4 ms·

Did you read to the end? He givers a physiology-based argument for a color wheel with four primary colors.

blueflow 2y ago
This is like watching some preschoolers argue whether the rainbow has 4 or 5 or 7 colors. The arguments do not make sense at all because they are actually some form of opinion.
- Daub 2y ago
  Well... I would say that you are half right. The whole notion of primary/secondary is very dependent on application, so your primary set would depend on your intended use. Psychology recognizes six primaries: red, green, blue, yellow, white and black. The author presents a four-color primary set. It clearly has its roots in Helmholtz's color physiology, but I confess I don't quite understand how it would be used. For me, I would miss the important blue/orange pairing.
  - zarzavat 2y ago
    The problem is that yellow is much lighter than blue. If you look at a perceptual pseudo-cylindrical color space such as CIECAM02 or (OK)LCh, yellow and blue are not on the same wheel because they exist at different lightness levels. So the author is correct that there are four opponent colors, but not correct that they can all be arranged on a wheel so easily. Color spaces are three dimensional geometries, attempts to reduce them to two dimensions are always wrong just like every 2D map projection creates distortions.
    - Daub 2y ago
      This is, at core, a demonstration of the complex dimensionality of color. Yellow is not lighter than blue, though I can understand why you say this. In the hue, saturation and lightness (hsl) understanding of color, hue terms such as Blue/yellow and lightness terms such as light/dark are separate entities. It is perfectly possible to have a blue and yellow of the same lightness value, but the yellow would be more dull. More dull = less pure = less saturated. The peak saturation to lightness of different hues is not consistent. Yellow is most yellow (ie most saturate/pure) when it is quite light. Conversly the peak saturationn of green or blue is in the dark range. As you rightfully state, One key issue with the color wheel is that it is two dimensional. There does not exist a remotely complete color model that does not recognize at least three dimensions. I am not dissing the utility of the color wheel, but it has application only to the ordering of hues.
  - chrisweekly 2y ago
    green as a "primary", when it is a combination of blue + yellow, seems strange to me
- red_admiral 2y ago
  Clearly 7, because 7 is the most powerful magical number. That's literally the level of Newton's justification, who invented the concept of a 7-colored rainbow. While we're here, one of my favorite unsong quotes, at roughly the same level of woo: > Then there was light. > > Beautiful, multicolored light, ten colors, the seven colors of the earthly rainbow and the three extra colors you only get in Heaven. Ten colors corresponding to the ten sephirot and the ten fingers and the Ten Commandments and the ten digits of the number system and the ten pip cards of the Tarot and all the other tens in all the correspondences of the world.
  - Daub 2y ago
    Where does this come from? The Torah I first assumed, but then I say mention of thebtarot.
    - red_admiral 2y ago
      "Torah fan fiction" by Scott Alexander at unsongbook.com . The sephirot part is from kabbalah.
- quietbritishjim 2y ago
  No, they are not just a matter of opinion (although there is some small subjective element to it). If our eyes somehow detected the exact wavelength of light directly then, yes, we would perceive a continuum of colours and it would be totally arbitrary how you divide it up. But we don't. As the article says, we have a discrete set of cone types that divide our perception of colours up into clusters (albeit overlapping and fuzzy). For that reason, for most people with normal vision, a rainbow appears to have 6 main colours.
