"it seems unnatural to look at a color and think of it in terms ... RGB"
March 31, 2013 6:57 PM Subscribe
Color Spaces:
Sex and Vision II: Color Appearance of Monochromatic Lights[PDF]:
It has been known for some time that colors can be described by three numbers. If I show you light of a certain color and ask you to match it by combining lights of three other colors and varying their intensities, you'll typically be able to find a combination that looks indistinguishable. But the wavelengths you combine might be very different from the wavelengths I showed you. Light of the wavelength corresponding to yellow and light of the right combination of red and green wavelengths will look the same, even though they are physically quite different.
Sex and Vision II: Color Appearance of Monochromatic Lights[PDF]:
: Because cerebral cortex has a very large number of testosterone receptors, we examined the possible sex differences in color appearance of monochromatic lights across the visible spectrum. There is a history of men and women perceiving color differently. However, all of these studies deal with higher cognitive functions which may be culture-biased. We study basic visual functions, such as color appearance, without reference to any objects. We present here a detailed analysis of sex differences in primary chromatic sensations.
Just curious, is this prompted by the recent thread in which atheists were likened to those with atypical color vision?
posted by jepler at 7:26 PM on March 31, 2013
posted by jepler at 7:26 PM on March 31, 2013
See also handprint.com's long and thorough article on Modern Color Theory (concepts), which is part of a longer discussion of color which is part of a set of writings on watercolor painting.
posted by sebastienbailard at 7:39 PM on March 31, 2013 [5 favorites]
posted by sebastienbailard at 7:39 PM on March 31, 2013 [5 favorites]
is this prompted by the recent thread in which atheists were likened to those with atypical color vision
This is prompted by my frantic efforts to clean out my Reader "starred" file before the end.
posted by the man of twists and turns at 7:43 PM on March 31, 2013 [10 favorites]
This is prompted by my frantic efforts to clean out my Reader "starred" file before the end.
posted by the man of twists and turns at 7:43 PM on March 31, 2013 [10 favorites]
Light of the wavelength corresponding to yellow and light of the right combination of red and green wavelengths will look the same, even though they are physically quite different. This structure is reflected in the retina.
This happens, of course, because yellow has a wavelength intermediate between red and green, so it stimulates both types of cone cells.
But what's interesting is that if you combine red and blue, you don't get a colour with a wavelength between the two! Any colour in the middle would also stimulate green cones. Instead you get purple, which doesn't correspond to any single wavelength at all. In this way purples are imaginary colours.
posted by vasi at 7:44 PM on March 31, 2013 [3 favorites]
This happens, of course, because yellow has a wavelength intermediate between red and green, so it stimulates both types of cone cells.
But what's interesting is that if you combine red and blue, you don't get a colour with a wavelength between the two! Any colour in the middle would also stimulate green cones. Instead you get purple, which doesn't correspond to any single wavelength at all. In this way purples are imaginary colours.
posted by vasi at 7:44 PM on March 31, 2013 [3 favorites]
Study data looked at 37 women and 21 men in one set of samples and in the other set 32 women and 15 men. The aggregate data set the total was 63 women and 33 men. Age ranges for the men was 16-61 while for the women is was 16-51. The study concludes
posted by humanfont at 7:56 PM on March 31, 2013 [1 favorite]
This we can speculate that there may be an extremely tiny, barely measurable difference in how men and women perceive colors. This is not as significant as the impact of sex linked colorblindness.
While the sex effects were small, the effect of sex was significant: F(1, 92) = 7.004, p = 0.010.
...
That is, regardless of the particular hue, males required, on average, a wavelength 2.2 nm longer than the wavelength needed to elicit the same sensation from females
posted by humanfont at 7:56 PM on March 31, 2013 [1 favorite]
Light of the wavelength corresponding to yellow and light of the right combination of red and green wavelengths will look the same, even though they are physically quite different. This structure is reflected in the retina.
