Talk:Unique hues

From Wikipedia, the free encyclopedia
Jump to: navigation, search
WikiProject Color (Rated Stub-class)
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.
Stub-Class article Stub  This article has been rated as Stub-Class on the project's quality scale.
 ???  This article has not yet received a rating on the project's importance scale.

Jacobolus and Colonel Warden persuaded me that the topic is notable and my argument about hue angles was wrong. But many problems remained. Incnis Mrsi (talk) 07:45, 19 January 2013 (UTC)

What is missing[edit]

  • Does existence of four unique hues really follow from opponent process theory? What namely the theory says about them? "A red which appears to have no yellow or blue in it" is something not scientific.
  • Which experiments can determine "unique hues"?
  • Do experiments show that the chromaticity of an unique hue is linear? I mean, that some radius from the white point may be determined.
  • Do opposite "unique hues" have necessarily to be complementary colors?
  • Do experiments show unambiguously that there are exactly 4?
  • How CIECAM02's unique hues and experimentally determined ones are related?
  • How Natural Color System's cardinal hues are related to all the above?

Incnis Mrsi (talk) 07:45, 19 January 2013 (UTC)

Let's try to answer this.
  1. No, from the existence of a opponent process it does not logically follow that there are unique hues, or that there are four of them. Or vice versa. That means it is strictly incorrect to make claims like "opponent colours predict the existence of four unique hues".
  2. Very good question :o). Generally, the experiments already assume that unique hues exist, that there are four of them and that they are red, yellow, green en blue. The problem is that most languages have no basic colour term distinction between blue and cyan and between magenta and "bright red" (calling them both "pink"). You might see that as an indication that cyan and magenta are not unique hues — or conversely that the cultural development of human societies is still wanting :o).
  3. Certainly not. This is due to the combined Abney effect and Bezold-Brücke effect: lightening and darkening changes the hue perception.
  4. No, unless you define them as such. Experimentally determined unique hues generally do not coincide with colorimetrical complementary colours.
  5. No, this has not been seriously tested. However, tests show quite convincingly that almost all "normal" human test subjects use a conceptual scheme including red, yellow and green and are pretty much in agreement about their loci. "Blue", cyan and magenta are highly problematical. Typically the tests do not allow for the possibility that there might be six unique hues and do not present the test subjects with a saturated magenta.
  6. The CIECAM loci are not defined as unique hues and do not coincide with experimentally determined unique hues, perhaps with the exception of "yellow".
  7. Well, the original NCS colours were phenomenological: the researchers by reflecting on their personal perception determined the loci (and then performed some circular testing, "affirming" said loci). Modern NCS colours are based on experimentally determined unique hues within the limits mentioned above: they essentially assume that unique hues exist, that there are four of them and that they are red, yellow, green en blue.
— Preceding unsigned comment added by MWAK (talkcontribs) 10:49, 21 January 2013‎ MWAK (UTC)
You slightly misunderstood my "linearity question". You answered about 1d subspaces in the 3d XYZ space. Yes, there are some changes where the brightness varies in several orders of magnitude, but I am not interested in how precisely is "chroma" defined. I ask about the definition of "hue" assuming than it is all right with chromaticity. Does an "unique hue" form a line segment in the plane of chromaticity which starts from the white point, or the colorimetric hue of the "unique hue" depends on saturation? If you prefer vector spaces over projectivizations, does an "unique hue" form something like a half-plane in XYZ? Incnis Mrsi (talk) 11:59, 21 January 2013 (UTC)
Oops, I captured Bezold–Brücke, but for some reason ignored Abney which is exactly about saturation. Though, all of this has little significance unless we clarify, which experiments can determine an unique hue. Just asking test persons to choose colors is an anti-scientific approach because it shows cultural prejudices, not psychological perception. Incnis Mrsi (talk) 14:12, 21 January 2013 (UTC)
That is an obvious concern. However, the operationalisation is even more of a problem. You can't ask test subjects "Indicate the loci that are the foci of the irreducible elementary qualia of your colour perception" for the only response you'll get will be a glazed look in their eyes :o). So these kind of questions are posed: "What red is least like blue or yellow? What yellow is least like red or green? What green is least like yellow or blue? What blue is least like green or red?". Remarkably, the result is never that cyan or magenta are unique hues :o). And because "red" is closer to yellow than to "blue" the test design will pull red towards magenta: hence the crimson. This again will pull yellow towards red. As the "blue" that is inquired after in fact includes both cyan and the additive primary blue the test subjects will choose some hue in between. This again pulls green towards cyan. Lo and behold: the four "unique hues" of the modern NCS.--MWAK (talk) 08:46, 22 January 2013 (UTC)
So, neither "opponent process" nor CIECAM02 provide a firm foundation? Incnis Mrsi (talk) 14:12, 21 January 2013 (UTC)
No, they at most represent a tantalising indication. By necessity there are several fundamental gaps between the physical processing of stimuli in the eye and a system purporting to describe the colour logic of phenomenal experience.--MWAK (talk) 08:46, 22 January 2013 (UTC)

