if we could tell red+green from true yellow. For all the beauty we can see, we're probably missing much more. For that matter, what if we could see, say, polarization? What if we had twice six primary color receptors? What would infrared look like? Etc.
posted by kindall at 4:36 PM on February 16, 2005

Humans can see polarization.
posted by NortonDC at 9:14 PM on February 16, 2005

This is where you get into opponent processing, and the red/green/blue become red-vs-green/blue-vs-yellow/black-vs-white.

I just want to jump in here for a second (I'm gonna get a bit technical here, but this subject is a pet of mine, the Wiki page is terrible, and people are saying all sorts of weird stuff. Not you, DevilsAdvocate!): Opponent processing isn't the final stage of the game. Most people think that there are three primary hues: red, yellow, and blue. Color scientists tend to think there are four: red, green, blue, and yellow, which make up two of the axes of our three-dimensional color sphere (black-white is the other). Any color we see can supposedly be located in this color space (though there are illusions you can try out that let you see 'hypercolors' existing outside the sphere). Color scientists posit four primary hues because they want to shoehorn our phenomenology into a nice isomorphic fit with the opponent processing going on in the ganglion cells and the lateral geniculate nucleus. The problem, unfortunately, is that the outputs of the ganglion cells don't correspond perfectly with the hues we perceive as primary. For example, the outputs that should correspond with yellow perceptions actually peak under wavelengths of 600 nm, which under normal circumstances looks reddish-orange to us. So there has to be a lot of color processing that goes on downstream of the ganglia. My favorite hypothesis put out along these lines is by Kimberly Jameson (and others). She suggests that positing five primary colors - red, green, yellow, blue, and purple - better fits our phenomenology and psychophysics.

Part of the reason I'm writing this is to point out that there isn't a simple correspondence between wavelengths and perceived color, or retinal cones and perceived color, or even opponent processing and perceived color. Some people in this thread have wondered what it would be like to see ultraviolet light... even if we had receptors that could pick up ultraviolet light, the rest of our brain wouldn't necessarily represent it any differently, so we might not see any new colors at all. Color is probably best identified with something like matrix activation in V4, but even that ignores a whole bunch of the fine-tuning that doubtlessly goes on in the frontal cortex.

To go back to the original question: red, green, and blue are used because their pure wavelngths roughly match the peak sensitivities of our three retinal cones. But we didn't have to use those wavelengths; the fact that we use those three wavelengths is actually kind of arbitrary. It's one color gamut among many that can be used in the CIE color space. In fact, different computer monitors might have different colored phosphors. The cool thing about our three-coned visual system is that any perceived color can be emulated by a huge number of (additive) mixtures of three different pure wavelengths. Different spectral stimuli that cause identical perceptions are called metamers. Metamers exist because of opponent processing: our ganglia send information to our visual system based on the differences between the spectra picked up by our cones... wavelengths don't get sent to the brain in their own private channel. That means you can feed all sorts of inputs in to get the same output (just as both 8-3 and 11-6 equal 5). This page describes the process pretty well, and also provides some calculations to transform one color gamut into another. (It doesn't tow the five-primary-hues line that I was espousing up above, but that's a pretty maverick hypothesis. The traditional view is that we have four primary psychological hues).
posted by painquale at 10:41 PM on February 16, 2005

Kindall: we can tell red+green light from pure yellow light. Try looking at a color photo under white light modified with a yellow gel filter; the photo will still look colored. Now go out walking at night, and look at the same photo under one of those yellow sodium vapour street lamps - it will magically have turned monochrome.

Something else you may care to play with is lighting a room at night with (a) an incandescent bulb (b) a compact fluorescent of the same nominal brightness. You'll see differences in the colours of things, because the incandescent bulb emits a continuous (black-body) spectrum and the fluorescent emits a peaky approximation of the same thing.
posted by flabdablet at 4:17 AM on February 17, 2005

Thanks, everyone.
Painquale, can you recommend any good books that discuss this? Technical is okay if reasonably contextualized.
posted by mdn at 7:47 AM on February 17, 2005

There's no such thing as THE primary colors. There are only primary colors for specific color systems. Within that system, the primary colors are those colors that don't have to be mixed.

So if I have a box of three crayons -- purple, red and yellow -- then the primary colors of that box are purple, red and yellow. I don't have to mix other colors to get those colors, but I do have to mix those colors to get other colors.

Of course, there are many colors that I couldn't make by mixing those three primary colors, but all color systems have their limitations. For instance, you can't make Coka-cola red with the Cyan, Magenta, Yellow print system. So when they print coke adds, they have to sneak in an extra, pre-mixed red -- an extra primary color! This is called a spot color.

A particular systems "gamut" is the range of colors you can get in that system my mixing that system's primary colors. RGB has a larger gamut than CMY.
posted by grumblebee at 7:52 AM on February 17, 2005

mdn: The best thing to pick up is probably a collection of papers edited by Alex Byrne and David Hilbert titled Readings on Color. It's a two volume set: the first volume discusses the philosopy of color, and the second volume is on the science of color. Color Perception: Philosophical, Artistic, and Computational Perspectives edited by Steven Davis is a great collection of papers from a conference. The papers are technical and reasonably up-to-date, but you could easily jump into them without any prior knoweldge.

If you want a self-contained book instead of a collection of papers, I'm not sure what would be best. Marr's Vision is a cornerstone of cognitive science, although it's more about object recognition than about color. It's the second most influential cognitive science book of the century according to this list. Kathleen Akins and Martin Hahn, two philosophers of color, are also writing a book that I'm sure will be excellent when it arrives -- watch for it.
posted by painquale at 10:03 AM on February 17, 2005

Ooh - just realized I forgot to mention C. L. Hardin's Color for Philosophers, which is a great book. It brought color science into philosophy's ambit and was one of the first to get me interested in this subject.
posted by painquale at 10:09 AM on February 17, 2005