  - itishappy 2y ago
    We have a discrete set of cones but they have continuous outputs so the space is indeed a continuum. The 6 colors of the rainbow are arbitrary and cultural (I've always learned there are 7). There's no physiological reason we say orange is a color but dark orange isn't. We perceive more than 7 colors.
    - adastra22 2y ago
      The rainbow as perceived by most humans has six distinct color peaks / bands that are noticeably distinct from each other. Physically it is an even continuum of wavelengths, but look at it with your own eyes and you will see clear banding. This is due to way retina information is filtered to produce color signals before being fed into the brain. Why seven colors? Because Isaac Newton felt it must be 7 due to it being a more mystical number, so he shoved indigo in there.
      - itishappy 2y ago
        > Newton, who admitted his eyes were not very critical in distinguishing colours, originally (1672) divided the spectrum into five main colours: red, yellow, green, blue and violet. Later he included orange and indigo, giving seven main colours by analogy to the number of notes in a musical scale. > The question of whether everyone sees seven colours in a rainbow is related to the idea of linguistic relativity. Suggestions have been made that there is universality in the way that a rainbow is perceived. However, more recent research suggests that the number of distinct colours observed and what these are called depend on the language that one uses, with people whose language has fewer colour words seeing fewer discrete colour bands. https://en.wikipedia.org/wiki/Rainbow https://en.wikipedia.org/wiki/Rainbow
card_zero 2y ago
That makes sense, since it separates out green from yellow, which is where we have two cones close together in the spectrum (nominally green and red cones) and the most response and discrimination (humans are good at seeing greens). This part is surely an error, though: > So there’s no such thing as “red with a little green”—there’s just a less intense red. This would mean that a monitor can't display orange (red with a little green).
- gilleain 2y ago
  Heh, meanwhile over here in paint world, I've been mixing red and green to make grey (specifically Royal Talens (362 Deep Rose) + (602 Deep Green)). I think of orange as yellow + small amount of red, but hey.
  - peterashford 2y ago
    Yeah, additive (light, monitors) and subtractive (paint) color systems work differently
- adastra22 2y ago
  I think you’re missing the point on orange. When I see aquamarine, I see blue and green at the same time. It feels like blue-green. Orange on the other hand looks and feels like a unique color, not red + green (although I know why that is). TFA is using physiology to explain why these color mixes are perceived differently.
  - card_zero 2y ago
    Yes, but at that point in the article, it's clearly talking about the amount of red, green, or blue signal being picked up by cone cells, and thus the wavelength or mix of wavelengths, which is additive mixing. It's claiming that red plus some green (in terms of stimulation of cone cells) equals dimmed red, which is factually incorrect. This part about "three filters" is overstated, I think, if based in fact at all. Presumably it's based on something the retina is known to do, but I hadn't heard before of red and green being preprocessed separately from blue, and given that we can see orange when presented with red and some green pixels, it's not exactly as stated.