I don't think this is entirely true. The same exact wavelength's of light can look very different depending on their surroundings:
Perceptual and Visual Illusions
posted by Golden Eternity at 7:58 PM on March 31, 2013 [1 favorite]
I don't think this is entirely true. The same exact wavelength's of light can look very different depending on their surroundings:
Perceptual and Visual Illusions
posted by Golden Eternity at 7:58 PM on March 31, 2013 [1 favorite]
In other color theory news - specifically enhancing photos - Dan Margulis, author of the L*A*B color book that has been my 'shop bible for years, has a new one out - The Picture Postcard Workflow. A propos nothing in particular in this thread; it's just been blowing my mind all weekend.
posted by notsnot at 8:05 PM on March 31, 2013 [5 favorites]
posted by notsnot at 8:05 PM on March 31, 2013 [5 favorites]
Folk who like the handprint's stuff on color will also get a kick out of James Gurney's notes and book along with Stapleton Kearn's notes.
posted by sebastienbailard at 8:17 PM on March 31, 2013 [2 favorites]
posted by sebastienbailard at 8:17 PM on March 31, 2013 [2 favorites]
But what's interesting is that if you combine red and blue, you don't get a colour with a wavelength between the two! Any colour in the middle would also stimulate green cones. Instead you get purple, which doesn't correspond to any single wavelength at all. In this way purples are imaginary colours.
Purple can't be produced with a single frequency*. An imaginary color cannot be produced by any particular spectrum (combination of frequencies). Since purple can be produced by a spectrum (a spectrum that includes high- and low-frequency light, but not much in the middle), it is not an imaginary color. It's actually a perfectly normal color that wishes everyone would get off its case.
*Btw, any signal that is not infinite in temporal extent includes a continuous interval of frequencies whose amplitude is non-zero (alternatively, an infinite number of frequencies). Every color (even those primary in your color space of choice) that you have ever seen in your life has been produced with light of more than one frequency.
posted by a snickering nuthatch at 8:23 PM on March 31, 2013 [4 favorites]
Purple can't be produced with a single frequency*. An imaginary color cannot be produced by any particular spectrum (combination of frequencies). Since purple can be produced by a spectrum (a spectrum that includes high- and low-frequency light, but not much in the middle), it is not an imaginary color. It's actually a perfectly normal color that wishes everyone would get off its case.
*Btw, any signal that is not infinite in temporal extent includes a continuous interval of frequencies whose amplitude is non-zero (alternatively, an infinite number of frequencies). Every color (even those primary in your color space of choice) that you have ever seen in your life has been produced with light of more than one frequency.
posted by a snickering nuthatch at 8:23 PM on March 31, 2013 [4 favorites]
My left eye sees different hues to my right. Not dramatically different, but enough for me to notice when I was about 13 and just getting into phenomenology (thank you, Dark Star).
I recommend this to anyone tempted to think colour exists outside the brain. Frequency, absolutely, and usefully so, but 'red'? As a pleasant side-effect, it induces intense relaxation about all those gorgeous Hubble images being so heavily tweaked.
posted by Devonian at 8:24 PM on March 31, 2013
I recommend this to anyone tempted to think colour exists outside the brain. Frequency, absolutely, and usefully so, but 'red'? As a pleasant side-effect, it induces intense relaxation about all those gorgeous Hubble images being so heavily tweaked.
posted by Devonian at 8:24 PM on March 31, 2013
This always makes me imagine what our world must look like to a species with different visual color spaces. Imagine if there were an alien race that had 5 different color receptors. A painting of an apple would look fine to us, but absolutely hideous to them. As would anything printed or on a video screen. They'd think we were idiots.
posted by gjc at 8:33 PM on March 31, 2013
posted by gjc at 8:33 PM on March 31, 2013
*Btw, any signal that is not infinite in temporal extent includes a continuous interval of frequencies whose amplitude is non-zero (alternatively, an infinite number of frequencies). Every color (even those primary in your color space of choice) that you have ever seen in your life has been produced with light of more than one frequency
Jpfed, If you are referring to the infinite set of frequencies that are produced at a discontinuity (as it might when the signal ends, a la the square wave) I don't think you can say that they meaningfully contribute to our perception of the frequency/wavelength present in that signal away from these discontinuities.
