Talk:Natural Color System

From Wikipedia, the free encyclopedia
Jump to: navigation, search
WikiProject Color (Rated Start-class, High-importance)
WikiProject icon This article is supported by WikiProject Color, a project that provides a central approach to color-related subjects on Wikipedia. Help us improve articles to good and 1.0 standards; visit the wikiproject page for more details.
Start-Class article Start  This article has been rated as Start-Class on the project's quality scale.
 High  This article has been rated as High-importance on the project's importance scale.

Early discussion[edit]

The changes on 5 July 2005 were made by me.--MWAK 5 July 2005 07:32 (UTC)

In view of the recent POV-edit, here some facts:

  1. The "Natural Color System" is not scientific: it isn't founded on serious empirical research — indeed the movement has never tried to falsify its tenets.
  2. It's in blatant contradiction to the present results of serious empirical research of color perception.
  3. The actual system can only work because it surreptitiously perverts its own underlying philosophy.
  4. The — overly expensive, may I add — instruments that go with it are themselves calibrated using the scientific system, for the simple reason the NCS is useless.
  5. The movement shamelessly exploits human ignorance: it is basically an attempt to profit from the confusion caused by the fact that outdated color theories are slowly replaced by the correct one.

--MWAK 10:28, 26 November 2005 (UTC)

