Having a background in sound I am probably coming at this all wrong but if you mix a 1kHz sine wave with a 2kHz sine wave at various amplitudes you will get various different sounding composite sounds but at no point will you be able to emulate say, a 1300Hz sound. How is it that mixing Red Light at 462 terahertz (or whatever) with green light at 545 terahertz (these are numbers I am just pulling off google) at the same amplitude can result in a perceived frequency equivalent to 516 terahertz or as we know it ‘yellow’?
Is it that the ‘yellow’ we experience from Additive colour mixing is not the ‘true’ yellow we see in the rainbow? Is it our eyes that make up the colour based on the input of two discrete light sources interfering with each other?
In: 10
Human perception of color is not quite detecting wavelength. Rather, our eyes have three types of color receptor: red, blue, and green. All color that we see is due to light stimulating these receptors to different degrees. For instance, what we see as yellow is a result of stimulating both the red and green receptors. This can be accomplished by combining both red and green light (as is done by a computer screen) or by producing a single wavelength in between which stimulates both receptors (such as the emission lines of sodium). Our eyes can’t tell the difference.
Some colors, such as pink, can only be produced though a mix of wavelengths.
Our ears have hundreds of tiny frequency-sensitive hairs with attached nerves in the cochlea, so we can very sensitively hear the difference between sound frequencies just fractions of an octave apart.
By comparison, our eyes are pretty much tone deaf. If our eyes were ears, we’d be able to hear just three very broad pitches: “low” (red), “medium” (green) and “high” (blue)
These three pitches would get mixed into different “chords” we call colours: purple is red + blue, yellow is red + green, orange is red + less green, white/grey is red+green+blue, and so on.
But since our eyes are so tone deaf, we can’t see the difference between, say,
* two tones, one “low” and one “medium” (red light + green light)
* a single “medium low” tone (single frequency yellow, eg from a sodium lamp).
These combinations look pretty much the same to us. A spectroscope could tell the difference – eg, a prism could separate red+green into a red band and a green band, but the identical-looking “pure” yellow would yield a single yellow band.
Whether red+green is “true yellow” depends on whether you’re talking about colours as things humans perceive (eg, if you’re choosing colours for a design or art piece), or about (say) “colours” as bands on a spectroscope (eg, if you’re doing analytical chemistry), but it sure looks yellow to us, and that’s often good enough.
It’s because the human eye only has red, green and blue colour receptors (or at least, we have receptors that are more sensitive in those areas) There is another one for luminosity eg- brightness, but we’ll stick with the colours for now.
Since we only really detect those three, all the screen needs to do is stimulate them in the same ratio as a natural colour would, and we get the same effect.
>Is it that the ‘yellow’ we experience from Additive colour mixing is not the ‘true’ yellow we see in the rainbow?
Yes, that’s right. If you shine a mixture of red and green light through an optical prism, the two colors will separate back out. If you shine true yellow light through the prism, there’s nothing to separate.
It’s just that your eyes can’t distinguish those two possibilities. True yellow light activates both “red” and “green” cones by being a color somewhat close to both red and green. The mix of red and green light activates both cones too, by having one component that activates one and another component that activates the other. (I’m simplifying here, but that’s the basic situation.) So you have the same experience in both cases.
Light waves do NOT interfere with each other in the same way sound waves do. Sound waves are waves that occur in a medium (such as air, water, or a solid). Light waves do not have a medium, so they cannot interfere with each other in such a way (though they can create an interference pattern, but that’s for another discussion).
An RGB monitor is NOT mixing any wavelengths. If you zoom in close enough you can see that there’s actually discrete Red, Green, and Blue dots. It cannot create Yellow, or any hues in between.
What’s actually happening is that the red or green or blue sensitive cones in your eyes within close proximity to each other are getting stimulated at the same time, and your brain is interpreting this as a portion of the screen being a specific color based on how stimulated each cone is.
The same thing is happening when you see any colors of the actual wavelength in the natural world. The cones in your eyes aren’t able to see, for example, yellow wavelength, but yellow is close enough to both red and green to where it activates both the red and green cones in your eyes, and your brain interprets this as yellow.
There’s no interference-pattern stuff going on, but you are correct: mixing red and green lights does not produce yellow light.
However, the way our eyes work is a total hack that we can exploit for fun and profit.
A light receptor that could report the full mix of frequencies hitting it would be complex and expensive and high-bandwidth; we simply never evolved anything that could do the job, and we’d be highly unlikely to ever do do.
Instead, each ‘pixel’ on the retina consists of a quad of different cells: one rod cell that just reports monochrome brightness, and three cone cells – one with a red filter, one with a blue filter, and one with a green filter. (there’s real physical blobs of pigment on each cell)
By comparing the relative brightness of the light reaching each of those cone cells, in combination with the overall brightness from the rod celll, your brain can infer the frequency of the light coming in.
(from an audio perspective, imagine having three different-sized baloons tuned to different points up the scale, playing a sound at them and recreating the input frequency from the relative amount of resonance on each)
If mostly the red is getting lit up, then the light must be red. If red and green, it’s somewhere in the yellow range, if mostly green then it’s green, if green and blue then it’s cyan, if mostly blue then it’s blue – and every variation along the way covers all the other colours.
However, you can totally fake those results – shine a red and green light together ferinstance, and it *produces the same response* as a yellow light does, and so you can’t tell the difference.
You could shine a redgreen light on a daffodil in a dark room – the light would look perfectly yellow, but the flower would look weirdly dark, because there’s no yellow light for it to reflect.
You can cover *almost* all of the visible spectrum this way, producing responses indistinguishable from the real thing, except for a couple of caveats:
* You can’t make a plausible orange. Go find the brightest, most vivid photo of an orange you can find on the internet, and compare it to an actual IRL fruit. The onscreen colour is sad and pathetic and dingy by comparison.
* On the other hand, magenta is completely fictional. There is no magenta on the spectrum; a wavelength that somehow hits the red and blue receptors at once, without lighting up the green in the middle… doesn’t exist. You can *only* produce that response by shining red and blue lights together – and that’s why the hue wheel wraps around; you’ve joined the two ends together round the back, as it were.
Latest Answers