from the I-see-your-true-colors-shining-through dept.
A new study corrects an important error in the 3D mathematical space developed by the Nobel Prize–winning physicist Erwin Schrödinger and others and used by scientists and industry for more than 100 years to describe how your eye distinguishes one color from another. The research has the potential to boost scientific data visualizations, improve TVs and recalibrate the textile and paint industries.
"The assumed shape of color space requires a paradigm shift," said Roxana Bujack, a computer scientist with a background in mathematics who creates scientific visualizations at Los Alamos National Laboratory. Bujack is lead author of the paper by a Los Alamos team in the Proceedings of the National Academy of Science on the mathematics of color perception. "Our research shows that the current mathematical model of how the eye perceives color differences is incorrect. That model was suggested by Bernhard Riemann and developed by Hermann von Helmholtz and Erwin Schrödinger — all giants in mathematics and physics — and proving one of them wrong is pretty much the dream of a scientist."
[...] In the study, which blends psychology, biology and mathematics, Bujack and her colleagues discovered that using Riemannian geometry overestimates the perception of large color differences. That's because people perceive a big difference in color to be less than the sum you would get if you added up small differences in color that lie between two widely separated shades.
[...] "We didn't expect this, and we don't know the exact geometry of this new color space yet," Bujack said. "We might be able to think of it normally but with an added dampening or weighing function that pulls long distances in, making them shorter. But we can't prove it yet."
Journal Reference:
Roxana Bujack, Emily Teti, Jonah Miller, et al., The non-Riemannian nature of perceptual color space, PNAS, 119, 2022. DOI: 10.1073/pnas.2119753119
(Score: 4, Funny) by zzarko on Friday August 12 2022, @08:00PM
... more irritating ads.
C64 BASIC: 1 a=rnd(-52028):fori=1to8:a=rnd(1):next:fori=1to5:?chr$(rnd(1)*26+65);:next
(Score: 4, Insightful) by looorg on Friday August 12 2022, @08:05PM (3 children)
So they don't know what the new geometry is and they can't prove it yet. Probably due to it apparently has to do with how things are perceived so there might be massive individual differences. So this new one could be wrong, or incorrect or remain un-proven for some time. But they know the old idea is wrong, or at least not complete. Yet somehow not so wrong that we can't have used it successfully for about 100 years give or take. Clearly we are at least in the ballpark.
That said I don't know if I really have a use for more vibrant colours. Not to mention that it will just be used to show more annoying commercials --- NOW WITH NEW VIBRANT COLOURS NEVER SEEN BEFORE!
(Score: 1, Interesting) by Anonymous Coward on Friday August 12 2022, @09:15PM
Came here to say about the same thing. There have been multiple different color spaces/models in use over the years, so it's not very surprising that someone is working toward yet another. I don't remember all the details we studied in a "Color and Design" class (art for engineers) many years ago, but this article could be a good starting point,
https://en.wikipedia.org/wiki/List_of_color_spaces_and_their_uses [wikipedia.org]
One of our dorm RAs was doing graduate research in color and her results were squelched by her advisor... Never learned the details, but it seemed like her research results disagreed with the theories of a large donor to the university (who had done color vision research during an earlier part of his career). That donor died not long after, but I think the RA had already moved her graduate work to another topic. It's not pleasant, but these things to happen, $$$$$ count in all sorts of unexpected ways.
(Score: 2) by krishnoid on Friday August 12 2022, @10:41PM
I bet some are so incredible they'd overwhelm your emotions [youtu.be]. It is the future, after all.
(Score: 2, Interesting) by Anonymous Coward on Saturday August 13 2022, @02:40AM
I can't say without reading their paper that the article headline matches reality in terms of "overturning" the 100 year old model. It puts it on par with Einstein correcting Newton, but the Schrodinger color theory was known to have issues for a very long time.
For example, a very interesting Schrodinger paper commentary is here [sunyopt.edu], where some issues are brought up as well as talking about some of Schrodinger's work. If you are comfortable with linear algebra, than it is an enjoyable read.