Firstly, your alleged "facts" are nothing more than a biassed bunch of ignorant babble. The NCS is based on color opponency, which was presented to the scientific community long ago, nowadays there being a pretty good scientific basis and a whole lot of scientific research supporting it (read something about bipolar and ganglion cells, will you?). Just because some editors here apparently know nothing buabout color opponency doesn't make the NCS "a sham", it only makes them (among them you) look utterly ignorant on the matter they dare talk about.
Secondly, that the instruments are calibrated using another color system that is based on the physical properties of light is not surprising; it should be obvious. NCS is all about the perceptual color space (the one that describes the range and organization of colors as experienced at the upper brain level, not at the lower retinal cone level); it does not deal with spectral light compositions nor with which retinal cones a certain light will stimulate.
Thirdly, that human trichromat vision is indeed perceptually based on the six basic percepts of white, black, red, yellow, green and blue (the elementary colors: two achromat poles and two pairs of opponent unary hues), should be obvious to anyone who cares to think for a moment about which colors do actually look "pure" or "unmixed", so that they can by no means be described as "looking like" a mix of other pure colors. It's no coincidence that these six elementary colors are those most frequently chosen to paint educational toys or for designs that try to appeal from their simplicity (such as in the Olympic flag and the Microsoft Windows logo). Orange, a binary hue, not only results from mixing red-looking and yellow-looking pigments, it does look indeed like "a mix of red and yellow". On the other hand, turquoise, another binary hue, looks blue and green at the same time, certainly not blue and yellow, even though it can be obtained by mixing blue-looking pigment with a bit of yellow-looking pigment. Finally, none of the four unary hues (red, yellow, green and blue) can be described as looking like a mix of other colors (you cannot say, for example, that red actually "looks like a cyan and magenta mixture", or that green actually "looks like a yellow and cyan mixture"), even though you can actually get them by mixing pigments or lights of different color appearance.
Fourthly, a quite different matter is that if you mix a blue-looking pigment with a yellow-looking pigment, you get a green-looking pigment; but here we are not talking about the perceptual color space at all, we are talking about what happens when mixing pigments. Or that if you mix a red-looking light with a green-looking light you get a yellow-looking light; which does not look like a red-green light at all, does it? (read about Crane and Piantanida's experiment if you want to know how to get to see a color that actually looks like a "red-green" mixture).
Fifthly, if you actually are unable to realize that saturated magenta is nothing but a mixed color that looks primarily reddish and secondarily bluish, then you do have a serious problem with your color perception.
Finally, I'm reverting all the ignorant negative bias from this article right now.
Uaxuctum 15:41, 31 January 2006 (UTC)
O, dear, you are a true believer, aren't you? :o) Let me try to deprogram you — of course I can't guarantee any results ;o).
  1. Yes, it's indeed very typical for the adherents of the tetrachromist movement to refer to all kinds of neurological research, just as the admirer of any pseudoscience is often able to cite long lists of articles he thinks are able to support his notions. Why is it that in this they are just as misguided as any creationist or astrologist? The reason is that they are making the very category mistake you seem to think I am making. As the NCS is supposed to be a description of the perceptual level, any results of neurological research have to be translated to that level. How that should be done however, is a complete and utter mystery to present science and philosophy alike. At best neurology can give some tentative suggestions. Tetrachromists make a careful selection from the available data, ignoring the most basic fact among them: that as the neural system simply processes signals from a trichromatic perceptual biological system, the eye, the most parsimonious hypothesis will be that the phenomenal result has a trichromatic structure too. And indeed there are no neurological research data contradicting this hypothesis. It seems to be true that the combined stimuli of two types, "red" and "green" are in opponency against "blue"; but to conclude therefore that the perceptual result of the combination, yellow, is to be placed on the same level with red, green and blue into a tetrachromic system is a simple non-sequitur. It's even worse to conclude that this would be the perceptual system. After all, why stop there? "Cyan" and "magenta" are analysed on a higher level again, which would make for a pentachromic and then hexachromic system. Should we then assume for the latter, two levels of colours: primary and secondary ones, biology, neurology and phenomenology are in blissful trichromic accordance. :o)
  2. Now you might retort: "But there is no accordance! Science may research what it will, but I know from my own subjective experience that the human perceptual system is tetrachromic. If you disagree, you are either colourblind or incompetent to judge!" Some scientists would answer to this that subjective experience is not a reliable source of information or even doesn't exist at all. I, however, will of course contend that subjective colour does exist but is experienced in a trichromic way. But then how can I explain that you think you perceive it differently? We can begin by looking at the historical development of tetrachromism. Early tetrachromists also had to compete with a trichromic colour system. But that was a different system from the present one. It was an inferior version based on the mistaken assumption that the primary additive colours were red, yellow and blue. It is noticable that present tetrachromism simply repeats most of these mistakes in that it holds the secondary colours red and blue to be "unary" colours: it is still nothing but the incorrect system with a mistake added: to see green as a basic colour also. But strangely enough this added error made the early tetrachromists in one way superior. The very reason they added green was that they correctly perceived that green is not a mixture of yellow and blue. It is in reality a mixture of yellow and cyan and they were aware that the RYB system failed in this respect. Had their intuitive reflection been even better — or had trichromatism not evolved by itself, blocking this road — they would have created a pentachromic system adding purple, which is after all not a mixture of red and blue (the situation for orange would obviously have been different). As even today the most common social conditioning regarding the ordering of colour concepts is in the form of the RYB system, it wouldn't be surprising if you experienced it as "natural"; the extra colour green would give it a sensation of more perfect completion and satisfy your need for intuitive understanding.
  3. But where then did the RYB system itself come from? Part of the answer is very convenient for my position: pure contingency. It so happened that the most saturated pigments available were vermillion and ultramarine, or red and blue. Indeed good magenta is so recent an invention that I doubt most people have a clear notion of it when they use the word. However this explanation is a bit too easy. Linguistic research, starting with Berlin, has shown that apparently there is another level apart from the biological, neurological and perceptual: the psychological. Human languages seems first to create a distinct word for red, then for green or yellow as second or third words, then for blue and only then for other colours. There seem, so it is claimed, to be four Urfarben. I have to admit this is at first blush an indication that the perceptual level might likewise be tetrachromic. However the indication is only slight. There is no real reason to speak of four basic psychological colours as pink ("magenta"), cyan, purple and brown have their own distinctive psychological value too and it is problematic to conceive these values as somehow derived from more basic ones. Also there is no good reason to see languages with three basic colour names as still incomplete, with four names as just right and with more names as exuberant. The number four is not one simply suggested by the data: it is clearly superimposed from a cultural bias, being our own RYB system. Even more troubling is the fact that the colourfoci do not simply coincide with either the trichromic or tetrachromic system: e.g. the focus for yellow is more orange than primary or unary yellow. It seems that a primate was equipped with a system to attend the organism to ripe fruit precisely by employing a psychological system that can be put in contrast with the perceptual one.
