New Look, Same Great Look
May 17, 2022 7:32 AM Subscribe
Art professor Kim Beil in Lapham's Quarterly on the advent of color photography and the complicated questions of the perception of color. Color is among the most challenging aspects of our experience to describe. Spectrophotometers and colorimeters can quantify light waves, yet their measurements have little impact on our feeling for color. As the philosopher Zeno Vendler put it, “Vincent van Gogh loved the color yellow—and certainly not because of its wavelength.” Color is infamous for its variability in language and perception. How can we know that what we are seeing is the same as what someone else sees? How can we separate what we are seeing from the thing itself? Or, as Ludwig Wittgenstein asked in his Remarks on Colour, “Where do we draw the line here between logic and experience?” In the Remarks, written the year before his death in 1951, the philosopher’s thoughts about color invariably lead back to the study of philosophy. What things are knowable? How are they known? What can be determined through philosophical reasoning? Wittgenstein reflected, “Colors are a stimulus to philosophizing.”
How can we know that what we are seeing is the same as what someone else sees?
Well, when people are babies, we spend lots of time telling them what color things are, and once they get it right often enough, they are on their own. If they don't get it right quickly enough, then we assume they have a vision problem. I assume if you had a really cruel parent and a cruel school system, you could switch things up and really mess with some poor child.
Philosophically though, I just don't see how visual mediums like the pictures shown would work if we all didn't see colors relatively similarly. Like the 2nd shot of the blue edging around the dress, the gold pendant - the technicolor shawl. If color-vision was specific to the individual I can't imagine how that piece would work as art.
posted by The_Vegetables at 9:05 AM on May 17, 2022
Well, when people are babies, we spend lots of time telling them what color things are, and once they get it right often enough, they are on their own. If they don't get it right quickly enough, then we assume they have a vision problem. I assume if you had a really cruel parent and a cruel school system, you could switch things up and really mess with some poor child.
Philosophically though, I just don't see how visual mediums like the pictures shown would work if we all didn't see colors relatively similarly. Like the 2nd shot of the blue edging around the dress, the gold pendant - the technicolor shawl. If color-vision was specific to the individual I can't imagine how that piece would work as art.
posted by The_Vegetables at 9:05 AM on May 17, 2022
As with most things Wittgenstein, it's not so much philosophy as semiotics. Objectively, the differences we see are differences in wavelength, but then we give each point on the spectrum that we care to distinguish a name, like dusky rose, or cerulean, and we try to reach a consensus on that name and its wavelength so we can communicate meaningfully about it. It gets hazy when each of us has an aesthetic preference that lies between familiar conventional names. Try to send someone to the paint store to get the color you want and you will likely be disappointed.
posted by OHenryPacey at 9:18 AM on May 17, 2022 [2 favorites]
posted by OHenryPacey at 9:18 AM on May 17, 2022 [2 favorites]
I took a graduate level class in color capture and reproduction; the whole process of taking a photograph and then displaying it with correct colors. It's wacky how complex this is, every part of the path the color data takes is subject to distortion and errors. And then there's how your brain actually perceives things with a whole gamut of fun optical illusions.
My favorite color illusion is chimerical colors: you overload your retina with one color and then expose it to a contrasting color. And then what you perceive is a phantom third color that does not exist in any physical sense. For instance "stygian blue", the perception of a blue that is intensely saturated blue but also very black. The effect only takes about 20 seconds, this is a good demo video.
On a totally different topic, capturing skin tone colors in photographs has been a topic fraught with racism for 100+ years. Google has a team working on improving this for its phone cameras, they recently released a new reference color scale ("Monk Skin Tone Scale") to improve how people measure and describe skin tones in digital imaging.
posted by Nelson at 9:29 AM on May 17, 2022 [10 favorites]
My favorite color illusion is chimerical colors: you overload your retina with one color and then expose it to a contrasting color. And then what you perceive is a phantom third color that does not exist in any physical sense. For instance "stygian blue", the perception of a blue that is intensely saturated blue but also very black. The effect only takes about 20 seconds, this is a good demo video.
On a totally different topic, capturing skin tone colors in photographs has been a topic fraught with racism for 100+ years. Google has a team working on improving this for its phone cameras, they recently released a new reference color scale ("Monk Skin Tone Scale") to improve how people measure and describe skin tones in digital imaging.
posted by Nelson at 9:29 AM on May 17, 2022 [10 favorites]
If color-vision was specific to the individual I can't imagine how that piece would work as art.