(Score: 1, Interesting) by Runaway1956 on Friday August 12 2022, @08:07PM (6 children)
What's the difference between lavender, lilac, clematis, salvia, heliotrope, iris, petunia, or any other flavor of purple? They all get lost in the greenery, right along with those red roses.
Abortion is the number one killed of children in the United States.
(Score: 4, Insightful) by Snotnose on Friday August 12 2022, @08:35PM (2 children)
If you're a guy it's whether or not your gay.
If you're a chick your normal.
If you're an interior decorator, $$$
I just passed a drug test. My dealer has some explaining to do.
(Score: 5, Informative) by Anonymous Coward on Saturday August 13 2022, @04:34AM (1 child)
Shut your piehole you violent liar.
https://blog.xkcd.com/2010/05/03/color-survey-results/ [xkcd.com]
(Score: 5, Touché) by FatPhil on Saturday August 13 2022, @06:30AM
XKCD supports the facts behind snotnose's post, even if it wouldn't support the presentation of them.
Great minds discuss ideas; average minds discuss events; small minds discuss people; the smallest discuss themselves
(Score: 3, Informative) by aafcac on Friday August 12 2022, @10:13PM (1 child)
It's purely about marketing. There are multiple incredibly efficient ways of specifying color using numbers. They're far more intuitive for most people once you understand how they work. As opposed to the names, which are almost completely meaningless without training. No reasonable person would know the difference between titanium and zinc whites. Most painters would know that one is warmer than the other.
(Score: 2, Interesting) by aafcac on Saturday August 13 2022, @04:00PM
I see I upset somebody in the marketing department at Behr Paints. Anybody that claims that the names have anything to do with anything other than marketing needs their head examined. Even for things where you've got a reasonable thing to model the color off of, there's no guarantee that it will be universally accepted across the entire world, and chances are it won't. Depending on what animal laid the egg, eggshell could have been any number of colors, some of them not even remotely white. I regularly see eggs at the local grocery store in brown and blue.
(Score: 2) by mcgrew on Saturday August 13 2022, @08:34PM
Also, if you place a cool color of the same intensity as a warm color next to it; say, red and blue, the cool color will recede. My knowledge of color is from an ancient physics class almost a half a century ago, which posited that color is determined by the size of the receptor, which determines what frequency of EMF is recepted.
This means that most likely, you and I don't see exactly the same colors. We see the same colors and call them by the same name, but I'd bet as soon as they can transplant eyeballs (not just corneas as now), the patient will see colors very differently, because it's unlikely any two people will have exactly the same sizes of color receptors.
I've wondered this since I was a kid; my dad and his brother were both color deficient and saw color, but couldn't tell red from green.
Carbon, The only element in the known universe to ever gain sentience
(Score: 2) by HammeredGlass on Friday August 12 2022, @08:27PM
is a welcome mat for disappointment 6 months down the road.
(Score: 4, Informative) by ChrisMaple on Friday August 12 2022, @08:58PM (9 children)
The most vibrant images are those with areas of monochromatic primaries or 2 or the three primaries, for emissive displays. Nothing in the article changes that; they haven't discovered new best wavelengths for the primaries.
What they've done is to discover new perceptual differences for 1+1 not = 2 that differ from the long-accepted differences. This might result in better predictions for the result of mixing pigments, or better color matching. Beyond that, it's little more than another curiosity in the understanding human perception.
Note that the results may not be equally applicable for all people, as there is some variability in the spectral response of human photo-receptors.