  4. Probably none of this will have convinced you. When I say: "Why, I clearly, evidently and obviously do experience green as a mixture of yellow and cyan and red as a mixture of yellow and magenta, which latter colour looks only to be in between red and blue because it is a primary component of both", you will in all likelihood merely be amazed by the power of human self-deception. My intuition cannot overcome yours. But perhaps you can be convinced by your own intuition. Indulge me by performing the following thought experiment. Imagine two cans of yellow paint. And a magenta and cyan one. Now if what you say is true, if red and green have no phenomenal yellow, cyan or magenta component whatsoever, it cannot be predicted by any of the conceptual qualities of the colours magenta and cyan what the result of their mixture with yellow would be. It would be a purely physical proces in which only previous experiment or understanding of the underlying physics could inform us of the outcome. Now add the magenta to the yellow paint in one can and imagine that the mixture turned green. And add the cyan paint to the yellow in the other and imagine that it turned red. Would you truly say that this result seems intuitively as natural to you as the usual opposite result? If not, could you unproblematically explain this as an effect of empirical conditioning? How would you know the difference between "true" intuition and conditioned intuition?
Apart from these problems with tetrachromism in general, it should be said that the NCS in particular is a very, very silly ofshoot of it. How is "blue" determined? Where has saturated magenta gone? What serious practical use could be conceivable for such a system (apart from enriching the sellers)?--MWAK 18:08, 9 April 2006 (UTC)
I give up trying to discuss with you. The color opponency model of color vision is a widely accepted theory in the scientific community, not some crackpot idea, and there is already quite a lot of research and knowledge about its underlying physiological mechanisms. And it's not at odds with the RGB model, it's complementary to it; they each describe a different part of the complex mechanism of human colour vision, neither describes it entirely. If you choose to doubt it, that's O.K., everyone's entitled to their own opinions; but it's your personal problem, so please keep it away from Wikipedia. This is not the place to publicize your personal doubts or misunderstandings about a widely accepted scientific theory, nor to try to unwarrantedly discredit or cast doubt on the validity of a standard color model used officially in several countries as well as by a very relevant international organization in the field of color trend forecasting. Uaxuctum 15:43, 15 May 2006 (UTC)
You stop discussing because you don't have any valid counterarguments? :o)The point is that the truth of colour opponency has little to do with the truth of the NCS "model". Indeed colour opponency is fully compatible with the RGB system, but the NCS isn't. However your text describes the NCS as if it were true; my edits merely instated the minimum of necessary NPOV by describing it as a hypothesis — as any hypothesis should. Besides your example proving that the NCS is more intuitive (comparing the mixing of red and green paint with the mixing of red and green light) is an apparent mistake on your side: it is logically flawed and wouldn't be used by the typical NCS-adherent. And no, I'm not pushing my personal POV here but simply reflecting the consensus: very few colour scientists see any value in the NCS. And surely any article should also mention that?--MWAK 06:10, 16 May 2006 (UTC)
No, simply your comments have already made it very clear you don't even understand color opponency, nor anything about how human visions works. Comments like that "as the neural system simply processes signals from a trichromatic perceptual biological system, the eye, the most parsimonious hypothesis will be that the phenomenal result has a trichromatic structure too" make you look like a complete ignorant of the matter. Study something about human vision, will you? There is much more to it than the retinal cones. To start with, the tri-stimulus produced by the cones doesn't reach any further than the retina. The bipolar and ganglion cells that receive their input quickly transform it into a very different kind of signal, one with an achromat component and two hue components, by comparing the relative intensities of those stimulus (the red-vs-green hue component for the relative intensities of the "red" versus "green" cones, the blue-vs-yellow hue component for the relative intensities of the "blue" cone versus the combined signal of the other two, and the achromat black-to-white component by measuring the overall intensity). The initial RGB-type "raw" signal as it is produced by the cones never reaches the brain, which means it cannot be experienced at the perceptual level (and explains why the RGB notations look so unintuitive), and which makes your claim that "I clearly, evidently and obviously do experience green as a mixture of yellow and cyan and red as a mixture of yellow and magenta, which latter colour looks only to be in between red and blue because it is a primary component of both" look utterly ridiculous and clearly shows how fanatic you are being in your attempt to discredit the NCS. Next you'll be claiming to "clearly, evidently and obviously" see black as a mixture of yellow, cyan and magenta. Secondly, color opponency and tetrachromacy are two completely different and unrelated things. Mono-/di-/tri-/tetra-/penta-chromacy refers to the number of different kinds of cone cells, it does not refer to the number of perceptual colors (despite what the -chromacy part of the name may misguidingly induce to believe). Maybe you'd first need to understand that cone signals and perceptual colors are completely different things, and that the absorption spectrum of the so-called "red" cone actually peaks at a point in the spectrum perceptually perceived as yellow; the "red" thing in the label "red cone" is a pure convention and has very little to do with what colors are prompted by the stimulation of that cone. For its part, color opponency deals not with the "raw" RGB-type signal produced by the cones, but with the signal resulting from the complex processing of that "raw" signal at a higher level by other kinds of cells, which compare the relative intensities of the different kinds of cones and take into account the "general picture" of larger areas of the retina. A monochromat sees the world in two elementary color percepts (the achromatic ones) because their retinas cannot compare signals from different wavelengths to perceive hue, but can only measure intensity; so they see the world in black and white (percepts corresponding to lack/presence of light). A dichromat (like John Dalton was), sees the world in four elementary color percepts, because their retinas can process visual input from two different wavelengths, so they can perceive one dimension of hue apart from intensity (the blue-vs-yellow dimension, that is, they can discriminate whether the light belongs on the shortwave or longwave part of the spectrum; if the input from both kind of cones is of similar intensity, then an achromat perception of black/white will be triggered instead; read the descriptions Dalton made of how the rainbow looked to him, and the similar descriptions made later by other dichromats, and also by persons who are dichromat on one eye only). A trichromat sees the world in six elementary colors percepts, because the additional kind of cone allows the retina to make further comparisons between the cone signals so as to produce an additional dimension of hue (the red-vs-green one, which allows finer discriminations in the light spectrum than merely short vs. long wavelengths). Expectedly, a tetrachromat (as some women are suspected, though not yet confirmed, to be) would see the world in eight elementary color percepts (the achromat scale of intensity and three pairs of opponent hues, with an additional pair of hues that is unknown to trichromats just like the red/green pair experienced by trichromats is unknown to dichromats). Uaxuctum 22:53, 16 May 2006 (UTC)