On the other hand, art is notoriously subjective; some pieces of art will "work" for some people and others will absolutely fail to move them. And I've always wondered if individual variations in how we perceive colors might not account for some of that.
posted by mstokes650 at 9:32 AM on May 17, 2022
On the other hand, art is notoriously subjective; some pieces of art will "work" for some people and others will absolutely fail to move them. And I've always wondered if individual variations in how we perceive colors might not account for some of that.
posted by mstokes650 at 9:32 AM on May 17, 2022
Well, when people are babies, we spend lots of time telling them what color things are, and once they get it right often enough, they are on their own. If they don't get it right quickly enough, then we assume they have a vision problem.
The issue is that you're talking about language, which Wittgenstein famously worked hard to show is not 1:1 with individual perception. The parent is teaching the child the labels for colors. If the child successfully learns to call [color parent calls green] "green", it doesn't necessarily mean that [color child perceives when looking at color parent is asking about] is the same.
posted by star gentle uterus at 9:32 AM on May 17, 2022 [4 favorites]
The issue is that you're talking about language, which Wittgenstein famously worked hard to show is not 1:1 with individual perception. The parent is teaching the child the labels for colors. If the child successfully learns to call [color parent calls green] "green", it doesn't necessarily mean that [color child perceives when looking at color parent is asking about] is the same.
posted by star gentle uterus at 9:32 AM on May 17, 2022 [4 favorites]
As someone who occasionally dabbles in photography I find color and its perception fascinating. Why does the color wheel work even though the visible spectrum is linear? “How do we know that the colors we see are the same as the colors someone else sees?” leads to the inverted spectrum problem and the wider concept of qualia.
There is also the theory of color espoused by Calvin’s dad.
posted by TedW at 9:39 AM on May 17, 2022 [3 favorites]
There is also the theory of color espoused by Calvin’s dad.
posted by TedW at 9:39 AM on May 17, 2022 [3 favorites]
Obligatory link to Margaret Livingstone's Vision and Art: The Biology of Seeing (non referral Amazon link)
I found her book helped to explain how our visual system works and thus why some combinations of light, color, and contrast create certain effects.
posted by SegFaultCoreDump at 10:12 AM on May 17, 2022 [2 favorites]
I found her book helped to explain how our visual system works and thus why some combinations of light, color, and contrast create certain effects.
posted by SegFaultCoreDump at 10:12 AM on May 17, 2022 [2 favorites]
Color is like language: because (almost) everyone experiences it, people grossly overestimate how much they know or can know about how it works via nothing more than introspection.
posted by Ivan Fyodorovich at 12:06 PM on May 17, 2022 [5 favorites]
posted by Ivan Fyodorovich at 12:06 PM on May 17, 2022 [5 favorites]
There was an interesting installation at the SF MoMA a dozen or so years ago. A series of rooms was illuminated with monochromatic sodium vapor lamps. When you first entered everything was amber-hued but after a short accommodation everything was simply black and white. IIRC, you were issued a white-light flashlight which you could shine on objects and when you did the "real" colors popped out. It was quite an experience.
posted by sjswitzer at 12:12 PM on May 17, 2022 [3 favorites]
posted by sjswitzer at 12:12 PM on May 17, 2022 [3 favorites]
We're all familiar with how computer displays simulate color using just three colors, red, green and blue, but it's interesting how these three colors simulate colors of the spectrum.
First, your color receptors (cones) are not tuned to those exact colors. In fact your "red" and "green" cones have a strikingly similar response curve since they evolved recently from the same cone (our ancestors had just two cones) with a small mutation. The three colors in displays correspond not to the peaks of the different cone responses but to their maximum differences. A spectral color is simulated by stimulating the cones so that their responses are the same as if the pure spectral color were present. This presents a problem of "gamut" since there are spectral colors (and intensities) that can't be perfectly simulated. This is not usually a problem because they occur in washed-out or dark areas where color discrimination is poor anyway. Conversely, we can see colors, magenta specifically, that don't correspond to a pure spectral color at all. There is no magenta in the rainbow. Magenta is something we see when an object absorbs mostly green, leaving mostly red and blue light.
In short, the colors we see are as much a product of how the human eye works as the objects themselves. Of course, not everyone has three different cone types, some (mostly XY) have but two and some (mostly XX) have four!
To further complicate things, there is individual variation in the pigments of the "standard cones" so that it may very well be the case that different people do indeed see colors differently (aside from chromaticity). Although they may identify colors the same as others do, their "standard" greens might be subjectively bluer or yellower, though there might be no way recognize or communicate this difference.