While the spectrum shows energies at all frequencies, they are not all activating our cone cells and contributing wholly to our perception of that wave as a certain color.
I would guess that when we talk colloquially about light at a certain frequency or wavelength, we do mean the frequency that has, by far, the most amount of energy, and in most of the duration that the light falls on our retina. That is an inherent approximation or loss of precision we have in both our sensory processes and language. (after all, the color detectors in the eyes implement only a crude approximation of a Fourier transform.)
I do think, however, that even though light of a single frequency can be generated and a certain subset of colors have the property of being single frequency, in reality we see multiple frequencies due to multi path reflections and also perturbations in the source (even in the case of lasers, although the bandwidth variations in normal light sources and hysteresis in visual systems are much larger). Again, the primary detected frequency would be dependent on the energy distribution at different frequencies and how much closer the neural detection process is to a winner-take-all mechanism.
I believe this property has an interesting parallel in the auditory domain with some modifications; the Shepard-Risset glissando illusion.
As far as real world computer monitors go, there is indeed a band of frequencies that the light would be emitted in, even if we are looking for spectral colors. How well they fall into MacAdam ellipses would determine how "pure" the color is going to look.
posted by ssri at 9:02 PM on March 31, 2013 [3 favorites]
Jpfed, If you are referring to the infinite set of frequencies that are produced at a discontinuity (as it might when the signal ends, a la the square wave) I don't think you can say that they meaningfully contribute to our perception of the frequency/wavelength present in that signal away from these discontinuities.
While the spectrum shows energies at all frequencies, they are not all activating our cone cells and contributing wholly to our perception of that wave as a certain color.
I would guess that when we talk colloquially about light at a certain frequency or wavelength, we do mean the frequency that has, by far, the most amount of energy, and in most of the duration that the light falls on our retina. That is an inherent approximation or loss of precision we have in both our sensory processes and language. (after all, the color detectors in the eyes implement only a crude approximation of a Fourier transform.)
I do think, however, that even though light of a single frequency can be generated and a certain subset of colors have the property of being single frequency, in reality we see multiple frequencies due to multi path reflections and also perturbations in the source (even in the case of lasers, although the bandwidth variations in normal light sources and hysteresis in visual systems are much larger). Again, the primary detected frequency would be dependent on the energy distribution at different frequencies and how much closer the neural detection process is to a winner-take-all mechanism.
I believe this property has an interesting parallel in the auditory domain with some modifications; the Shepard-Risset glissando illusion.
As far as real world computer monitors go, there is indeed a band of frequencies that the light would be emitted in, even if we are looking for spectral colors. How well they fall into MacAdam ellipses would determine how "pure" the color is going to look.
posted by ssri at 9:02 PM on March 31, 2013 [3 favorites]
the color detectors in the eyes implement only a crude approximation of a Fourier transform
It's not even a crude Fourier transform; it's a filter-bank (with three filters for most of us; four for bees, goldfish, or the rare female human tetrachromats; more for certain shrimp, etc).
On the one hand, what with the quantized nature of both light and the process by which our retinas detect light, you can say that even a brief impulse of light can have a finite number of frequencies. On the other hand, mathematically, the non-narrowband nature of finite signals and the Heisenberg uncertainty of a photon's energy/wavelength are practically the same phenomenon. So, in conclusion, color is a land of contrasts. Thank you.
posted by hattifattener at 9:34 PM on March 31, 2013 [3 favorites]
It's not even a crude Fourier transform; it's a filter-bank (with three filters for most of us; four for bees, goldfish, or the rare female human tetrachromats; more for certain shrimp, etc).
On the one hand, what with the quantized nature of both light and the process by which our retinas detect light, you can say that even a brief impulse of light can have a finite number of frequencies. On the other hand, mathematically, the non-narrowband nature of finite signals and the Heisenberg uncertainty of a photon's energy/wavelength are practically the same phenomenon. So, in conclusion, color is a land of contrasts. Thank you.
posted by hattifattener at 9:34 PM on March 31, 2013 [3 favorites]
It's not even a crude Fourier transform; it's a filter-bank (with three filters for most of us; four for bees, goldfish, or the rare female human tetrachromats; more for certain shrimp, etc).