Thank you for your argumentation. I fear I haven't made myself sufficiently clear. I will try to elucidate my point of view:

  1. First of all, while talking about "tetrachromism" I wasn't referring to the condition of tetrachromy but to the doctrine that there are four fundamental hues. So you are a tetrachromist, though, I suppose, not a tetrachromat ;o).
  2. I get the impression — but correct me if I'm wrong — that your standpoint is based on a fundamental misunderstanding: that somehow the retinal processing of the trivariant conal input involves a loss of information. But that's not true: the cortex receives all the information needed to create a perception corresponding to the raw RGB input. And the signals are not "very different" from that input. The green-red opponency gives the brain, well, red and green info; and the blue in the blue-yellow opponency gives the brain blue info; and the "yellow" in the latter opponency basically analyses the relative strength of red and green in relation to blue. So the signals are the RGB system with the retina already doing some hierarchical ordering. And therefore I concluded that the most parsimonious hypothesis would be that the subjective perception is ordered along the CMY model. And so I'm not that fanatical in stating that I really perceive red as a mixture of yellow and magenta — and would be greatly surprised if mixing yellow and magenta paint would render green. Again I ask you: wouldn't you? :o)
  3. In all of this I'm presuming that you do not simply equate the retinal activity with the subjective perception and that you're not simply assuming Herring was vindicated by the facts. You would agree e.g. with what is written on a site like this: ?
  4. The reason that the RGB notation is counterintuitive is simply that it is an additive system. But is the derived subtractive CMY system counterintuitive too? Yes, yellow does not really look like a mixture of red and green. But if you mix red and green paint the mixture doesn't look yellow either: so in this respect NCS and CMY are in accordance and the example can't be given as proof of the intuitive superiority of NCS. The examples I gave however are pertinent: according to the NCS mixing cyan and yellow paint is wondrously counterintuitive and we should be flabbergasted to see the demonic concoction turn green. On the other hand it should be completely inexplicable why a mix of red and blue paint doesn't turn bright pink. Again of course parsimony is on the side of the CMY :o)
  5. Whatever the truth of all this, the simple fact is that the NCS or the CIELAB systems are contested. Hård's original colour space of 1968, not even showing saturated magenta or cyan, is taken serious by no one today. The present NCS has a very unclear conceptual description, can indicate no calibration and seems neither to give rise to or to be based on, empirical research. Even if all that you say is true, it can't save the NCS as it functions in the 21st century: as pure pseudoscience. Therefore it should not be presented by a Wikipedia article as established fact. So my edits are the very minimum needed to achieve a modicum of NPOV.--MWAK 19:42, 17 May 2006 (UTC)
Firstly, color opponency is not a "doctrine", it's a widely accepted scientific theory (in the very webpage you mention it is stated clearly that "the Opponent Color Theory of the 19th century physiologist Ewald Hering [...] derived by the analysis of subjective human color vision is in general correct"). Whether you choose to accept this general consensus of the scientific community or not is irrelevant, but please stop talking about color opponency as if it were some sort of delirious crackpot idea or some strange cult doctrine, because it is NOT: it is a scientific theory widely accepted by scholars and researchers of human vision as correct in its general statement (that human vision works at the perceptual level with the black-white, blue-yellow and red-green pairs of elementary percepts, and that the latter two are opponent, that is, that each member of the pair excludes the other). Of course, the functioning of human vision is extremely complex, and involves complex processing of the signal not only at the retinal level, but also at the cortex level, so you shouldn't be surprised to find that we do not have a complete understanding of all the details underlying the opponent channels (the author of the webpage does explain some of the difficulties of studying neural responses), but this doesn't mean their existence is questioned. Secondly, I tell you again that the RGB cone signal doesn't reach the brain. The red-green opponent channel is not the same as the RG components. I told you already that the R component does not correspond to color red, are you unable to understand such a basic fact? The R cone's absorption peak is at the yellow part of the spectrum, not at red (in that webpage you can find a chart showing the absorption peaks of each cone against the spectrum). No cone signal has any direct correspondence with a color perception. White is an elementary color percept, which means a pure, completely unmixed color perception that cannot be described in terms of others (so much so that this color is widely considered the epitome of purity), yet this color perception is prompted when all three cones are equally stimulated (that is, when supposedly, according to your misunderstanding of human vision, one should perceive a color of a mixed hue of red, green and blue, that is, looking as some kind of greenish magenta). The elementary color percepts experienced at the perceptual level are the final result of a very complex processing of the tristimulus signal originating at the cones (that the brain has no direct access to). Our conscious mind (our subjective experience of vision) only has access to the final result of that complex processing, which is full of corrections, adjustments, interpretations and visual effects, and whose color dimension is composed of the six elementary color percepts (the colors actually experienced as looking pure and unmixed), not of raw RGB-type tristimulus retinal cone components. It seems that, even after reading the above webpage that explains the concepts pretty clearly, you haven't yet managed to understand the basics of color opponency and human vision, nor what NCS is about and what it is intended for (CIELab, RGB, CYMK, Hexachrome, NCS, Munsell, Pantone, etc., each have a different design goal and a different purpose, so it's stupid to try to judge one with the parameters of another; leaving aside that Wikipedia is not the place to judge them, but to describe them). Uaxuctum 22:50, 21 May 2006 (UTC)
Allow me to answer:
  1. The theory about the biological mechanism of colour opponency is indeed so well established that we may call its subject a "fact". What had better been called a doctrine however is the supposed implication that there are four basic colour percepts. There is no consensus on this, if only because many psychologists do not believe in percepts :o).
  2. Even among those who do, there has been an at least equally strong tradition that there are three basic percepts: red, yellow and blue; and that these were basic because they were the subtractive primaries. The terms "red" and "blue" were always somewhat imprecise; scientific research then provided the more exact hues, naming them "magenta" and "cyan". According to this hypothesis the cortex creates from the unintuitive additive RGB input a colour perception based on an intuitive inverted subtractive CMY-system. This way there is a congruency between reality as we mostly perceive it happen — when e.g. mixing paint — and how we analyse it by the CMY. Seems quite plausible from a Darwinistic point of view :o).
  3. The CMY met with strong resistance. Some turned to the alternative RYGB as it seemed more akin to the old RYB, forgetting that in the other system three of the basic percepts are the old secondaries: the old "orange" should always have been "red" and the old "violet" is identical to "blue". This system is in fact closer to the raw RGB input and thus, if the CMY prediction is correct, more unintuitive, except for yellow: the example you used was thus not to the point because in this case the RYGB and CMY agree.
  4. The known facts about the biological system allow for the truth of both the RYGB and the CMY. Therefore only psychological research can decide between them. I know of no data about the subjective perception of colour relations that unequivocally corroborate the RYGB. Even those researchers working in a NCS framework before it solidified into dogma, failed to get the desired results, though asking suggestive questions about the "pure hues" red, green and blue: their subjects put magenta opposite to green, not red. And violet not opposite to yellow, showing that reality doesn't simply conform to either the RYGB or the CMY. That's why the present NCS has such a strange blue, remember? But replacing red by magenta was too much ;o). Perhaps we should let the colours speak for themselves:
According to the NCS these fysically correct transitions from yellow to magenta and from yellow to cyan:
should be, both unintuitive and a priori as likely as these transitions:

Also according to the equation of the biological opponency system with the basic percepts (the NCS is more shrewd in this) cyan would be perceived as the simple mix of green and (violet)blue, without loss of saturation, but not merely in between because it is a component of both and thus identical to an unsaturated mix:

Likewise, according to the NCS magenta would be a simple mix of red and blue, but not a component of both:

Hmmm, you appear to be confused. There's really no such thing as "primary colors" - when you take a prism and split sunlight into it's components, the violet is electromagnetic radiation with a wavelength of 400nm, not some combination of electromagnetic radiation of any three other "fundamental" wavelengths, be they red, green, blue, cyan, magenta, or yellow. That's not how visible light works naturally. The cones in the eyes are not, in fact, three narrowspectrum instruments that pick up only pure red (~700nm), green (~530nm), and blue (~470nm) wavelength radiation, blind to all else, with our supposed perception of colors being nothing but mixtures of the three. No, each reacts to a broad range of colors - the M and L cones pick up nearly the entire spectrum, ranges of ~400nm-660nm and ~430nm-700nm respectively, while the S cone is more narrowspectrum, with a range of ~400nm-550nm. They each have peak sensitives - at 570nm (muddy yellow), 540nm (yellowish green), and 420nm (violet-ish) for the M, L, and S cones respectively, but these aren't some primary colors from which all other colors are nothing but components of, indeed, you'd get very poor results trying to use them as primaries. What the brain does is to compare how strongly the cones are firing off, and use that to triangulate the true wavelength of the color in question - this is possible precisely because they measure a broad swath of mostly overlapping frequencies, but at different sensitivities for various parts of the spectrum.
However, this is not a precise process, since it does not do a true hyperspectral analysis of the whole spectrum, and only has three points of analysis, it can be tricked. A mixture of blue and yellow paint is not, in fact, reflecting green electromagnetic radiation - but, the simultaneous sensitization of the cones by cyan and yellow light happens to cause the brain to say "Green!". This does not mean that green is some false color. No, when you use a pure green pigment, actual green wavelengths are reflected by it. But blue and yellow can produce the same perception. Now, there are other colors, like brown, that indeed don't exist as anything but mixtures, but everything in the spectrum of light - the seven distinct shades, and the infinite number of small variations - can exist on its own. When you pick your "primaries" in your color system, all you're trying to do is choose ones that can give you the widest range possible of perceived colors, not find the mystical colors from which derive all others. In fact, no three colors can actually reproduce the entire spectrum - CMY and sRGB both do particularly bad with regards to the green spectrum, and can't represent much of the possible green shades. You can do better if you abandon the slavish devotion to three primaries and expand somewhat - Pantone, for instance, has a process with a full 13. But then you get into boyish mudslinging, "Nuh-uh, CMY make everything! You can only see three colors idiot!", or pedantic arguments where someone goes through colors and points out which mix into which and thus are "only" some mixture of other colors. Look, that Cyan and Yellow may give you a Green, but is that the Green you want to be stuck with for the rest of your life? Three is just the smallest thinkable number of colors that gives you something basically workable, it's not the end all be all, you can get a much wider spectrum than that. This is not somehow a claim that the brain has more than three kinds of color pathways, again, the cones pick up a large range of colors, not just three, using that as the basis of an argument for three primary colors just shows a misunderstanding of how the color system works. (talk) 13:41, 21 April 2014 (UTC)
Well, I'm not quite sure whether I understand you correctly. Perhaps we can break down the problem into some relevant aspects.
  1. Indeed, on the purely physical level there are not three wavelengths in the light spectrum that are somehow "elementary". Wavelengths can only become elementary in relation to some perception system.
  2. A good way to realise this, is indeed to remember that the three human cone types, part of a biological perception system especially relevant to human beings, are each sensitive to a range of wavelengths. Of course, those ranges too are only elementary in relation to some perception system.
  3. That each cone type can be stimulated by a range of wavelengths does not imply that there are not three primary colours. At first blush there even seems to be no connection between the one and the other because the cone does not seem to convey the information by which exact range it was stimulated. That a "green" cone can be activated by a mixture of yellow and cyan light as well by a narrow band of green light does not imply that green is not an additive primary. "Green" is first of all some effect of green cone stimulation; only secondarily, in relation to the effect it has on this part of the perception system, can a wavelength be called "green".
  4. From a certain practical point of view, namely that of the art of creating colour perceptions by e.g. illuminating objects or mixing paint, "primary colours" have no absolute functional value. Using two colours to mix paint is poor but often still useful; using the "true primaries" allows for a far superior colour range but will still be imperfect. Using four is better, thirteen more so.
  5. However, the discussion above is not pragmatist. It addresses the problem whether a phenomenal colour experience exists (more or less whether there is a conscious perception of a special way in which colours be) and if so (by far the more interesting possibility) whether there are elementary colours (irreducible "simple" hues of which other hues are "complex" combinations) and if so whether there are three or four of them in human colour experience (let's not pretend we can speak for other species). Or eight as you seem to suggest (seven shades plus magenta?). The Natural Color System is based on the premiss that there are four elementary colours. In the phenomenological sense.--MWAK (talk) 16:57, 21 April 2014 (UTC)


I do not want to add to the discussion but just to finalize it. The claim of NCS: the hue of colors ar best described in RGBY. To prove that claim only (verifiable hopefully) experimentation with test persons are conlusive. Such a test could be showing colors to persons and asking them to describe it in RGBY. Other test persons would be asked to describe it in RGB, CMY, the 6 colors according to Itten (bauhuas) color theory, the 5 colors of Munsell or any other number of colors. To me (but this is my personel opinion) the best results will be with colors that most test persons are familiar (have in memory) with (probably the colors in their native language as not everyone gives names to the same colors) an secondly it seems logical that the more colors test persons are familiar with the better they would score. In that case I would bet on red, green, orange, yellow, blue, purple as these are the colors I experience most people know. So is there any reference to such a test?

One could now try to explain the results on a biological basis. From the article as it is now written conclusions are drawn from biological insights. Without the above experiment such explanation could be right or wrong, you just don't know (you didn't test). Such a conclusion would be evenly right are wrong as the conclusion that the primaries are GYB are favored because of how the eye responds to light. My guess is that both are wrong because there is even a higher level: our memory. The more colors we can memories the better we will score on the test.

So I repeat: I don't want a discussion before the question is answered. Does anyone know of such a test? --BartYgor 13:12, 17 July 2007 (UTC)

I'm going to add some thoughts. Let's say it is true that RGBY are biolocially printed in our brain. Then why would we limit a couloursystem to these four? We also have are memory to help as, and if we can imagine purple and practise on it, then accuraccy would rize on naming colors. IN fact that is exactly was NCS does: they don't stick to 4 colors; but add 36 in their colour circle. People who practise will have a good mental picture of what G40Y looks like. So proficient NCS users would probably score good on the above test. But so would a system with 5 basic colors and some numerically defined in betweens. And is it's true that RGBY are biolocially printed in our brain. Then we could easily test that: we would offer a multitude of testpersons a choose between a lot of reds yet all with the same luminance and saturation and ask them to choose that what seems to be the most reddish red. If the claims were true we would see a peak in numbers to favor one type of (brain)red and if NCS really makes claims come true then their basic red should conform whit that peak! If not such a peak exist or even two or three peaks than all claims are falls. So again has anybody heard of such a study? --BartYgor 16:18, 18 July 2007 (UTC)

My two cents: A color space is some arrangement of some colors. The only thing a proper color space needs is a proper definition, so people can reproduce it and work with it. If this arrangement is useful to some people, or has interesting properties, we should accept it; no matter if it was invented by a drunk Van Gogh, it has been developed by the NASA, or its rules had been found inside a 4000 year old pyramid. All this things are noise. Nonetheless, if some color space claims to have some property, it must be demonstrated. Notice mathematical propierties are easier to demonstrate than neurobiological ones. At the end, color spaces are like world maps; there is no obvious, natural or perfect solution. So any useful contribution is a partial solution.