This is like watching some preschoolers argue whether the rainbow has 4 or 5 or 7 colors. The arguments do not make sense at all because they are actually some form of opinion.

Daub 2y ago
Well... I would say that you are half right. The whole notion of primary/secondary is very dependent on application, so your primary set would depend on your intended use. Psychology recognizes six primaries: red, green, blue, yellow, white and black. The author presents a four-color primary set. It clearly has its roots in Helmholtz's color physiology, but I confess I don't quite understand how it would be used. For me, I would miss the important blue/orange pairing.
- zarzavat 2y ago
  The problem is that yellow is much lighter than blue. If you look at a perceptual pseudo-cylindrical color space such as CIECAM02 or (OK)LCh, yellow and blue are not on the same wheel because they exist at different lightness levels. So the author is correct that there are four opponent colors, but not correct that they can all be arranged on a wheel so easily. Color spaces are three dimensional geometries, attempts to reduce them to two dimensions are always wrong just like every 2D map projection creates distortions.
  - Daub 2y ago
    This is, at core, a demonstration of the complex dimensionality of color. Yellow is not lighter than blue, though I can understand why you say this. In the hue, saturation and lightness (hsl) understanding of color, hue terms such as Blue/yellow and lightness terms such as light/dark are separate entities. It is perfectly possible to have a blue and yellow of the same lightness value, but the yellow would be more dull. More dull = less pure = less saturated. The peak saturation to lightness of different hues is not consistent. Yellow is most yellow (ie most saturate/pure) when it is quite light. Conversly the peak saturationn of green or blue is in the dark range. As you rightfully state, One key issue with the color wheel is that it is two dimensional. There does not exist a remotely complete color model that does not recognize at least three dimensions. I am not dissing the utility of the color wheel, but it has application only to the ordering of hues.
- chrisweekly 2y ago
  green as a "primary", when it is a combination of blue + yellow, seems strange to me
red_admiral 2y ago
Clearly 7, because 7 is the most powerful magical number. That's literally the level of Newton's justification, who invented the concept of a 7-colored rainbow. While we're here, one of my favorite unsong quotes, at roughly the same level of woo: > Then there was light. > > Beautiful, multicolored light, ten colors, the seven colors of the earthly rainbow and the three extra colors you only get in Heaven. Ten colors corresponding to the ten sephirot and the ten fingers and the Ten Commandments and the ten digits of the number system and the ten pip cards of the Tarot and all the other tens in all the correspondences of the world.
- Daub 2y ago
  Where does this come from? The Torah I first assumed, but then I say mention of thebtarot.
  - red_admiral 2y ago
    "Torah fan fiction" by Scott Alexander at unsongbook.com . The sephirot part is from kabbalah.
quietbritishjim 2y ago
No, they are not just a matter of opinion (although there is some small subjective element to it). If our eyes somehow detected the exact wavelength of light directly then, yes, we would perceive a continuum of colours and it would be totally arbitrary how you divide it up. But we don't. As the article says, we have a discrete set of cone types that divide our perception of colours up into clusters (albeit overlapping and fuzzy). For that reason, for most people with normal vision, a rainbow appears to have 6 main colours.
- itishappy 2y ago
  We have a discrete set of cones but they have continuous outputs so the space is indeed a continuum. The 6 colors of the rainbow are arbitrary and cultural (I've always learned there are 7). There's no physiological reason we say orange is a color but dark orange isn't. We perceive more than 7 colors.
  - adastra22 2y ago
    The rainbow as perceived by most humans has six distinct color peaks / bands that are noticeably distinct from each other. Physically it is an even continuum of wavelengths, but look at it with your own eyes and you will see clear banding. This is due to way retina information is filtered to produce color signals before being fed into the brain. Why seven colors? Because Isaac Newton felt it must be 7 due to it being a more mystical number, so he shoved indigo in there.
    - itishappy 2y ago
      > Newton, who admitted his eyes were not very critical in distinguishing colours, originally (1672) divided the spectrum into five main colours: red, yellow, green, blue and violet. Later he included orange and indigo, giving seven main colours by analogy to the number of notes in a musical scale. > The question of whether everyone sees seven colours in a rainbow is related to the idea of linguistic relativity. Suggestions have been made that there is universality in the way that a rainbow is perceived. However, more recent research suggests that the number of distinct colours observed and what these are called depend on the language that one uses, with people whose language has fewer colour words seeing fewer discrete colour bands. https://en.wikipedia.org/wiki/Rainbow https://en.wikipedia.org/wiki/Rainbow

card_zero 2y ago

That makes sense, since it separates out green from yellow, which is where we have two cones close together in the spectrum (nominally green and red cones) and the most response and discrimination (humans are good at seeing greens). This part is surely an error, though: > So there’s no such thing as “red with a little green”—there’s just a less intense red. This would mean that a monitor can't display orange (red with a little green).

gilleain 2y ago
Heh, meanwhile over here in paint world, I've been mixing red and green to make grey (specifically Royal Talens (362 Deep Rose) + (602 Deep Green)). I think of orange as yellow + small amount of red, but hey.
- peterashford 2y ago
  Yeah, additive (light, monitors) and subtractive (paint) color systems work differently
adastra22 2y ago
I think you’re missing the point on orange. When I see aquamarine, I see blue and green at the same time. It feels like blue-green. Orange on the other hand looks and feels like a unique color, not red + green (although I know why that is). TFA is using physiology to explain why these color mixes are perceived differently.
- card_zero 2y ago
  Yes, but at that point in the article, it's clearly talking about the amount of red, green, or blue signal being picked up by cone cells, and thus the wavelength or mix of wavelengths, which is additive mixing. It's claiming that red plus some green (in terms of stimulation of cone cells) equals dimmed red, which is factually incorrect. This part about "three filters" is overstated, I think, if based in fact at all. Presumably it's based on something the retina is known to do, but I hadn't heard before of red and green being preprocessed separately from blue, and given that we can see orange when presented with red and some green pixels, it's not exactly as stated.