Is it supposed to be a frequency of light or a combination of all colors or both? What does it mean when a prism splits up a color?
In: 0
Colour is complicated.
Light is any kind of EM radiation; *visible* light is EM radiation in the part of the spectrum that our eyes respond to.
Every photon has a wavelength; any tiny patch of light can have a vast mix of photons at different wavelengths and intensities.
Think of a violin; there’s no frets on it, so you can play not just any note, but anywhere between notes, as loud or quiet as you want.
Now imagine you can have a million strings on your violin, playing any conceivable monstrous chord, each string at its own pitch and its own volume.
And that’s one million-string violin *per pixel* (or per retinal cell).
There’s just too god damn much information. There’s no way to make a cell that can report all the incoming wavelengths and their exact relative proportions – and even if you could, you’d need a brain the size of an office building to process it all. This is big data on a scale that’s just plain *stoopid*.
So, we cheat by massively simplifying the whole thing down to just four numbers per pixel.
Take four different kinds of retinal cells per tiny patch of retina, three with a response-curve that peaks at a different part of the visible light spectrum, and one that measures overall brightness.
From the relative strength of each cell’s response in each quad, you can reconstruct a three-string approximation of each insane monster chord. It’s not the same thing, not at all, but there’s a consistent translation between the two – which is all we really need.
Something’s only hitting the long-wavelength response? We call that red.
And by comparing that response with the total brightness, we can guess at the mix of frequencies. If the total brightness and low-frequency response are similar, then the overall mix is up around the red-peak in the spectrum, and that’s your bright fire-truck red.
If we’re only seeing a little red response but a bunch of total brightness, then it must be mostly far down the low end, because it’s hitting the tail of the curve – and that’s your deep burgundy territory. If it were hitting the high-end tail of the red response, the mid-frequency (green) cells would be responding too (and that’s where you start getting into oranges and yellows).
Rinse and repeat for the other two peaks, and you can infer a whole bunch about the original mix of wavelengths.
What we call ‘white’ is when *all three* peaks are getting hit equally. Usually this means a broad sweep of frequencies right across the spectrum, a huge blaring _**HNORT**_ of every single string from eg: the giant thermonuclear fireball we call the sun.
However, that doesn’t *have to* be the case: if you peer really close at a white part of your screen (you may need a magnifying glass or a water droplet), you’ll see that it’s actually just red, green and blue light mixed together equally – just three single strings being played in the first place, one wire-skinny little *ploink*.
Our eyes use the same three-note chord to represent *both* cases; we literally can’t tell the difference between them.
And your screen replicates all the other colours the same way. Just three colours out can generate the same visual response as virtually any spread-spectrum smear of wavelengths you want, because we bought our colour perception from Wish.
Except orange. Screens *suck* at producing a believable orange, because maths. There’s just one place where the formula for replicating colours falls down badly, and it’s right there. Don’t believe me, image search the brightest, most vibrant orange you can find anywhere, and hold an actual orange up next to it. It sucks.
Latest Answers