RGB NCS Conversion tools[edit]


I think that the links in the article to NCS/RGB conversion tools should be removed and here is why: RGB (depending on the algorythm used to get from tristimulus to RGB) do not cover the entire 'space' of visible colors. Therefore the calculation of NCS to/from RGB cannot be good as for instance the uttermost red is more red than the R in the RGB. The 100% red in NCS however is probably somewhere in between. Without knowning the XYZ (or any other scientific color space) values of the red, yellow, blue, green and (yes) even the white and black youo cannot calculate the RGB of vice versa. There are even probably NCS colors that do not have a valid RGB value.

Besides this what I would like to know is: what are the references of the NCS system ?

TheBlob 13:46, 20 February 2006 (UTC)

I agree that unless NCS is defined in terms of uncalibrated primaries, it makes no sense to offer "conversions" to "RGB". Perhaps sRGB would make sense, but the problem seems to be that very few Wikipedia authors understand which color spaces are absolute and which are not, and that you cannot convert without a framework. Notinasnaid 14:00, 20 February 2006 (UTC)
Furthermore, no two of the tools cited show the same result (tested with NCS S 3030 Y60R); you are forced to 'believe' in one of four different results. Also no information/distinction found on differences between NCS and NCS S (second edition). —Preceding unsigned comment added by (talk) 19:41, 28 March 2009 (UTC)
A little more research throws up this very interesting comment, from a developer attempting this "It has been difficult to find two descriptions of NCS that agree. It has been impossible to find a description of the mathematical details of NCS." [1] Given that I have to conclude that NCS is a reference-based system - more akin to Pantone. This isn't a negative thing in itself, but for color scientists it makes it very hard to work with, and almost meaningless to convert. I couldn't find any technical information about its points of reference (e.g. "What is Red") on the NCS web site. The web site does talk of "deep frozen reference samples". Given that, the article should probably not talk about conversions, except to warn of these points. Notinasnaid 14:59, 20 February 2006 (UTC)
I fear that you are taking the entire NCS far too serious. Asking questions like these is akin to asking your astrologer where Sedna is in the horoscope. :o)--MWAK 18:17, 9 April 2006 (UTC)

Deleted text[edit]

I reinstated the text "While the science may be quite difficult to understand most users of the system will simply be choosing or matching colors with printed reference cards." because there was not a clear explanation of why it was deleted. This text had nothing to do with discrediting anything, but was an attempt to make it clearer to the casual reader how NCS can be used by them. I did this because I read the article and could not understand how NCS fitted into other color spaces, so I did a little research to improve the article. Is this statement inaccurate? I don't know much about NCS, but it seemed to be accurate. Notinasnaid 19:00, 15 May 2006 (UTC)

Sorry, I didn't notice that part. I was trying to revert the edits by the other user, not yours. Uaxuctum 22:53, 16 May 2006 (UTC)

NCS to RGB conversion[edit]

Is there a mathematical formula for calculating RGB values from NCS? (Obviously not an exact one, as this would be dependent on things like monitor gamma) HymylyT@C 19:16, 19 June 2006 (UTC)

See "RGB NCS Conversion tools" above... Notinasnaid 20:07, 19 June 2006 (UTC)
Silly me ;) HymylyT@C

Olympic Flag[edit]

I've removed the Olympic Flag logo, as it's copyrighted, and can only be used in certain articles per Wikipedia's Fair Use guidelines. Perhaps someone can just mention the Olympic Flag in the article? ...Sorry :o\ tiZom(2¢) 02:05, 23 June 2006 (UTC)

If green is opposite to red they why isn't orange opposite to blue and purple opposite to yellow?[edit]

If green is opposite to red they why isn't orange opposite to blue and purple opposite to yellow? (talk) 14:58, 2 April 2010 (UTC) KGB

