I am aware that there is nothing inherently inevitable about the way we call colours; the names for the colours are simply social or linguistic conventions to name electromagnetic waves of a certaian wavelength that we can perceive with our eyes as light. What I’m interested in is whether the colours can be thought of as contiguous, uninterrupted segments of the spectrum, in a way such that (for instance) 400 to 500 nanometres would be “blue”, 501 to 600 nanometres would be “green”, and so on. Or is the spectrum full of enclaves and isolated areas in the sense that “green” might comprise several non-contiguous segments of the spectrum separated from each other by other segments that would, culturally, be identified as a different colour.
I am fully aware that there might be cultural differences and that the whole thing is, to a large extent, subjective - somebody might call a turquoise-like colour green which someone else perceives as blue, for instance. Yet, it’s possible that such differences only concern the precise delineation of the segments (e.g. whether the line that separates green from blue runs at 503, 507 or 510 nanometres), without affecting the question more fundamentally.
The colors of the rainbow are contiguous by definition -but- there are other elements to what we call “color” than the wavelength of the associated light. Brightness, saturation, etc. White, grey and black don’t involve different wavelengths, they involve different intensities.
If we’re restricted to spectral colors (a color that could be perceived from a single sharp emission line at some wavelength), then they’re definitely continuous: For any two wavelengths that are perceived as being in the same color category, the wavelengths in between will also be.
Take away that restriction, though, and it’s not necessarily so. Most people will regard a mixture of red and blue to be at least very similar to violet, and might even use the same name for both, even though red is very far away from violet in the spectrum.
On the other hand, classical Japanese lacked a distinction between blue and green, calling them both aoi. The sky and the grass were the same color. Different shades of aoi, yes, but still aoi. So if you were hired by some daimyo to make him an “aoi” flag, pretty much any green or blue flag would do unless he qualified it like by saying aoi-like the grass, or aoi-like the sky.
For the most part spectral colours are spread in a continuous range with no voids or lumps. But not quite. There is a tiny lump in the response of the eye that is seen as red in the range 330 to 480nm - which is in the far blue. This means that far blues are seen as a mix of blue and red - which is why violet looks purplish. The eye-brain maps pure far blue to the same colour as a mix of red and blue.
If you talk about pure spectral colours this doesn’t matter - there is still a one-to-one mapping of wavelength to perceived colour - but there is a set of pure spectral colours that are ambiguous with a colour made from a mix of wavelengths.
When you talk about non spectrally pure colours it is a different matter. The way the eye works means that you can pick any perceivable colour (except the except for the ends of the perceivable spectrum) and you can create it with an infinite number of combinations of spectrally pure light. (Which isn’t to say you can make any colour out of any other.) A critical point is that this combining is indeed continuous. There are no lumps or bumps.
That’s interesting, and it’s something that’s always puzzled me, for two reasons:
The cones in our retinas come in three varieties. There are the long-wave cones, sensitive to reds, oranges, and yellows; the short-wave cones, sensitive to blues, greens, and violets; and the intermediate cones, mainly sensitive to yellows and greens. (I know this is an oversimplification, but that’s the basic concept.) So why does purple (a mix of red and blue) stimulate the same brain response as violet? Do the long-wave cones also respond to far blues?
Why should a mixture of red (which absorbs blue) and blue (which absorbs red) inks have any color at all? I would expect it to appear black, rather than purple, since it’s absorbing both ends of the spectrum.
The transparent bands in the rainbow come from shepherd moons, IIRC.
Cartoonacy - the earlier post by FV mentioned - the red cones see a small bump in the high blue frequencies, so hence the perception of purple, as a mix of red and blue.
But yes, if we are talking pure spectrum then the colours are continuous. But we need to consider 2 things - the source or composition of a colour, and how we see it.
Colour could be a single pure frequency, like a laser. Or, it could be a mix of frequencies, such as a red and a blue laser, or a smear of frequencies, like an incandescent or fluorescent light or a phosphor.
We perceive colour by what proportion of red, green, and blue hits the assorted cones of our eyes. From far enough away, a colour TV type display of red, green, and blue lights (each relatively pure smears) can look like white, or any colour, based on relative intensity that each rod in the retina perceives. Similarly, you can get the same effect from a collection of coloured floodlights (something stage lighting is common for). What we perceive as white is generally equal to what sunlight produces, a “normal” balance of RGB intensities.