Yes, this is definitely more correct. In fact there is some research which looks into modeling extraction of color information using the long short and medium cones as kernels convolved with the light signal. This is further complicated by the fact that color cones in the eyes are not ordered nicely like, say a Bayer Filter.
So, in conclusion, color is a land of contrasts.
To slightly expand on what I wrote, the neural detection process is complicated at multiple levels, with color contrast enhancement happening at the thalamic level(Parvocellular), then at the primary visual cortex(V1) in humans, and then of course at higher levels (V2, V4) where overall chromaticity and luminance are taken into account affecting color constancy.
posted by ssri at 10:27 PM on March 31, 2013 [3 favorites]
Yes, this is definitely more correct. In fact there is some research which looks into modeling extraction of color information using the long short and medium cones as kernels convolved with the light signal. This is further complicated by the fact that color cones in the eyes are not ordered nicely like, say a Bayer Filter.
So, in conclusion, color is a land of contrasts.
To slightly expand on what I wrote, the neural detection process is complicated at multiple levels, with color contrast enhancement happening at the thalamic level(Parvocellular), then at the primary visual cortex(V1) in humans, and then of course at higher levels (V2, V4) where overall chromaticity and luminance are taken into account affecting color constancy.
posted by ssri at 10:27 PM on March 31, 2013 [3 favorites]
My left eye sees different hues to my right.
I've noticed this about my eyes, too. My right eye sees things a bit warmer than my left.
posted by Thorzdad at 4:41 AM on April 1, 2013
I've noticed this about my eyes, too. My right eye sees things a bit warmer than my left.
posted by Thorzdad at 4:41 AM on April 1, 2013
Jpfed, If you are referring to the infinite set of frequencies that are produced at a discontinuity (as it might when the signal ends, a la the square wave) I don't think you can say that they meaningfully contribute to our perception of the frequency/wavelength present in that signal away from these discontinuities.
Indeed; I'm just poking a stick at the nonsense of privileging the "single-wavelength-of-light" perspective, because such signals never actually exist.
While the spectrum shows energies at all frequencies, they are not all activating our cone cells and contributing wholly to our perception of that wave as a certain color.
Eh, for it to matter to color perception (and I'm not positing that it does) it just has to have nonzero energy in bands that more than one cone cell is sensitive to; the fact that there is EM that we're not sensitive to at all is a bit of a colorless herring.
I do think, however, that even though light of a single frequency can be generated and a certain subset of colors have the property of being single frequency, in reality we see multiple frequencies due to multi path reflections and also perturbations in the source (even in the case of lasers, although the bandwidth variations in normal light sources and hysteresis in visual systems are much larger). Again, the primary detected frequency would be dependent on the energy distribution at different frequencies
I'm not quite sure I'm grasping what you're talking about here. When you say "primary detected frequency", are you referring to the frequency that would in some ideal case produce the same color percept as the actual stimulus?
and how much closer the neural detection process is to a winner-take-all mechanism.
This is the really weird part. I'm aware of lateral inhibition within a color channel. But for the bandwidth of the signal to affect the "primary detected frequency" (if you mean what I think you mean) via a winner-take-all mechanism, there would have to be inhibition between color channels. If that is the case, shit be cray! Such a mechanism would be pushing all of our color percepts towards the primary components of the color space (whichever one is being used where the inhibition occurs).
posted by a snickering nuthatch at 6:18 AM on April 1, 2013
Indeed; I'm just poking a stick at the nonsense of privileging the "single-wavelength-of-light" perspective, because such signals never actually exist.
While the spectrum shows energies at all frequencies, they are not all activating our cone cells and contributing wholly to our perception of that wave as a certain color.
Eh, for it to matter to color perception (and I'm not positing that it does) it just has to have nonzero energy in bands that more than one cone cell is sensitive to; the fact that there is EM that we're not sensitive to at all is a bit of a colorless herring.
