Talk:Primary color/Archive 2

Archive 1

Archive 2

Archive 3

Archive 4

Article is too confusing

The recent reversion to the changes I introduced are unfortunate, the lead paragraph is inconsistent with the rest of the article e.g.

" For human applications, three primary colors are typically used, since human color vision is usually trichromatic."

What is a "human application"? This sentence is confusing since we use 4 pigments in a CMYK print cartridge. Oil painting can use 4 "primaries" (e.g. an ochre, ultramarine blue, cad red and titanium white). There are 6 colors in the "psychological primaries" section in the same article.

"Any particular choice for a given set of primary colors is derived from the spectral sensitivity of each of the human cone photoreceptors;" But the psychological primaries apparently have much more to do with relative responses of photoreceptors as opposed to their sensitvities.

I think the article should really open with the broad idea that there no specific pigments or wavelengths that are "true" primaries, all color reproduction applications balance gamut size with cost in specific contexts. Understanding why specific wavelengths/pigments are selected generally requires a good understanding of color vision and the specific application.Maneesh (talk) 04:07, 3 January 2016 (UTC)

You had inserted this paragraph in before the WP:LEDE paragraph:

Primary colors refer to the notion that the very large number of color sensations can be effectively simulated by combining a small number of stimuli made from a physical source such as a filament, illuminated pigment mixture etc. Primary color notions are useful in understanding the trichromatic basis of human color vision, but do not completely explain all aspects of color perception (i.e. the opponent process). Red, green and blue phosphors of a pixel on a computer monitor and cyan, magenta and yellow pigments in a printer cartridge are both examples of using a small set of "primaries" to mix a large range of colors; but the specific primary colors used in each of these processes are an outcome of multiple factors (including cost). Understanding why specific primaries are chosen in an application generally involves an understanding of the detailed physics, chemistry and biology of color vision. In summary there are no "true" primary colors defined in terms of wavelengths, pigments or their mixtures; only established heuristics that perform well for color reproduction in different applications (e.g. using a palette of ultramarine blue, burnt umber, cadmium red, yellow ochre and titanium white for oil painting).

There may be some useful ideas there, but it's not a decent lede. Starting with "X refers to" is always wrong. Better to say what x is, as the current lede does. And what do you mean by simulated? I don't understand the point about "notions are useful in understanding the trichromatic basis of human color vision". Seems more the other way around. Nobody is suggesting that primary colors are about explaining or understanding color vision. Why the scare quotes on "primaries", and what is a "true" primary color, which doesn't exist anyway, and why mention it in the lede if it doesn't exist? What introduce a palette of oil paints here?

I recommend you carefully try to work some of these ideas into the article, which is summarized by the current lede. If there's a problem with the lede, please clarify what that problem is.

To answer "what is a human application?", I think it means an application intended for viewing by a human, such as a color display or print. Dicklyon (talk) 06:42, 3 January 2016 (UTC)

It is regrettable that such a fundamental concept of color is, I feel, explained so poorly in this article (as it is in many places). I meant "stimulated" where I wrote "simulated" (easy to see by the occurrence of "stimuli"). Understanding "X Y Z" as the true "imaginary" primaries makes sense, and can only be understood with a basic understanding of trichromatic color perception. If you are not familiar with Bruce MacEvoy's handprint site: http://www.handprint.com/HP/WCL/color6.html, please do look at it to understand why the asserting the existence of primary colors is something that needs to be done with care (hence the scarequotes). The current lede simply isn't consistent with other information in the article (as I've already shown, e.g. number of primaries), it really needs to be redone.Maneesh (talk) 00:55, 4 January 2016 (UTC)

The lead says, "Primary colors are sets of colors that can be combined to make a useful range of colors." You seem to be accusing it of something it doesn't say. And there's nothing special about the XYZ system; it's just an arbitrary convention. Also your "combining a small number of stimuli" concept doesn't work at all with subtractive/pigment colorants. Dicklyon (talk) 01:02, 4 January 2016 (UTC)

Let's start again. The existing article doesn't explain the notion of primary colors well (consider the handprint link that shows we must consider primary colors as either "imaginary" (XYZ) or imperfect (cannot be mixed in some way to realize the entire gamut). The handprint article describes this overall idea near the top of the article:

A major theme is that "primary" colors are either imaginary or imperfect. That is, primary colors are either imaginary sensations you cannot see — and "colors you can't see" aren't really colors — or they are actual lights or paints that cannot mix all possible colors, which means they aren't really "primary".

This article should do the same, given the confusion about primary colors.

The existing text says:

For human applications, three primary colors are typically used, since human color vision is usually trichromatic.