You'll have to explain what you mean. The NCS is premised on a red–green opposition and a yellow–blue opposition, first proposed by Hering. That’s the basis of the model. Why do you think orange should be opposite blue? –jacobolus (t) 19:57, 2 April 2010 (UTC)
The question is presumably based on the notion that Hering based his system on the concept of complementary colours. If the traditional complementary pairs had been used, apart from red-green, indeed yellow-purple and orange-blue should have been present. However, his system was based on the, rather confused, line of thought that because a pure yellow cannot be mixed from red and green an additional, by him then presumed to be chemical, mechanism should account for the perception of yellow. He assumed that one chemical reaction accounted for the differentiation between red and green and another for that between yellow and blue. Of course, if Hering would have fully thought his system through, he should have noticed that a pure cyan and magenta cannot be mixed from other colours either and he might well have made these a third opposition pair — or rearranged the whole into yellow-blue, red-cyan and green-magenta, thus reaching the present scientific system with three true complementary colour pairs. Sadly he didn't.--MWAK (talk) 06:27, 3 April 2010 (UTC)
Your understanding is incorrect. “Red”, “yellow”, “green”, and “blue” are the four so-called “unique hues”, about which there has been a tremendous amount of psychological research (go search the academic literature for “unique hue”). Human observers will tell you that “cyan” (a term that wasn’t in wide popular use until pretty recently; in Hering’s time color printing was rare and color mixture was as far as I know mostly studied in terms of paint or dye mixtures, etc.; cyan was just a greek word for blue that was adopted to fit the sort of greenish blue used for printing, to distinguish it from what someone would call a pure blue) is just a greenish blue, and likewise that magenta (a term invented because the process ink wasn’t quite red or purple) is a mix of some blue and mostly red. By contrast, no human who hasn't been trained will claim that a bright blue or red is a mix of two more basic colors. The present “scientific” system is a great deal more complicated than you claim, and your yellow-blue red-cyan green-magenta system is pretty arbitrary from the perspective of human vision. If you want to learn about how color vision works, I suggest Bruce MacEvoy’s page about it, which is clear, readable, and quite detailed. Incidentally, the most recent standardized “present scientific system” model is CIECAM02, whose Jab form has an a dimension representing red–green opponency, and a b dimension representing yellow–blue opponency. The two are rough correlates of the color difference signals computed in the neurons of the eye and sent to the brain through the optic nerve. –jacobolus (t) 13:32, 3 April 2010 (UTC)
Oh, I see now that Uaxuctum tried to explain this to you in 2006. Well, again, I urge you to go read MacEvoy’s page. It does a far better job than one of us will be able to do in quick summary on this talk page. If you read carefully, I trust you’ll find it a quite convincing explanation. (After that, you might try reading Mark Fairchild’s 2005 book (well, the 2nd edition was in 2005) Color Appearance Models, or one of the classic color science textbooks by Hunt or Wyszecki & Stiles.) –jacobolus (t) 13:44, 3 April 2010 (UTC)
Well, I haven't changed my opinion over the years. I still think the NCS is based on a fatal mix of poor phenomenology with even poorer science :o). You seem to have convinced yourself that there is some solid empirical basis for all of this; I can only say I strongly disagree. What are the results of a "unique hue" experiment capable of testing (instead of affirming) the system, e.g. by offering the subjects samples of magenta and cyan in varying degrees of saturation to see whether they are equally capable of analysing these hues in terms of red/blue and green/blue when the saturation increases? Can subjects analyse an unsaturated yellow (i.e. a bronze green) in terms of red and green? Which colour perception experiments have even used a sufficiently saturated cyan and magenta? Why are the "unique hue" foci strongly divergent from the stimulus optima? Why is the ganglion system presented as strong proof when there is a 0,5 chance it would be arranged this way, irrespective of the subjective experience it might cause? These are the kind of questions a critical science should ask itself.--MWAK (talk) 15:54, 3 April 2010 (UTC)
“Solid empirical basis” for color opponency and the existence of “unique hues”? Yes, there are dozens if not hundreds of papers about it, starting with Jameson & Hurvich’s work in the 1950s . You can find experiments using colored lights, using Munsell swatches, etc. The results of these experiments are (1) to confirm the general theory, and (2) to show that it is quite a simplification of the extremely complex process that is color vision. Yes, RYGB are in some sense more “fundamental” than “cyan” or “magenta”. However, different observers pick somewhat different unique hues (i.e. your “unique green” is a different hue from my “unique green”, and so on), and people’s color perception also changes as they age. Men’s vision differs somewhat from women’s. Obviously many people have various visual deficiencies. &c. The behavior of the ganglion system is an explanation for the experimental results we see; it is not a “proof” of anything. As for NCS – I don’t find it especially useful for any of my own work. –jacobolus (t) 04:08, 4 April 2010 (UTC)
Well, I do not deny that the biological mechanism referred to exists, nor that people can indicate "unique hues" under some operationalisation of that concept. However, the NCS uses some very specific philosophical concept of elementary colours, as irreducible and unanalysable qualia of human colour experience. The biological mechanism of colour opponency is not a sufficient explanation of this. Especially so when the measured "unique hues" do not at all coincide with the predicted ones. Western test subjects will call the opponency "blue" purple: it does not even fall within their blue colour sector. The NCS has changed its colours over the years for some reason! I agree that red, yellow, green and blue (at least some blue) are somehow special. But is this equal to their being "elementary"? Or is it some psychological mechanism reflecting ape priorities: the red of blood, the dark yellow of ripe fruit, the green of vegetation, the blue of the sky? Perhaps they are Urfarben, as Hering would put it, but not Reinfarben. And a "unique hue" might not represent a philosophical simplex but a complex of psychological, cultural and phenomenological factors. Of course, admitting that will cost you some parsimony, but then very much parsimony is gained by allowing yellow, cyan and magenta to be the elementary colours.
And since when do Munsell swatches show a saturated magenta ;o)?--MWAK (talk) 19:26, 6 April 2010 (UTC)
‘"unique hues" do not at all coincide with the predicted ones.’ — By whose prediction? They seem to vary dramatically from one observer to another. ‘Western test subjects will call the opponency "blue" purple’ — What is “opponency blue”? ‘And a "unique hue" might not represent a philosophical simplex but a complex of psychological, cultural and phenomenological factors.’ — I’m not sure anyone is arguing otherwise.... ‘The NCS has changed its colours over the years’ — Really? That should be probably described in detail in this article. I don’t actually know all that much about the full history of the NCS. –jacobolus (t) 03:03, 7 April 2010 (UTC)
Because the essence of the NCS is that their elementary colours are simplices, admitting that "unique hues" can be determined by a whole range of factors makes "unique hue" research a less than ideal confirmation of the NCS tenets. Certainly the link with colour opponency becomes very tenuous. If that mechanism were dominant, this would e.g. predict that about 450 nm would be called "blue" (so I indicated this as "opponency blue"). Hering himself from intuition estimated it at 470 nm. Now look at the image in the article; the blue there is typical of the new 1995 NCS-standard. To compare the two in what I hope is a fair approximation given the limits of screen technology:
"opponency blue":        1995 NCS elementary blue:       
A not inconsiderable difference I would say.--MWAK (talk) 19:16, 9 April 2010 (UTC)
Well, Jameson/Hurvich’s experimental evidence suggests something like 470 nm, with quite a bit of variation from one observer to the next. If we use the CIECAM02 model (according to the CIE 1931 standard observer definition, and assuming a D65 white point), that’s a hue angle of 245°. At lightness J=50, the most colorful color of the same hue that the sRGB gamut will admit is (which to be honest looks a touch greenish to me). I have no idea where you got from, but that’s the same hue as a spectral color at about 420 nm. –jacobolus (t) 15:28, 10 April 2010 (UTC)
Just to be complete, at the same hue as the spectral colors, and at lightness J=50, we have: 430 nm ; 440 nm ; 450 nm ; 460 nm ; 470 nm ; 480 nm . –jacobolus (t) 15:39, 10 April 2010 (UTC)
By the way, the CIECAM02 model defines its unique hues (R, Y, G, B) to be at angles 20.14°, 90°, 164.25°, 237.53°. In sRGB, some possible colors (again, max chroma, J = 50) at these hues are , , , . See here for detailsjacobolus (t) 19:01, 27 April 2010 (UTC)
A rather belated response :o): you had an excellent point in that I myself did not use a coherent notion of opponency colour, implicitly defining it both as a stimulus optimum and as the "purest" (in Hering's sense) stimulus available. In the latter case the difference is of course enormous. But even is very different from ...--MWAK (talk) 14:59, 20 January 2013 (UTC)
These hot Scandinavian guys apparently use some color terms, especially "red", in their unique way. Their "unique red" is, actually, our crimson. Then, its opposite color is something like turquoise, which they describe as "green". In light of these deviations, their claims about "simplicity" look humorous. Incnis Mrsi (talk) 11:50, 17 January 2013 (UTC)
Well, I'm very critical of the test designs usually applied at such experiments — but they are nevertheless that: scientific experiments. So these "unique hues" are those indicated by test subjects, not simply arbitrary claims. But indeed, it seems unlikely they exhibit much simplicity. The simplest explanation would be that they represent some synthetic experience of Urfarben and the real simple colours: yellow, magenta and cyan.--MWAK (talk) 14:59, 20 January 2013 (UTC)
I'm very glad to see a critical guy in this jungle. Could you help me a bit with Talk:Complementary colors #red and green, talk:Unique hues and talk:RYB color model, please? Incnis Mrsi (talk) 16:17, 20 January 2013 (UTC)