I do think, however, that even though light of a single frequency can be generated and a certain subset of colors have the property of being single frequency, in reality we see multiple frequencies due to multi path reflections and also perturbations in the source (even in the case of lasers, although the bandwidth variations in normal light sources and hysteresis in visual systems are much larger). Again, the primary detected frequency would be dependent on the energy distribution at different frequencies
I'm not quite sure I'm grasping what you're talking about here. When you say "primary detected frequency", are you referring to the frequency that would in some ideal case produce the same color percept as the actual stimulus?
and how much closer the neural detection process is to a winner-take-all mechanism.
This is the really weird part. I'm aware of lateral inhibition within a color channel. But for the bandwidth of the signal to affect the "primary detected frequency" (if you mean what I think you mean) via a winner-take-all mechanism, there would have to be inhibition between color channels. If that is the case, shit be cray! Such a mechanism would be pushing all of our color percepts towards the primary components of the color space (whichever one is being used where the inhibition occurs).
posted by a snickering nuthatch at 6:18 AM on April 1, 2013
My left eye sees different hues to my right.
I've noticed this about my eyes, too. My right eye sees things a bit warmer than my left.
Also a member of Team Saturation Setting Slightly Out of Sync
posted by ennui.bz at 7:06 AM on April 1, 2013
I've noticed this about my eyes, too. My right eye sees things a bit warmer than my left.
Also a member of Team Saturation Setting Slightly Out of Sync
posted by ennui.bz at 7:06 AM on April 1, 2013
the man of twists and turns: "is this prompted by the recent thread in which atheists were likened to those with atypical color vision
This is prompted by my frantic efforts to clean out my Reader "starred" file before the end."
WHY MUST YOU REMIND ME? (I am so trying to avoid thinking about it).
posted by symbioid at 11:08 AM on April 1, 2013
This is prompted by my frantic efforts to clean out my Reader "starred" file before the end."
WHY MUST YOU REMIND ME? (I am so trying to avoid thinking about it).
posted by symbioid at 11:08 AM on April 1, 2013
My left eye sees different hues to my right.
I mentioned something like this to a noted vision researcher once and was told that it was not possible. Well, OK then.
posted by mimo at 12:18 PM on April 1, 2013
I mentioned something like this to a noted vision researcher once and was told that it was not possible. Well, OK then.
posted by mimo at 12:18 PM on April 1, 2013
My eyes sometimes see things differently, but it's because I have a bad habit of using one eye for reading and closing the other (not consistently). They adapt to dark independently, so when I have both open again the one that's been closed for a while has the iris open wider until it adjusts, and that affects the look of things.
Your eyes may vary.
posted by Segundus at 2:50 PM on April 1, 2013
Your eyes may vary.
posted by Segundus at 2:50 PM on April 1, 2013
yes, the worst time I ever had with a vision mismatch between my eyes was after using an electronic-viewfinder camera for a few hours. My viewfinder eye was seeing a processed (and, notably, white-balanced) image, while my other eye (when it was open) was seeing normally. However long it was I kept at this, the difference in color perception between my two eye was quite startling. Luckily, after a very short time (hours, probably?) the sensation went away.
posted by jepler at 3:09 PM on April 1, 2013
posted by jepler at 3:09 PM on April 1, 2013
I'm not quite sure I'm grasping what you're talking about here. When you say "primary detected frequency", are you referring to the frequency that would in some ideal case produce the same color percept as the actual stimulus?
Yes, that was sloppy. I was referring to perception of a spectral color, not at the level of the photoreceptor, but at the percept level which might be at the primary/secondary visual cortex (at the psychophysical level of a MacAdam ellipse). Again, I was trying to refer to the entire neural detection process (which would be implemented in a distributed fashion starting from the photoreceptor to V1/V2) the end result of which would be closer to a winner-take-all process (resulting in choosing the same color, I am guessing, within the just-noticeable-difference boundary). That is my understanding.
This is the really weird part. I'm aware of lateral inhibition within a color channel. But for the bandwidth of the signal to affect the "primary detected frequency" (if you mean what I think you mean) via a winner-take-all mechanism, there would have to be inhibition between color channels. If that is the case, shit be cray! Such a mechanism would be pushing all of our color percepts towards the primary components of the color space (whichever one is being used where the inhibition occurs).