Tangentially, you've probably seen amber-colored images from the James Webb telescope recently. Of course infrared waves aren't amber so this is just a choice. The image could equally well be black and white. But as it happens there's a reason they use amber. In a black and white image, it's difficult to see detail in the lightest and darkest parts of the image. So instead, the image goes from black to amber and then from amber to white. This gives at least twice the dynamic range and apparently helps us resolve detail in the extremes (possibly by messing with the gamma curves).
posted by sjswitzer at 12:42 PM on May 17, 2022 [10 favorites]
First, your color receptors (cones) are not tuned to those exact colors. In fact your "red" and "green" cones have a strikingly similar response curve since they evolved recently from the same cone (our ancestors had just two cones) with a small mutation. The three colors in displays correspond not to the peaks of the different cone responses but to their maximum differences. A spectral color is simulated by stimulating the cones so that their responses are the same as if the pure spectral color were present. This presents a problem of "gamut" since there are spectral colors (and intensities) that can't be perfectly simulated. This is not usually a problem because they occur in washed-out or dark areas where color discrimination is poor anyway. Conversely, we can see colors, magenta specifically, that don't correspond to a pure spectral color at all. There is no magenta in the rainbow. Magenta is something we see when an object absorbs mostly green, leaving mostly red and blue light.
In short, the colors we see are as much a product of how the human eye works as the objects themselves. Of course, not everyone has three different cone types, some (mostly XY) have but two and some (mostly XX) have four!
To further complicate things, there is individual variation in the pigments of the "standard cones" so that it may very well be the case that different people do indeed see colors differently (aside from chromaticity). Although they may identify colors the same as others do, their "standard" greens might be subjectively bluer or yellower, though there might be no way recognize or communicate this difference.
Tangentially, you've probably seen amber-colored images from the James Webb telescope recently. Of course infrared waves aren't amber so this is just a choice. The image could equally well be black and white. But as it happens there's a reason they use amber. In a black and white image, it's difficult to see detail in the lightest and darkest parts of the image. So instead, the image goes from black to amber and then from amber to white. This gives at least twice the dynamic range and apparently helps us resolve detail in the extremes (possibly by messing with the gamma curves).
posted by sjswitzer at 12:42 PM on May 17, 2022 [10 favorites]
Great essay.
I used to work with Dolby Vision-process video content. Dolby Vision is a vertically-integrated method to display images with a broader gamut of colors and brightnesses than previous methods, with hooks in every stage of the image pipeline, from camera to obsolete mobile phone screens. It has a rigorous psychosensory base. Even for entry-level workers in this field, DoVi introductory texts tend to upend years of technical expertise these workers thought they had (e.g. the typical human eye does not have individual structures that react with red, green and blue light per se; they manifest as structures sensitive to relatively long, medium and short wavelengths of light that in the population generally, but not in any given individual exactly, correspond to what we think of as red, green and blue).
Anyway, Dolby Vision is more sophisticated and in some ways simpler than the lossy video compression schemes of the past, which all used various permutations of color subsampling, pixel block quantization, motion substitution, and other forms of inter- and infra-frame entropy to recreate the fluid motion of the universe within the limited data bandwidth available to human information systems. TV and film have always thrown out unimportant parts of action to capture what filmmakers believe are the important parts for exhibition. Dolby Vision operates almost entirely on a color subsampling/limiting basis; it analyzes a single picture frame encoded with 16 bits of relative precision for every independent red, green and blue component value per pixel and computes equivalent dependent values that can be accurately represented in a picture that only requires 12 bits of relative precision within an absolute range. It then passes this new image with its changed optical-electric transfer function down the image pipeline. This feat of perceptually-lossless-when-compared-to-nature image compression is largely thanks to over a century of publicly-available research on the relative imprecision and inaccuracy of the ocular light-sensing capabilities of human beings.