What was doubtful/contested[edit]

The things from Special:PermanentLink/487322572 which were not likely supported by sources, were synthetic or non-neutral.

  • Chenrezig Sand Mandala.jpg – which namely hues does it use?
  • Color names for these four colors blah-blah-blah… – false. Color names such as "green" or "blue" cover broad range of hues.
  • Leonardo da Vinci – what exactly he painted? He was a painter AFAIK.

Incnis Mrsi (talk) 07:45, 19 January 2013 (UTC)

In any case, da Vinci was not the first to propose a system with four main colours, that was Leone Battista Alberti. Da Vinci merely observed that several distinctive colours can be distinguished.--MWAK (talk) 09:09, 22 January 2013 (UTC)

How can we fix this?[edit]

There is a great deal of variability when defining unique hues experimentally.[3] Often the results show a great deal of interobserver and intraobserver variability leading much debate on the number of unique hues.[4] Another source of variability is environmental factors in color naming. Despite the inconsistencies...

This reads more like an opinion piece than an introductory explanation. It is vague since it doesn't say what is varying.

One persistent criticism of unique hues is that the relationship between them and wavelengths is highly variable. If that was what is intended, it should say so. However, I don't think it goes in the lede, since it's one of the more glib and uninformed criticisms. Vision researchers long ago discovered that the hue-to-wavelength relationship can vary due to a number of factors, with new ones being discovered all the time. Among them Bezold–Brücke shift, Abney effect,Metamerism (color), age-related yellowing of the lens, reduction in MPOD (macular pigment optical density), normal genetic variation, body temperature, etc.

There are actually more substantive criticisms. Here for example. Zyxwv99 (talk) 23:36, 25 May 2014 (UTC)

