I subscribe to the physics definition of color and distinguish it from "perception of color." Perception of color is subjective, in that what each person actually perceives may be subject to how their individual brain is wired. Olo is a new perception of color, and perhaps each of the five people who have seen it experienced it differently.
Similar to how we have a discrete number of types of photoreceptors to measure light, we have a discrete number of types of olfactory receptors to measure odor. Similar to how we usually see a combination of light with different wavelengths instead of a single color, we generally smell a combination of chemicals instead of a mono-molecular odor. There is a high variance in how smells are perceived by different people!
Quite true, and certainly the physics definition of color == wavelength is a useful one. But it does lead to some counterintuitive situations...
For example, if you have a flower that reflects red and green wavelengths, we will perceive the flower as yellow. But by the strictly physics definition, you'd say "no that's not actually yellow - yellow has its own (range of) wavelength(s) which that flower doesn't reflect at all". You can imagine a surrealist Magritte painting saying "The flower is not yellow" that creates a little absurdist frission – of course the flower is yellow! (Even if the paint used to re-create that yellow reflects different wavelengths than those of the flower itself!)
You could call that a semantics problem. Detailed understanding of the human visual system is quite a recent thing and so our language doesn't have a natural separation between "wavelength-yellow" (a specific wavelength) and "perceived-yellow" (a slightly fuzzy perception that can be created for most people with different combinations of wavelengths).
Understanding that difference is one of the reasons I wanted to write this in fact! We tend to think that we see the world in an objectively accurate way, but there is quite a lot of construction of our experience happening in our brains - not just for color, but things like perception of 3 dimensional form, separation of objects, recognition of faces, even recognizing living vs. inanimate things.
I tend to grant perceived-color a lot of linguistic ownership over words like "yellow", since perception is what drives these concepts in our brains. The flower is truly yellow and we may associate it with sunshine and lemons even if the wavelengths are very different.
When we want to speak precisely about wavelength-color in a physics context, I'd tend to just use wavelengths as our names. This also has the benefit of breaking our tendency of thinking of visible light as somehow a fundamentally different thing from other wavelengths of light that we can't see and so have no color words – but that's a whole other article I want to write...
Reminds me of an article I read a few years ago (which had graphs for the sensitivity of RGB across wavelengths), sadly I can't remember where - but I like your presentation even more, I think it's more accessible, and I love your graphics - the RGB gauges and the saturated colour bands instead of graphs for the RG sensitivity.
Yeah you'll find the sensitivity charts everywhere, although I was surprised how much those charts weren't explained, which could be quite confusing. And plus trying to wrap your head around a plot like that doesn't necessarily build intuition unless you're already deep into this stuff.
One source of confusion with these charts is that SML and RGB sensitivity are *quite different things*, and for some reason the pop-science explanation of the eye is that we have RGB sensors in our eyes. Even wikipedia has this chart talking about "blue/green/red cones", which is pretty wrong: https://en.wikipedia.org/wiki/Photoreceptor_cell#/media/File:1416_Color_Sensitivity.svg
What's the deal? It turns out those charts are showing two different things, even though they're usually both confusingly presented as something like "color sensitivity of the eye".
The latter chart (with SML labels on the curves) is in fact the true representation of the sensitivity of your actual photoreceptor cones. The former chart (with the blue/green/red labels on the curves) is showing something that's useful to color scientists – a mapping of eye sensitivity to "RGB colorspace".
Basically, real SML sensors in the eye are *not* RGB, but you can mathematically create a pretend RGB eye that (mostly) can sense combinations of colors that match what a real eye sees – mapping the SML curves into equivalent RGB curves. Color scientists do that because display technologies *are* RGB and so it's convenient to pretend there's just a simple match between the wavelengths a display makes and the wavelength sensitivities of the eye – it just makes the math simpler.
And when you do that mapping, you create that kick-up in red sensitivity. Why? Look at the sensitivity curve of S – it's way down into violet – and compare it to the sensitivity curve of blue in the other chart – it's (duh) much more blue and doesn't reach down into violet as much. If you want to recreate the sensitivity of a real SML eye with the pretend RGB eye, you need to create some unique combination of RGB that can create violet and if your blue can't get there on its own, the only way you can do it is to introduce a little red. So that's what the pretend RGB eye chart shows.
But in all of the pop-science stuff, that chart is just shown without explaining that it's just a mathematical convenience for color scientists, not actually what the eye actually sees.
I enjoyed reading this article. Thank you!
I subscribe to the physics definition of color and distinguish it from "perception of color." Perception of color is subjective, in that what each person actually perceives may be subject to how their individual brain is wired. Olo is a new perception of color, and perhaps each of the five people who have seen it experienced it differently.