It's ironical[edit]

I never heard before about a color system so fixed on cultural prejudices and thence, completely unnatural. They postulate that the spectral red is a crimson-like hue with a small admixture of yellow… which crank may invent something better? Incnis Mrsi (talk) 11:50, 17 January 2013 (UTC)

Problem is some authors dragging in the psychological aspects of the scientific theory into an article about engineering standards. I'll rewrite this, properly sourced this time. BP OMowe (talk) 18:59, 8 April 2013 (UTC)
Incnis Mrsi, are you saying that sRGB(255, 0, 0) looks pure to you or did you not calibrate your monitors, because NCS elementary red sure does not look like there is any yellow it in at all. (talk) 15:56, 6 July 2014 (UTC)
Ah, but spectral red is not the modern NCS "pure red". And to anyone it should be obvious that both are subtractive mixtures of yellow and magenta :o).--MWAK (talk) 19:34, 6 July 2014 (UTC)

Name of company & name of system[edit]

While I realise that there are folks out there who get most frightfully upset over the use of US vs UK spellings in Wikipedia articles, I must confess it's something that bores the hell out of me. In other words, I couldn't care less whether an article is written in whichever of the two spelling variants is chosen by the user creating the article, provided, of course, that the spelling is consistent throughout and, as I point out below, depending on whether we're talking about "common" nouns or proper nouns.

As pointed out in the corresponding subsection of the Wikipedia policy regarding article titles which, funnily enough, gives the precise example that "both color and colour are acceptable and both spellings are found in article titles (such as color gel and colour state)". Said subsection also gives the example "Australian Defence Force, United States Secretary of Defense". Thus, regardless of our own leanings, Wikipedia will naturally respect the "correct" spelling in the case of proper nouns.

This article is clearly not about a common noun but about a standard or product created by a company. According to that company's website, both the company (NCS Colour AB) and their standard, product, call it what you will, Natural Colour System, uses that -our variant (sources: Background & About us). In other words, they chooses the name, so they gets to decide which variant of English they wish to use. And Wikipedia simply gets to reflect that. I personally can never get used to including that there hyphen in Coca-Cola, but that's the way it is, and that's how Wikipedia writes it.

Furthermore, while the file "New Practice for Specifying Color by the Natural Color System (NCS)" at ASTM International does use the -or spelling in its title and content, it also states the following: "1.7 Acknowledgement – ASTM International has obtained a license from the Scandinavian Colour Institute AB (SCI) to use the NCS System. NCS-Natural Colour System is a trademark of the SCI, Stockholm, Sweden and is protected by copyright ( All rights reserved."

Curiously, the IP user who created the article back in 2002, stated that "Yes, they spell "color" differently in the name of the standard and the English translation of their name)."...

There's no doubt whatsoever in my mind that the article title should be Natural Colour System –crystal clear– but I've been asked to broach the matter here, so the ball's in the Community's court. Regards, --Technopat (talk) 00:23, 21 August 2014 (UTC)

A couple of things argue against your claim that "Colour" is more correct here. First, if you check at and do a trademark search, you'll find "Natural Color System" and "NCS Natural Color System" registered to a Swedish company (I didn't look close enough to say it's the right Swedish company, but I'm guessing so; please check); I find no corresponding US trademark with the British spelling. Second, usage in English sources is about evenly split, so we should WP:RETAIN what we have. Dicklyon (talk) 04:38, 21 August 2014 (UTC)