Ha! Can't wait for a follow-up article, if you'll make one!
I went into a rabbit hole and discovered this YT channel : Ooqui https://www.youtube.com/@ooqui . Lots of experiments with imagining different trichromacy or tetrachromacy, and using software or VR to visualize them. And lots of other stuff too, like using VR to display a different color for each eye, and so trying to induce impossible colors.
(I already made a comment below with this link, but I suppose you didn't had a notification. If you've seen it, sorry for the double comment)
I have deuteranomaly, the illusion works for me, but from the examples in the article the green on purple has a much stronger effect than the others. The one on reddish-orange was quite weak.
Mild protan here (protanomaly - colorblindness of a type common in US males. Have a woman dress you, guys.)
I guess I see a little something around the red circle, but on the whole, kinda meh. As is my relationship to color generally. Those without color deficiency, note: I can _see_ color, to my mind all the colors of the rainbow, but I realize after testing that there's something I'm missing.
Lucky for me the explanation I got after my first Ishihara test was woefully inadequate: "You might have some trouble if you work with electronics for a living: resistors are color coded".
Not believing I had any real color vision problem, I embarked on a visual-design-heavy series of career choices, not knowing the surprises in store for me. Had I really understood my deficiency, I probably would've punted to something else. Sure glad I didn't.
I'm not sure of how this works, but I suspect that by screwing around with the colors appropriately you might be able to get a stronger effect. But not being a protan myself, it's hard for me to tune them!
I've always wanted to know what people with really good visual memory "see" when "remembering" optical illusions (I'm aphantasic myself so I see nothing)
I'm on the opposite side of the spectrum, in that I have a vivid (though not extremely vivid) mind's eye. There are times when I'm thinking about an object with my eyes closed and the color appears as literally as if my eyes were open and looking at the object.
Love this! If you cross/unfocus your eyes, even before the circle starts getting smaller, you get A Lot more of "the color". At least, it worked for me.
Wow, these are striking. The red inside, reddish-orange outside one I could see even before the circle started shrinking, and then once it did the "green" area was massive, larger in width than the radius of the circle.
It turns out to be really hard to stay fixated in one point! If you really force yourself to do that (and don't blink) then the colors all sort of blend together during the static period. But I almost always seem to move my eyes around a little which leads to that effect around the boundary.
Perhaps rather than using a screen with all its RGB limitations, we could try staring into a strong monochromatic light (say red) and then switching to the target light (say green), also monochromatic.
I was thinking about that, but making light of specific wavelengths seems hard! (Typically I think that's a job for lasers?) The easiest method I could think of would be to use optical filters, so you could start with a range of frequencies and then only let a few through? Maybe you could pair that with some material that mostly only emits frequencies in some band? Not sure...
I was looking into this recently for an unrelated project; there are tools for making monochromatic light of tunable wavelengths, but they're in the $10,000+ range.
Heating up a salt of some element that has an emission line only in the visual part of the spectrum should do. Methanol has a mostly colorless flame. It can be temporarily colored like in this video: https://www.youtube.com/watch?v=7hC-KyFraCQ
Perhaps a different approach could be to filter white light. Narrow-band filters exist for astronomy, though we need to match the wavelength and for it to be affordable
I'm incapable of watching videos, sorry. :) But note that we'd really need a green flame in just the right spot, at around 508 nm.
Also, for reasons I don't understand, suddenly switching colors instead of using a shrinking circle doesn't seem to work as well. You can sort of see this by setting the shrink time to something really small like 5 seconds:
She’s just dipping a stick in a solution and then placing it in the flame, which in her case is likely propane or butane though. The need of having a line at exactly 508 nm looks like a stumbling block though.
Edit: realized I should maybe explain why: title of the talk is “Category Theory for Neuroscience.” After introducing lots of math, toward the end the guy explains using category theory to dissolve the inverted spectrum problem. (Idea that “maybe my red is your green whooooaaa”)
I only had time to watch the last 7 or so minutes where he talks about the inverted spectrum problem, but I have a vague familiarity with category theory and yoneda's lemma due to learning Haskell.
I'm not entirely convinced this dissolves the problem. Is the implication that if the spectrum were inverted for some subjects, it would be experimentally detectable by having that subject report their subjective experience of the distance between the different colors? He kind of makes it sound like thinking about the problem in these terms makes the entire "inverted spectrum idea" nonsensical, but I don't see it.
Also, while it might be true that it would be detectable if you merely spun the "color wheel" around, I'm absolutely not convinced you couldn't spin it around and then stretch the relevant parts according to the color gamut to make all the relations line up again like they experimentally do! If that weren't possible, you're saying that there exists no isomorphism and that there is exactly and only one way for the colors to exist; that seems like an extremely strong claim! (and certainly not supported by the evidence given)
I'm pretty sure you've seen this Wikipedia page, but because you didn't mention it: Impossible Colors: https://en.wikipedia.org/wiki/Impossible_color
The illusion you showed is called a "Hyperbolic color" (impossibly highly saturated), in the class of "Chimerical colors". There are other kinds too!
I didn't see it. Which is a shame because apparently this whole post is about https://en.wikipedia.org/wiki/Impossible_color#Chimerical_colors !
Ha! Can't wait for a follow-up article, if you'll make one!
I went into a rabbit hole and discovered this YT channel : Ooqui https://www.youtube.com/@ooqui . Lots of experiments with imagining different trichromacy or tetrachromacy, and using software or VR to visualize them. And lots of other stuff too, like using VR to display a different color for each eye, and so trying to induce impossible colors.
(I already made a comment below with this link, but I suppose you didn't had a notification. If you've seen it, sorry for the double comment)
I did see it, but I just find it almost impossible to watch videos, sorry!