As noted above, video delivery systems often make use of further compression schemes, which is why Dolby Cinema films still have mosquito noise visible around edges of objects in frames and color banding over large homochromic areas of a shot. But Dolby’s key innovation is measuring, quantifying and emulating a more precise model of how we see and sense color than the simple logarithmic “gamma” curves of the last century and applying that model to every step of filmmaking. The article notes that these gamma curves, or transfer functions, are informed by and often conform to the requirements and desires of white supremacy and hegemony. In stark 10,000-to-1 contrast, Dolby Vision demo reels revel in the accurate reproduction of the skin tones of all of humanity, not just beautiful white American girls in red swimsuits throwing red, green and blue beachballs.
tl;dr Advances in color science, capture and exhibition allow artists to depict Others closer to how they actually are, which is, as always, more like us than our rulers would have us believe.
posted by infinitewindow at 12:44 PM on May 17, 2022 [2 favorites]
I used to work with Dolby Vision-process video content. Dolby Vision is a vertically-integrated method to display images with a broader gamut of colors and brightnesses than previous methods, with hooks in every stage of the image pipeline, from camera to obsolete mobile phone screens. It has a rigorous psychosensory base. Even for entry-level workers in this field, DoVi introductory texts tend to upend years of technical expertise these workers thought they had (e.g. the typical human eye does not have individual structures that react with red, green and blue light per se; they manifest as structures sensitive to relatively long, medium and short wavelengths of light that in the population generally, but not in any given individual exactly, correspond to what we think of as red, green and blue).
Anyway, Dolby Vision is more sophisticated and in some ways simpler than the lossy video compression schemes of the past, which all used various permutations of color subsampling, pixel block quantization, motion substitution, and other forms of inter- and infra-frame entropy to recreate the fluid motion of the universe within the limited data bandwidth available to human information systems. TV and film have always thrown out unimportant parts of action to capture what filmmakers believe are the important parts for exhibition. Dolby Vision operates almost entirely on a color subsampling/limiting basis; it analyzes a single picture frame encoded with 16 bits of relative precision for every independent red, green and blue component value per pixel and computes equivalent dependent values that can be accurately represented in a picture that only requires 12 bits of relative precision within an absolute range. It then passes this new image with its changed optical-electric transfer function down the image pipeline. This feat of perceptually-lossless-when-compared-to-nature image compression is largely thanks to over a century of publicly-available research on the relative imprecision and inaccuracy of the ocular light-sensing capabilities of human beings.
As noted above, video delivery systems often make use of further compression schemes, which is why Dolby Cinema films still have mosquito noise visible around edges of objects in frames and color banding over large homochromic areas of a shot. But Dolby’s key innovation is measuring, quantifying and emulating a more precise model of how we see and sense color than the simple logarithmic “gamma” curves of the last century and applying that model to every step of filmmaking. The article notes that these gamma curves, or transfer functions, are informed by and often conform to the requirements and desires of white supremacy and hegemony. In stark 10,000-to-1 contrast, Dolby Vision demo reels revel in the accurate reproduction of the skin tones of all of humanity, not just beautiful white American girls in red swimsuits throwing red, green and blue beachballs.
tl;dr Advances in color science, capture and exhibition allow artists to depict Others closer to how they actually are, which is, as always, more like us than our rulers would have us believe.
posted by infinitewindow at 12:44 PM on May 17, 2022 [2 favorites]
Studying colour across language shows how differently people see things even when we think we recognise the words:
most of the world’s languages have five basic colour terms. Cultures as diverse as the Himba in the Namibian plains and the Berinmo in the lush rainforests of Papua New Guinea employ such five term systems. As well as dark, light, and red, these languages typically have a term for yellow, and a term that denotes both blue and green. That is, these languages do not have separate terms for “green” and “blue” but use one term to describe both colours, a sort of “grue”
And again, what this says about how much we all - individually, culturally, linguistically, semiotically - create our world when we often think we are just reflecting it back at each other is fascinating.
A last point: I have no visual imagination. It's a thing, but I don't "see" things mentally. Even if I saw them the same way as most of you, I couldn't remember them the same way. So yeah. Brains are bloody wonderful.
https://theconversation.com/amp/the-way-you-see-colour-depends-on-what-language-you-speak-94833
posted by onebuttonmonkey at 12:46 PM on May 17, 2022 [1 favorite]
most of the world’s languages have five basic colour terms. Cultures as diverse as the Himba in the Namibian plains and the Berinmo in the lush rainforests of Papua New Guinea employ such five term systems. As well as dark, light, and red, these languages typically have a term for yellow, and a term that denotes both blue and green. That is, these languages do not have separate terms for “green” and “blue” but use one term to describe both colours, a sort of “grue”
And again, what this says about how much we all - individually, culturally, linguistically, semiotically - create our world when we often think we are just reflecting it back at each other is fascinating.
