I’ve been puzzling over this one for a little while:
We perceive a mixture of red and green light as yellow; it isn’t yellow, it’s red and green because it can be split back up into red and green - there is such a thing as pure yellow light that cannot be split.
Yellow lies between red and green in the visible spectrum, it makes sense that our eyes/brains might be ‘averaging’ the mixture of red and green to settle on yellow.
All well and good, but why do we perceive a mixture of blue and red as purple/magenta, which does not lie between red and blue.
Also, is it just me or does it look as though both ends of the visible spectrum (as it is depicted here) are quite similar in terms of perceived colour?
What is actually going on when we perceive colour? (both in the receptive cells in the retina and in the brain)
Are we saying then, that the perception of yellow (when actually a mixture of red and green is present) is not a case of the eyes/brain ‘averaging’ the wavelength?
i.e. that the perceptive colour mapping system is two-dimensional -C, Y and M at corners of a triangle, mixing to white in the centre as opposed to visible spectrum, which is linear and one-dimensional?
In the argument over subtractive vs. additive colors, I think you’re missing the point of Mangetout’s question, which seems to be howcum our eyes perceive different mixes of colors to be the same. In other words, can we perceive “colors” which are the results of different spectral distributions as if they were the same, and why?
The answer is that yes, this is the way we see color. Apparenmtly we have three degrees of freedom in our vision (whether you have additive or subtractive colors, you only need three primaries). This is the way our eyes have developed, and I don’t know why it should be. It’s the basis of our tristimulus color theory, and the reason our color TVs have three different pigfments, and our color printing uses three inks.
Occasionally I hear people say that we have three different sets of cones, each responding to a different color, like the color-separated detectors in a color TV, but I’ve never been taught this in any optics class. The way the eye responds to color is a complex thing. Just when you think you have it figured out, along comes Edwin Land with the Land Effect and his Retinex Theory of Color Vision, and suddenly you’re confused again.
Regardless of the physical mechanism, tristimulus color seems to work pretty wellwith our eyes, and it was the basis upon which James Clerk Maxwell produced the first color photo way back about 1870, and we’ve been using it ever since.
This is fundamentally different from the way we perceive sound – you can’t mix two different frequency sound waves and expect your ear to hear a third, completely different frequency (“beats” don’t count) – our ears aren’t built on a tristimulus basis. Two sounds with different frequency spectra will not be perceived as the same, unlike colors. I don’t see why we couldn’t have been built with four-stimulus or five-stimulus eyes. I always thought it could make for an interesting science fiction story – the Aliens of Triskelion or wherever are upset with our color TV sets because our sets are tristumulus, but their eyes are quatrastimulus, and all our colors are “off”.
The two-dimensional thing is that venn diagram-like thing you see where circles of the three primaries are overlapped -you need a two-dimensional surface to do this, but you can represent all electromagnetic wavelengths in only one dimension.
Libertarian – I’m afraid you’re confused. Any object you see reflects or transmits the different colors in different ways. A plot of the amount of reflection or transmission as a function of wavelength is the spectrum of the object. “white” light contains all the wavelengths of visible light (although not all in the same amount), from about 350 to about 700 nm. only VERY rarely does something transmit or reflect only one wavelength out of the whole batch --Lasers and Narrow-Band Interference Filters being the two chief (and expensive) examples. With everything else, the spectrum has a range of colors in it, with some predominating over others.
So when you see a RED apple, you’re actuallty looking at something that has not only red light in it (you’re 622 to 780 nm – although if you can see light beyond 700 nm that isn’t REALLY bright, your eyes are a LOT better than most people’s), but also quite a bit of orange, yellow, green, and even blue.
What’s more confusing is that I can produce an entirely different spectral makeup that looks exactly the same to your eye. In most cases I can even come up with a single spectral line of color (as from a Laser or Filter) that will look the same as well. Youir eye hasn’t got the capasbility of distinguishing them.
Furthermore, NOT all colors can be duplicated by a single spectral line – the “spectral locus” on a CIE Chromaticity diagram shows the locations of pure spectral colors on a tristimulus chart. Glaringly missing are “purples” (as opposed to spectral violet or indigo). “purple” requires a mixture of blue and red.
The two-dimensional thing is not, Mangetout, that Venn-diagram-like thing you see. That is a two-dimensional representation of basically a three-coordinate system (there are three colors in that “Venn Diagram”, after all. The CIE Chromaticity diagram is two dimensional only because it normalizes the three components of color – it sacrifices telling you the overall Intensity of the color (the absolute amounts of each of the tristimulus values) and presents them normalized to unit intensity – so that the sum of the amount of each adds up to one.
This is a long and involved topic. See any good comprehensive book on Optics (like the Handbook of Optics, or Warren Smith’s Modern Optical Engineering), or a good book on color theory.
Isn’t part of the reason that violet and red seem somewhat similar is that they’re close to being an “octave” apart? Violet is almost double the frequency of red, isn’t it?
Isn’t the human retina sensitive to ultraviolet? I remember reading that the real reason we can’t see UV is that the lens of the human eye can’t focus red and UV to the same spot. According to this theory, if we saw both red and UV at the same time, the chromatic aberration would be fierce, kind of like when I look through the corner of my glasses and see the refracted images of the green and red traffic signals displaced oppositely from the image of their yellow enclosure… So the lens or cornea of the eye is opaque to UV, and we get clearer vision at the expense of fewer colours.
I’ve heard that, back during the Second World War, cataract patients (who had had their natural lenses replaced with glass) could see UV. They were recruited by the Brits to sit offshore and watch for normally-invisible morse-code signals that were sent by members of the French resistance using UV flashlights…
I’d dearly love to find out whether that story is in fact true.
Imagine: a totally new colour to see! What would it look like? What would that do to our color wheels?
I appreciate what you’re saying, but I don’t know why you say I’m confused. I know that objects in the real world have many colors and that many wavelengths are transmitted from their surfaces. I know that an apple is not purely red, but I don’t know why you thought I didn’t.
But Mange was asking about “red”, not apples. Or so I thought, at least. Maybe I’m confused about that.
Libertarian – sorry if I misquoted or misunderstood. My point is that, unless you’re dealing with a monochromatic source (laser or interference filter or output of a monochromator), any real-world color you see is going to be made up of a range of colors with varying amounts of each. In other words, “red” is going to be a composite of a whole spectrum of colors, so my arguments above are germane.
I’ve heard it said that one of the few places you can see “pure red” (almost spectrally pure single wavelength light) in nature is in a rainbow. It makes sense.
Sound frequencies that are close (I am not sure how close is close) actually are perceived as a third frequency that is the average of the other two, with amplitude pulsing at a rate of the difference of the two frequencies (those pulses are the beats you refer to). (I do not mean that the beats themselves are the third frequency.) [/tangent]