There's also the Ooqui YT channel which plays a lot with colors, including impossible ones, using VR or stereoscopy https://www.youtube.com/@ooqui
Saw a new-ish color! Great post.
I have deuteranomaly, the illusion works for me, but from the examples in the article the green on purple has a much stronger effect than the others. The one on reddish-orange was quite weak.
Thank you! The coolest illusion I've seen.
Mild protan here (protanomaly - colorblindness of a type common in US males. Have a woman dress you, guys.)
I guess I see a little something around the red circle, but on the whole, kinda meh. As is my relationship to color generally. Those without color deficiency, note: I can _see_ color, to my mind all the colors of the rainbow, but I realize after testing that there's something I'm missing.
Lucky for me the explanation I got after my first Ishihara test was woefully inadequate: "You might have some trouble if you work with electronics for a living: resistors are color coded".
Not believing I had any real color vision problem, I embarked on a visual-design-heavy series of career choices, not knowing the surprises in store for me. Had I really understood my deficiency, I probably would've punted to something else. Sure glad I didn't.
Interesting. Can you try this variant?
https://dynomight.net/img/colors/generate.html?inside=ff0000&outside=b06400
I'm not sure of how this works, but I suspect that by screwing around with the colors appropriately you might be able to get a stronger effect. But not being a protan myself, it's hard for me to tune them!
I've always wanted to know what people with really good visual memory "see" when "remembering" optical illusions (I'm aphantasic myself so I see nothing)
I'm on the opposite side of the spectrum, in that I have a vivid (though not extremely vivid) mind's eye. There are times when I'm thinking about an object with my eyes closed and the color appears as literally as if my eyes were open and looking at the object.
...And do you see optical illusions? Even the ones that seem to move?
Love this! If you cross/unfocus your eyes, even before the circle starts getting smaller, you get A Lot more of "the color". At least, it worked for me.
Wow, these are striking. The red inside, reddish-orange outside one I could see even before the circle started shrinking, and then once it did the "green" area was massive, larger in width than the radius of the circle.
> even before the circle started shrinking
It turns out to be really hard to stay fixated in one point! If you really force yourself to do that (and don't blink) then the colors all sort of blend together during the static period. But I almost always seem to move my eyes around a little which leads to that effect around the boundary.
Perhaps rather than using a screen with all its RGB limitations, we could try staring into a strong monochromatic light (say red) and then switching to the target light (say green), also monochromatic.
I was thinking about that, but making light of specific wavelengths seems hard! (Typically I think that's a job for lasers?) The easiest method I could think of would be to use optical filters, so you could start with a range of frequencies and then only let a few through? Maybe you could pair that with some material that mostly only emits frequencies in some band? Not sure...
I was looking into this recently for an unrelated project; there are tools for making monochromatic light of tunable wavelengths, but they're in the $10,000+ range.
I thought that perhaps you could use physical materials, and just make the illustration on a page? E.g.
https://www.thermofisher.com/us/en/home/life-science/cell-analysis/fluorophores/oregon-green.html
https://www.thermofisher.com/us/en/home/life-science/cell-analysis/fluorophores/alexa-fluor-700.html
But I'm not sure... Neither of those look *that* strongly peaked.
Maybe you could find a common laser pointer with a similar enough wavelength?
Heating up a salt of some element that has an emission line only in the visual part of the spectrum should do. Methanol has a mostly colorless flame. It can be temporarily colored like in this video: https://www.youtube.com/watch?v=7hC-KyFraCQ
Perhaps a different approach could be to filter white light. Narrow-band filters exist for astronomy, though we need to match the wavelength and for it to be affordable
I'm incapable of watching videos, sorry. :) But note that we'd really need a green flame in just the right spot, at around 508 nm.
Also, for reasons I don't understand, suddenly switching colors instead of using a shrinking circle doesn't seem to work as well. You can sort of see this by setting the shrink time to something really small like 5 seconds:
https://dynomight.net/img/colors/generate.html?inside=ff0000&outside=008080&size=400&countdown=5&shrink=5&mode=shrink
I'm not sure why this is, but arranging two colors of flames seems extra tricky...
She’s just dipping a stick in a solution and then placing it in the flame, which in her case is likely propane or butane though. The need of having a line at exactly 508 nm looks like a stumbling block though.
This is super cool; somewhat related, this lecture is pretty interesting: https://m.youtube.com/watch?v=4GJ4UQZvCNM
Edit: realized I should maybe explain why: title of the talk is “Category Theory for Neuroscience.” After introducing lots of math, toward the end the guy explains using category theory to dissolve the inverted spectrum problem. (Idea that “maybe my red is your green whooooaaa”)
I only had time to watch the last 7 or so minutes where he talks about the inverted spectrum problem, but I have a vague familiarity with category theory and yoneda's lemma due to learning Haskell.
I'm not entirely convinced this dissolves the problem. Is the implication that if the spectrum were inverted for some subjects, it would be experimentally detectable by having that subject report their subjective experience of the distance between the different colors? He kind of makes it sound like thinking about the problem in these terms makes the entire "inverted spectrum idea" nonsensical, but I don't see it.
Also, while it might be true that it would be detectable if you merely spun the "color wheel" around, I'm absolutely not convinced you couldn't spin it around and then stretch the relevant parts according to the color gamut to make all the relations line up again like they experimentally do! If that weren't possible, you're saying that there exists no isomorphism and that there is exactly and only one way for the colors to exist; that seems like an extremely strong claim! (and certainly not supported by the evidence given)
I don’t usually react strongly to optimal illusions but this one had me (literally) giggling euphorically while watching it. Pretty cool.