Yeah, no, I don't think the lateral inhibition process at the retinal photoreceptor level results in color resolution at that level itself.
At the photoreceptor level, the long and medium cones actually respond (of course, not equally) to light from the entirety of the visible spectrum, while the short cones respond to a smaller spectrum. The long and medium cones' peaks are separated by ~30nm. When the signal from the L or M cones increase, at the receptor level, they don't know whether it is because more photons from an insensitive part of the spectrum impinged or if fewer photons from a more sensitive wavelength hit the photoreceptor. The cones are "color-blind" and the only way to know the chromatic information is to include the surrounding receptors' signals.
There are at least two stages of linear transformation of the S, M and L signals at the LGN and the V1 and the color basis are no longer red, green and blue. It is closer to a color space with a luminance component, and an approximate red-green and an approximate blue-yellow component. Further non-linear transformations can take place at V1 that appear to hint at a finer honed color discrimination/detection process with more specificity than is allowed by the RG/BY system.
posted by ssri at 8:36 PM on April 1, 2013 [2 favorites]
Yes, that was sloppy. I was referring to perception of a spectral color, not at the level of the photoreceptor, but at the percept level which might be at the primary/secondary visual cortex (at the psychophysical level of a MacAdam ellipse). Again, I was trying to refer to the entire neural detection process (which would be implemented in a distributed fashion starting from the photoreceptor to V1/V2) the end result of which would be closer to a winner-take-all process (resulting in choosing the same color, I am guessing, within the just-noticeable-difference boundary). That is my understanding.
This is the really weird part. I'm aware of lateral inhibition within a color channel. But for the bandwidth of the signal to affect the "primary detected frequency" (if you mean what I think you mean) via a winner-take-all mechanism, there would have to be inhibition between color channels. If that is the case, shit be cray! Such a mechanism would be pushing all of our color percepts towards the primary components of the color space (whichever one is being used where the inhibition occurs).
Yeah, no, I don't think the lateral inhibition process at the retinal photoreceptor level results in color resolution at that level itself.
At the photoreceptor level, the long and medium cones actually respond (of course, not equally) to light from the entirety of the visible spectrum, while the short cones respond to a smaller spectrum. The long and medium cones' peaks are separated by ~30nm. When the signal from the L or M cones increase, at the receptor level, they don't know whether it is because more photons from an insensitive part of the spectrum impinged or if fewer photons from a more sensitive wavelength hit the photoreceptor. The cones are "color-blind" and the only way to know the chromatic information is to include the surrounding receptors' signals.
There are at least two stages of linear transformation of the S, M and L signals at the LGN and the V1 and the color basis are no longer red, green and blue. It is closer to a color space with a luminance component, and an approximate red-green and an approximate blue-yellow component. Further non-linear transformations can take place at V1 that appear to hint at a finer honed color discrimination/detection process with more specificity than is allowed by the RG/BY system.
posted by ssri at 8:36 PM on April 1, 2013 [2 favorites]
My left eye sees different hues to my right.
I mentioned something like this to a noted vision researcher once and was told that it was not possible. Well, OK then.
Different hues is not the same as different vision. My eyes aren't perfectly white balanced with each other, but they both have the same relative sensitivity. I think what the researcher was saying is that you aren't going to get different colors in different eyes. IE, brightness, saturation and contrast can be different, but hue can't. That's my interpretation anyway.
posted by gjc at 3:52 PM on April 7, 2013
I mentioned something like this to a noted vision researcher once and was told that it was not possible. Well, OK then.
Different hues is not the same as different vision. My eyes aren't perfectly white balanced with each other, but they both have the same relative sensitivity. I think what the researcher was saying is that you aren't going to get different colors in different eyes. IE, brightness, saturation and contrast can be different, but hue can't. That's my interpretation anyway.
posted by gjc at 3:52 PM on April 7, 2013
« Older First, the tried to make moons | up close & personal Newer »
This thread has been archived and is closed to new comments
posted by knapah at 7:22 PM on March 31, 2013 [4 favorites]