Magenta

Magenta doesn't exist. Really, it doesn't.

Mindblowing conclusions abound here, but I figured I'd write about this topic since it's so damned interesting. Okay, there's a light spectrum, that runs from red to violet. You probably know this, and it looks like this:

Red is the longest wavelength and violet is the shortest, with lengths of about 750 nm down to 380 nm. The perception of color in your eye has to do with the three types of "cones" in your eye. Each type of cone has peak sensitivity in a different wavelength - generally separated into red, green, and blue. In actuality, however, these are closer to yellow, green, and violet; perception of the other colors are due to differences in perception in different type of cones. For example, cyan (the light blue between blue and green) has no particular cone cell particularly for it. When the violet cone and yellow cones give roughly the same response to the brain, and the green cone is a bit higher, the brain concatenates this information and creates cyan. Thus, all colors are perceived as a mixture of wavelengths.

Here's another bit of an experiment. Stare at the little black dot in the middle of the yellow and cyan circles below and then stare at the black dot to their right.


For yellow, you should see its opposite in the afterimage: violet. For cyan, you should see red. This is due to cone cells getting "tired" in your eye. Look at it like this:

When you stare at the yellow circle, the yellow cones get very tired, the green cones get pretty tired, and the violet cones don't get very tired at all. The color white is made up of all the constituent colors; violet and yellow and green, it essentially stresses all the cones equally. Since the yellow cones are really tired, they drop out of the equation for a second or two; same goes, to a lesser extent, for the green ones. That chiefly leaves the violet cones to express white to the brain, and the best they can do to pure white is violet. Hence, why the afterimage appears to us as violet.

Now try it with this one:

This time, green receptors get very tired, yellow quite tired, and violet not so tired at all. But if you look at the spectrum above again, you'll notice that green is closer to violet than yellow is, so the violet receptors get a bit more tired than with yellow. Just as with yellow and cyan, one would expect the complement of the color should be perceivable when you stare at the black dot on the right. For green, that complement is magenta, the color halfway between red and violet.

But hold on a second. Try and find magenta on the spectrum:

You'll find it's not on there. What gives? Well, in a lot of image processing programs, the spectrum is like a big circle, with the left side looping back to red eventually.

As previously mentioned, the red-to-violet spectrum goes from longest to shortest wavelengths of light. It's a line, not a wheel, and at its respective ends we have other types of electromagnetic radiation: infrared and ultraviolet. The "northeast" region of the wheel above is actually not on the spectrum. So how can we see it?

When the brain perceives violet and red light wavelengths coming from the same object, it has to process them somehow. It could, similar to what it does for colors like blue or cyan, is just take the "average" of the color and just pick something in the middle. In the linear spectrum, that would cause it to process the color into a shade of green. The other option is just to invent a color with no actual wavelength. The brain takes the second option, and perceives this as magenta, even though magenta has no wavelength.