Similar to how we have a discrete number of types of photoreceptors to measure light, we have a discrete number of types of olfactory receptors to measure odor. Similar to how we usually see a combination of light with different wavelengths instead of a single color, we generally smell a combination of chemicals instead of a mono-molecular odor. There is a high variance in how smells are perceived by different people!
I am looking forward to more articles on color!
Quite true, and certainly the physics definition of color == wavelength is a useful one. But it does lead to some counterintuitive situations...
For example, if you have a flower that reflects red and green wavelengths, we will perceive the flower as yellow. But by the strictly physics definition, you'd say "no that's not actually yellow - yellow has its own (range of) wavelength(s) which that flower doesn't reflect at all". You can imagine a surrealist Magritte painting saying "The flower is not yellow" that creates a little absurdist frission – of course the flower is yellow! (Even if the paint used to re-create that yellow reflects different wavelengths than those of the flower itself!)
You could call that a semantics problem. Detailed understanding of the human visual system is quite a recent thing and so our language doesn't have a natural separation between "wavelength-yellow" (a specific wavelength) and "perceived-yellow" (a slightly fuzzy perception that can be created for most people with different combinations of wavelengths).
Understanding that difference is one of the reasons I wanted to write this in fact! We tend to think that we see the world in an objectively accurate way, but there is quite a lot of construction of our experience happening in our brains - not just for color, but things like perception of 3 dimensional form, separation of objects, recognition of faces, even recognizing living vs. inanimate things.
I tend to grant perceived-color a lot of linguistic ownership over words like "yellow", since perception is what drives these concepts in our brains. The flower is truly yellow and we may associate it with sunshine and lemons even if the wavelengths are very different.
When we want to speak precisely about wavelength-color in a physics context, I'd tend to just use wavelengths as our names. This also has the benefit of breaking our tendency of thinking of visible light as somehow a fundamentally different thing from other wavelengths of light that we can't see and so have no color words – but that's a whole other article I want to write...
A great explanation!
Reminds me of an article I read a few years ago (which had graphs for the sensitivity of RGB across wavelengths), sadly I can't remember where - but I like your presentation even more, I think it's more accessible, and I love your graphics - the RGB gauges and the saturated colour bands instead of graphs for the RG sensitivity.
Thanks Matt!
Yeah you'll find the sensitivity charts everywhere, although I was surprised how much those charts weren't explained, which could be quite confusing. And plus trying to wrap your head around a plot like that doesn't necessarily build intuition unless you're already deep into this stuff.
One source of confusion with these charts is that SML and RGB sensitivity are *quite different things*, and for some reason the pop-science explanation of the eye is that we have RGB sensors in our eyes. Even wikipedia has this chart talking about "blue/green/red cones", which is pretty wrong: https://en.wikipedia.org/wiki/Photoreceptor_cell#/media/File:1416_Color_Sensitivity.svg
That chart (which is repeated in lots of other places) was driving me crazy when I was learning about this stuff. Look at that little kick-up in the "red cone" sensitivity down in the blue range. You don't see that in many other sensitivity charts like this one: https://en.wikipedia.org/wiki/Cone_cell#/media/File:Cone-fundamentals-with-srgb-spectrum.svg
What's the deal? It turns out those charts are showing two different things, even though they're usually both confusingly presented as something like "color sensitivity of the eye".
The latter chart (with SML labels on the curves) is in fact the true representation of the sensitivity of your actual photoreceptor cones. The former chart (with the blue/green/red labels on the curves) is showing something that's useful to color scientists – a mapping of eye sensitivity to "RGB colorspace".
Basically, real SML sensors in the eye are *not* RGB, but you can mathematically create a pretend RGB eye that (mostly) can sense combinations of colors that match what a real eye sees – mapping the SML curves into equivalent RGB curves. Color scientists do that because display technologies *are* RGB and so it's convenient to pretend there's just a simple match between the wavelengths a display makes and the wavelength sensitivities of the eye – it just makes the math simpler.
And when you do that mapping, you create that kick-up in red sensitivity. Why? Look at the sensitivity curve of S – it's way down into violet – and compare it to the sensitivity curve of blue in the other chart – it's (duh) much more blue and doesn't reach down into violet as much. If you want to recreate the sensitivity of a real SML eye with the pretend RGB eye, you need to create some unique combination of RGB that can create violet and if your blue can't get there on its own, the only way you can do it is to introduce a little red. So that's what the pretend RGB eye chart shows.
But in all of the pop-science stuff, that chart is just shown without explaining that it's just a mathematical convenience for color scientists, not actually what the eye actually sees.
Right! I should have said SML instead of RGB. Sorry, can't believe I made that mistake in my comment!
The separation of these concepts is very helpful for understanding!