(Score: 2) by crafoo on Friday August 12 2022, @09:55PM (4 children)
Color perception is pretty weird. Here is someone on gfx design stack exchange asking something similar: asking for a mathematical relationship to understand color perception given surrounding colors. https://graphicdesign.stackexchange.com/questions/75773/correction-for-color-constancy [stackexchange.com]
(Score: 3, Interesting) by aafcac on Friday August 12 2022, @10:20PM (2 children)
It is, and because of individual variations in perception it's difficult to come up with a comprehensive model. My cones seem to be spaced a bit oddly, so colors in the greens are hard for me to tell apart, but I do see colors slightly outside the normally visible range. Which is super helpful at night, but sometimes I struggle to see things that haven't been heated up. And, the only time I'm ever completely unable to see is when everything is the same temperature, like in a walk in freezer.
It's part of why people should be skeptical that everybody is perceiving the same color. Chances are that we're not, it's just usually close enough.
(Score: 2) by hendrikboom on Saturday August 13 2022, @01:58AM (1 child)
Could your vision be tetrachromatic?
(Score: 1) by aafcac on Saturday August 13 2022, @03:49PM
From what I understand, probably not. I'm more likely an anomalous trichromat where the cones have sensitivity that's offset a bit more than normal. Normal trichromatic color vision will have a bit of duplication over the green bit of the spectrum, presumably it's because that color is both rather important and it guarantees that you don't have weird holes in the middle of the Spectrum.
I think it's rather interesting that there's a bunch of things like this in the general population that don't get much attention, but could be of vital importance if society ever fails and we need to be able to do things like keep watch at night without modern technology. Fortunately, that's not a likely outcome.
(Score: 3, Interesting) by hendrikboom on Saturday August 13 2022, @02:02AM
Years ago I heard of an experiment in which people were looking from a room lighted in one colour through a hole where they saw into another room in another colour.
Within experimental error (about 10%) they discovered that the shift in perception of the colour seen through the hole could be modeled by the Lorentz transformation.
Wish I still had a reference.
(Score: 1) by aafcac on Friday August 12 2022, @10:24PM (2 children)
No, but if the model is off enough, there would be better ways of slicing it up than what we're currently doing. Bayer pattern arrays are based on the notion that green is the most important color for midtones and midway between blue and red in terms of sensitivity. If either of those things isn't true, then giving twice as many pixels to green as either red or blue might not make sense. For example, a different shape of a layout where green isn't given equal weight to blue and red combined could be a reasonable situation.
(Score: 0) by Anonymous Coward on Saturday August 13 2022, @02:19AM (1 child)
I believe they've measured the snot out of the photopic curve, so I doubt that would be the case.
(Score: 1) by aafcac on Saturday August 13 2022, @03:54PM
I don't disagree, I think that if it were that far off, we'd know. More likely would be if the change could create a new and improved way of compressing images similar to how MP3 works, but even that is probably a stretch as we'd probably be noticing significant issues with the current technology by now.
(Score: 0) by Anonymous Coward on Saturday August 13 2022, @06:21AM
I think one has to consider the fact that the brain tends to emphasize contrast. It looks for what's different between two objects. The whole point is that if everything looked the same then we couldn't tell anything apart.
Take the Checker shadow illusion
https://en.wikipedia.org/wiki/Checker_shadow_illusion [wikipedia.org]
If you have a relatively dark object with a darker background (or perhaps if you have a colored object in a background with a similar color) it appears brighter than if you placed that same object in a brighter background (or if you placed that colored object in a different background of a different color this might alter your perception of the object's color so as to improve contrast). The perceived shade/tint (or perhaps color) of the object changes so as to improve contrast.
Notice that with the checker shadow illusion with the connecting line that the two squares do appear to be the same shade (the image linked below only has one connecting line). If you try to cover the connecting line with your finger notice how at least some of perceived shade difference seems to reappear. Move your finger out of the way again and the perceived shade difference disappears.
https://en.wikipedia.org/wiki/Optical_illusion#/media/File:Grey_square_optical_illusion_proof2.svg [wikipedia.org]
However, it can be difficult for someone to remember the exact shade/tint/color of an object when compared to another object. If we see two cars at two different times we might be able to tell that they are a different shade of blue. But if the blues are too similar we might not be able to do a great job remembering the difference if we see them at different times (unless we maybe saw them multiple times?). Put the car side by side closely enough and it's easier for us to spot slightly different shades because now our eyes exaggerate the contrast. Things that may also affect our perception are lighting conditions and how the surrounding environment looks. This may contribute to why it's more difficult for us to remember the differences between two cars that are only slightly different in color, because our brain knows that they may appear different in different conditions so it only retains an approximate color. Put them side by side and our brain knows they are in approximately the same lighting condition with approximately the same environmental variables and so it'll exaggerate the differences so as to point them out to us.