Regarding the current wording of the article, as well as the comments by MWAK above, they seem to be misrepresenting the literature on the topic, where the existence of the four unique hues at a phenomenological level is widely accepted and not something under serious question. Note that even the "substantive criticism" in the above cited article reads as: "The results of both experiments question subjects' abilities to identify certain hues as unique" (note that they say their experiments question the "the subjects' abilities to identify", and note also that they do not seem to be clarifying or even seem to be really taking into account the quite relevant distinction between the realm of perceptual chromatic qualia and the realm of linguistic color labels), "Red, green, blue, yellow, and white have been distinguished from other hues as unique" (white is obviously not a "hue", but this mistake aside, note they word it as "have been distinguished" and not as "have been purported" or "have been claimed"), "hue scaling, which has often been used to support the existence of unique hues, but has never been attempted with a set of non-unique primaries" (note that here they themselves refer to orange, purple, teal, and lime, as "non-unique", and do not seem to be trying to actually claim these as perceptually unique; rather their experiments seem to be trying to criticize some of the methodologies used to "scientifically prove" the existence of the unique hues, while not actually addressing why the very idea that certain particular hues appear to be phenomenologically "unique" while certain others appear to be phenomenologically "mixed" arose in the minds of people and prompted the most scientifically-inclined among them to feel the need to find ways to "scientifically prove" this intuitive notion). I wonder how test subjects might have felt when tasked with trying to describe, say, the color of the sky, or the color of the soviet flag, in terms of "orange", "purple", "teal" and "lime"; one really needs to resort to linguistic creativeness in order to come up with things like "the soviet flag is a purplish orange" or "the sky looks a light purple-teal" (which are like the only minimally feasible descriptions in terms of those terms, and even so seem totally conceptually crazy). I mean, are you kidding me? These four hues are so obviously non-unique, so clearly perceptually mixed-looking, that it sounds ludicrously nonsensical even to entertain the thought of trying to use orange as a primary to describe red (at a perceptual/phenomenological level) rather than viceversa—the uniqueness of red and the non-uniqueness of orange being so perceptually in-your-face obvious. But if you ask test subjects to try and describe red in terms of orange, purple and other non-uniques, they will attempt to complete their assigned task as best as they can (they are being paid to undergo the test, after all, so I don't think any test subject would have the gall to just cry foul and point to the king's new clothes). I have never seen any literature on unique hues actually claiming, or trying to prove, cyan or magenta as unique hues. These are perceptually regarded (and linguistically describable) as just a kind of "light blue" (i.e., blue+white) and a kind of "light [bluish] red" (i.e., red+white+some blue), and I don't think even the most CMYK-obsessed people would seriously attempt to dispute these perceptual descriptions and try to call blue a "magentish cyan" or red a "yellowish magenta" (or some other similar nonsense). Mistaking cyan or magenta for "basic" colors would show a serious misunderstanding of the nature and distinction between the phenomenological primaries and the sensory and the color-reproduction primaries (and, by the way, calling the latter kinds primary "colors" is really a misnomer, since they are primary inks or primary light-mixing wavelengths or primary cone spectral absorption peaks, and not properly "colors" per se since colors as we experience them are really only a product of the mind and therefore the term "color" only has its true sense in a phenomenological dimension, and at the physical and biological level there are only light wavelengths, reflectance and absorption spectra, and electrical impulses, and the relationship between these and the phenomenological dimension is highly complex as well as widely variable among individuals and between species and also depending on environmental factors such as the illuminant, viewing conditions, visual context, etc). The intersubject variability in the placement of the loci for some of the unique hues (particularly green and blue) has nothing to do with actually "doubting" the existence of these as unique hues, and this variability is easily explainable taking into account well-known phenomena (such as that the blue end of the spectrum is far larger than the red end due to the distribution of absorption spectra, so that the ability of our visual system to make fine-grained color-spectral distinctions near the red end is far greater than towards the blue end; that the exact absorption spectra of cones varies from person to person for several reasons, since there are several genetically distinct kinds of retinal pigments for the "same" function but with slightly different absorption spectra, and the density distribution of each kind of cone is also variable from person to person, and the lens in the eye yellows with age, thus altering the spectrum of the light passing through it by shifting blue towards green, etc.). I think the rewording of the lead of this article in the direction of casting doubt on the wide consensus about the four unique hues, is totally unwarranted, and basically appears to be reflecting a quite bizarre personal view and not at all what the literature on the subject says (which, moreover, should be quite obvious to the mind of anyone who experiences "normal" color perception; not to mention that the special, salient status of these four hues, together with the achromatic poles, as "more basic" colors than other colors, seems widely recognized even at a cultural level, where these six are by far the most frequently chosen, as in flags, educational toys, logos, etc., even though Europen languages have 11 "basic" color terms so one would expect colors like gray, pink, brown, orange, or purple, to be more or less as equally preferred as white, black, red, yellow, green, and blue, but they are not). Uaxuctum (talk) 23:11, 8 May 2015 (UTC)