Note that the human capacity for interpretation also plays a role. In a relatively coloured environment, our mind will weed out some of the tint. Electronic eyes and film are not as forgiving. Where we see white after a while in the light, cameras will tell you that sunlight is white, outdoor shade in sunlight is blue (from the sky glow), incandescent light bulbs are orangeish (the light bulb is not quite hot enough to be “white-hot”) and fluorescent lights are greenish due to the frequency the phosphor. Our minds adjust but cameras do not.
But that still leaves my second question. Subtractive color mixing is the result of combining pigments that absorb different segments of the spectrum, so that our eyes only perceive the wavelengths that both pigments reflect. So why is a mixture of red and blue perceived as purple? The pigments should absorb both ends of the spectrum, and appear as black. Do red pigments also have a “bump” that reflects part of the high blue frequencies?
Subtractive coloring doesn’t remove ALL the light in that band. So colors don’t appear “black”.
There’s nothing magic about the color names we’ve chosen, or about their number. Newton originally thought there were five basic colors. Then he made the analogy between colors and the notes of the scale, and figured there should bge seven, correspobding to the seven notes of the scale (ignoring the octave as a repeat of “do”). There was some support for the idea of this, because things like Newton’s rings ran the spectrum, then started at red again and repeated. Newton associated each color with a particular length (although he stubbornly resisted the idea of light as a wave, so he couldn’t call it a “wavelength”), and could show that each such length was in the same harmonic ratio as the notes in the scale. That meant, to his mind, that there was a color for each. Voila! Suddenly Orange, whgich had until then been a combination color (red and yellow) became a color of the spectrum, added to the original five (red, yellow, green, blue, violet). But he still needed one for the “note” between blue and violet, where most people don’t actually seem to perceive a different color Newton grilled his friends and convinced them there MUST be a color there, which he called “Indigo”, and thus the seven colors of the rainbow were born.
Except most people still don’t see “Indigo” as a color. The Universal Code for resistors uses the rainbow for its color = number banding, but ignores “indigo” (which I suspect everyone would have seen as “blue”). The “Rainbow flag” either only has six colors, with no indigo:
Agreed; I’m deliberately oversimplifying. My point is not so much that it should be black, but that it shouldn’t be purple. Let me rephrase.
Red pigment (or ink, or dye, or what have you) absorbs short-wave blue light. Blue pigment absorbs long-wave red light. By the rules of subtractive color mixing, a blend of red and blue should reflect neither red nor blue light. So why do we perceive it as purple?
And while I agree that some red and blue light will be reflected, the same is true of green, yellow, and orange. Yet we don’t see the mixture as white. Why do we see it as purple?
I withdraw my question. I just realized my own mistake.
Mixing *red *and *blue *does not produce purple. Mixing *magenta *and *cyan *(in the proper proportions) does, because both of these colors reflect high-end wavelengths.
Are you looking at the color matching functions? Please note that those curves are x, y, z not r, g, b. They are not representative of your photoreceptors. The spectral sensitivities of your cones do not show this. The bump is a result of the problem mentioned in the OP, but doesn’t occur at the retinal level. In order to create this color by mixing only 3 lights, yet you cannot get this hue without adjusting red. In short, the bump is a result of the proportion of primaries (partially those selected way back due to technological limitations), but the cones in our eyes have single peaks.
Found an in-depth explanation.
Yes, I was referring to the matching functions. Although there is no physical realisation inside the body of these curves, they are the easiest to understand in thinking about how colours work.
The bump does come from the cone’s sensitivity, but is hard to see when you look at the curves. Whilst it is true the cones have a single peak each, what matters is the difference in sensitivities between them, and at the blue end you see that the difference between the sensitivities has a glitch - both red and blue are dropping down towards zero, but the red sensitivity shallows out slightly, and thus the difference between red and blue is less, with the result that the effective processed colour information has the bump in it.
Another feature of our colour vision is that there are two ways to make yellow.
It exists as both a discrete wavelength of EM radiation, and also can be made by combining red and green frequency light.
This is a quirk of how our colour vision works. If we had fully 3 degrees of freedom of colour vision red and green should make another colour. This link gives a nice intro (towards the end explaining about yellow), and at the very bottom there’s a link to much more information than you could ever possibly want about human colour vision.