"That's because people perceive a big difference in color to be less than the sum you would get if you added up small differences in color that lie between two widely separated shades."
Because if you put two small different colors/shades together your eyes exaggerate the difference so as to emphasize contrast. That perceived exaggerated difference equals what you would see had you put two objects with stronger contrasting colors/shades together.
(Score: 3, Interesting) by unhandyandy on Saturday August 13 2022, @02:40AM (4 children)
(Score: 1, Insightful) by Anonymous Coward on Saturday August 13 2022, @03:59PM (3 children)
I think human perception depends on the relative differences of two objects being compared. If we are comparing two very different objects we don't need to amplify the differences as much. If we are comparing two similar objects then we amplify the differences much more.
Let's say we are comparing the distance between the earth and the sun and the distance between Mars from the sun.
The (average?) distance between earth and the sun is 94.177 million mi
The (average?) distance between Mars and the sun is 131.02 million mi
Let's say the exact distance between the earth and the sun is 94.177187231 while the exact distance between Mars and the sun is 131.021347921 million mi. If our aim is to approximately compare the distance between Earth and the sun to the distance between Mars and the sun we would drop a few of those significant figures and round. 131 million miles vs 94 million miles. Our brain isn't going to magnify the differences and it's not really going to remember them either.
Now if we were comparing the distance between the moon and the sun to the distance between Earth and the sun at a given time all of a sudden more significant figures becomes relevant. So we amplify the differences and we may be more inclined to remember more significant figures.
If we are comparing two very similar objects we amplify the differences so that we can make a more relevant comparison and we are more inclined to remember more significant figures. If we are comparing two very different objects our brains do a lot of rounding and approximating because all of those significant figures become less relevant.
The same thing can be said about comparing different shades/tins and colors that we are looking at.
(Score: 0) by Anonymous Coward on Wednesday August 17 2022, @01:18AM (2 children)
A related video I found is the following
Future Computers Will Be Radically Different (Analog Computing)
Mar 1, 2022
https://www.youtube.com/watch?v=GVsUOuSjvcg [youtube.com]
Notice how in 19:10
You can convert from the analog domain (while the answer is still close enough to reliably extract a correct digital result) to the digital domain (basically, you are correcting the error, kinda like how auto-tune will correct a pitch that's only slightly off and bring it to the correct pitch because it knows the intended pitch is likely the nearest octet. Analog computations create more and more errors the more computations you have to do but if the intended computation is clear in between computations you can correct it to the exact intended value before sending it off to the next computation) and send the digitally corrected result through the next analog computation.
You don't even have to necessarily do the error correction in between every analog computation. If, say, after 5 computations the result is still close enough for the intended value to be reliably inferred and exactly correct to (because your analog hardware spits out results precisely enough to not lose a lot of precision/accuracy after five computations) you can then make the digital correction after those five computations before sending it off to the next five analog computations. The more corrections you do in between fewer analog computations the slower the process will be (but the better the end result), the fewer corrections you do in between more analog computations the more accuracy you are likely to lose depending on how precise/imprecise your analog hardware is built.
(Score: 0) by Anonymous Coward on Wednesday August 17 2022, @01:24AM
errr ... sp: octet should read octave *
(Score: 0) by Anonymous Coward on Wednesday August 17 2022, @01:37AM
(in addition to digital error corrections of analog outputs there would also be nothing wrong with doing certain digital computations in between certain analog computations).