A last point: I have no visual imagination. It's a thing, but I don't "see" things mentally. Even if I saw them the same way as most of you, I couldn't remember them the same way. So yeah. Brains are bloody wonderful.
https://theconversation.com/amp/the-way-you-see-colour-depends-on-what-language-you-speak-94833
posted by onebuttonmonkey at 12:46 PM on May 17, 2022 [1 favorite]
sjswitzer’s description of the LMS visual system is much, much clearer and better than mine!
posted by infinitewindow at 12:46 PM on May 17, 2022 [1 favorite]
posted by infinitewindow at 12:46 PM on May 17, 2022 [1 favorite]
How can we know that what we are seeing is the same as what someone else sees?
Well, try working at digital service bureau for any extended period of time, and you'll be able to answer that question pretty quick.
When you're dealing with customers who were paying for images that required color matching, all your high-falutin semiotic bushwah can get tossed right out the window.
Here's what I learned in my stint in the color trenches:
I could explain the color inconsistencies of slide dupes fairly convincingly, but when it came to prints: trying to explain that the color model of slides (transparency color) and prints (reflective color) were completely different was often extremely challenging.
More than once I had an artist bring in their original artwork to show me how the colors of the print didn't match the original.
The fact that the slide they provided me didn't match the painting to begin with was often a bitter pill to swallow.
posted by jeremias at 1:29 PM on May 17, 2022 [4 favorites]
Well, try working at digital service bureau for any extended period of time, and you'll be able to answer that question pretty quick.
When you're dealing with customers who were paying for images that required color matching, all your high-falutin semiotic bushwah can get tossed right out the window.
Here's what I learned in my stint in the color trenches:
- I got pretty damn good at color matching, but the first few months as I learned the ropes were stressfull as hell. It's both art and science, and the science part has a lot of variables to deal with.
- I also became extremely good at explaining to customers why true color matching was going to be impossible. Ideally this conversation happened before, not after, the project was submitted.
I could explain the color inconsistencies of slide dupes fairly convincingly, but when it came to prints: trying to explain that the color model of slides (transparency color) and prints (reflective color) were completely different was often extremely challenging.
More than once I had an artist bring in their original artwork to show me how the colors of the print didn't match the original.
The fact that the slide they provided me didn't match the painting to begin with was often a bitter pill to swallow.
posted by jeremias at 1:29 PM on May 17, 2022 [4 favorites]
jeremias! Printing is truly at the intersection of art and science and the fact that it can be done well at all is a testament to your skills (and struggles)!
I was once dragooned into an attempt to commodify print services. The whole thing was insane. Aside from the fact that there were more hands than cookie jar in the collaboration (always a sign of a bad business plan), it's clear that unless you're doing something that's already commodified, there's such a huge amount of shared experience and trust relationships between print services and customers that you can't commodify it. The customer needs to understand the limits of the process and the provider needs to understand the wants and needs of the customer. There are so many free variables that can never be quantified, but an iterative relationship between the customer and provider can produce good results as they learn to understand each other's wants and capabilities. The worst result is a loading dock full of unusable product.
Needless to say, this project went nowhere and my objections were (mostly) ignored. I like to think that maybe we spent less time pursuing a pointless venture than we might have otherwise, but who knows?
posted by sjswitzer at 3:06 PM on May 17, 2022 [2 favorites]
I was once dragooned into an attempt to commodify print services. The whole thing was insane. Aside from the fact that there were more hands than cookie jar in the collaboration (always a sign of a bad business plan), it's clear that unless you're doing something that's already commodified, there's such a huge amount of shared experience and trust relationships between print services and customers that you can't commodify it. The customer needs to understand the limits of the process and the provider needs to understand the wants and needs of the customer. There are so many free variables that can never be quantified, but an iterative relationship between the customer and provider can produce good results as they learn to understand each other's wants and capabilities. The worst result is a loading dock full of unusable product.
Needless to say, this project went nowhere and my objections were (mostly) ignored. I like to think that maybe we spent less time pursuing a pointless venture than we might have otherwise, but who knows?
posted by sjswitzer at 3:06 PM on May 17, 2022 [2 favorites]
« Older Welcome to cluttercore | Newton's cradle, Gangnam style Newer »
This thread has been archived and is closed to new comments
In a previous life I dabbled in neuroscience (hence the username, which I've been using on various sites since ~1988). I imagined going back to my alma mater and teaching a course that studied vision in all domains: physics, biology, psychology, art, etc; a course that would draw students from a wide array of majors. I suspect this article would make great material for that course. And now I'm wanting to pull out my copy of David Marr.
posted by neuron at 8:53 AM on May 17, 2022 [2 favorites]