"Human applications" seems remarkably specific, I'm not aware of significant applications of color outside humans. The article then goes on an describes NCS, where there are 6 primaries and CMYK where there are 4. How does this sentence make sense?Maneesh (talk) 01:23, 4 January 2016 (UTC)

But when the handprint guys say "they aren't really 'primary'", they are redefining what "primary" means, or using a definition at odds with what this article is about. They assert that "some art school graduates develop a rigid attachment to primary colors and the formulaic approach to color mixing that goes with them." Is that really true? If so, why do art stores stock so many different paints and pigments? I think they made this up. Everyone who knows anything about primary colors knows they are "imperfect", in the sense of not making a complete gamut. We don't need a different definition to make that clear. If the color triangles don't make it clear, add something. No need to redefine. As for 3 primaries for humans, yes, that's the usual, because humans are trichromatic. That doesn't mean one can't find some advantages from using more than 3 sometimes. Dicklyon (talk) 04:09, 6 January 2016 (UTC)

Some of your assertions would pull the discussion far out of scope, e.g. addressing "why do art stores stock so many pigments?" would require you to understand who art stores sell to . The handprint site is exceptionally detailed, well referenced and correct. You need only look at this talk page to understand that there is a great deal of confusion about primary colors, look at this widely sold color wheel (read the definition of "primary colors"): http://www.dickblick.com/products/artists-color-wheel/?clickTracking=true&wmcp=pla&wmcid=items&wmckw=04951-0000&gclid=COv_pqGMlsoCFUVgfgod-BUErA#photos. Your statement about 3 primary colors doesn't resolve why the color spaces identified in this article use more than 3, this can be resolved by opening with the idea that there are no 3 colors of light or pigments that are "ideal" primaries. Explaining that "ideal" primary colors don't exist should be one of the first thing this article does, that is define primary colors against common (incorrect) notions like we see on the color wheel I've linked to. Maneesh (talk) 21:15, 6 January 2016 (UTC)

I think writing an article by centering around a supposed common misconception is backwards. And color spaces are 3-dimensional because human vision is trichromatic. That is not at all at odds with there being no perfect real primaries, nor with the use of more than 3 primaries sometimes. Dicklyon (talk) 21:59, 6 January 2016 (UTC)

You need only look at essentially any "how to paint" book to understand that the misconception is not merely "supposed", it is endemic. I'm only asking you to resolve the sentence that claims that trichromacy is reason why three colors are used in "human applications" (what others are there? you still haven't mentioned) followed by the description of NCS that has 6 primary colors (is NCS an exception?) and CMYK which has 4. You don't find this confusing at all? Why does trichromacy not explain the numbers of primaries in these cases (2 out of 5 or so color models mentioned on this page)? Color appearance is certainly many dimensional (color contrasts, adaptation); worth mentioning since color appearance is what we generally mean by perception (not just controlled matching / discrimination) and is a function of a many, many neurons not just 3 opsins (or 4 if you consider rod effects at mesopic light levels). In summary the article should give the overall picture that there are no real primaries (wavelength or pigments), just arbitrary ones used in specific applications (giving examples of small sets of light/pigments used in different applications). It should also allude to the tremendous complexity of color appearance, making it clear that the actual number of colors a set of primaries can mix can be difficult to define very precisely. Would be great if other would chime in on this discussion.Maneesh (talk) 22:41, 6 January 2016 (UTC)

The opponent or NCS colors are not primaries as defined in the article, which states that they might be called "psychological primary colors", because any other color could be described in terms of some combination of these. OK, they might be called that. But it only takes 3 to span a 3D gamut of human-distinguishable colors (the three pairs of opponent or NCS colors define the 3 axes in a 3D color space). The K in CMYK is also not a primary, just an optimization of how to get a sharper deeper black. It also "might be called" a primary if you like, but the article does not say it is one. Where is the problem with this treatment? Dicklyon (talk) 01:12, 7 January 2016 (UTC)

Third opinion

A third opinion has been requested. Since the discussion above is lengthy, will one of the editors please state concisely (in one paragraph) what the question is? Robert McClenon (talk) 02:01, 7 January 2016 (UTC)ل

It is my impression that User:Maneesh wants to rewrite the article, or at least the lede, around a description of a wrong idea about what primary colors are, to debunk a (supposedly) widely held misconception. My opinion is that the current lede, which simply states what primary colors are, without reference to that misconception, is the more normal way to go. I see no errors or great problems with the article, but I would not object to putting some of his concepts into a section (not in the lede) if they are not already in there. His proposed prepended lede is shown in this version. When he prepended that lede he also removed a paragraph that I agree is highly questionable (in this edit); that came back with my revert, and we can work on either removing or correcting it. I invite Maneesh to state this case briefly before a third opinion is rendered. Dicklyon (talk) 03:31, 7 January 2016 (UTC)

Not crazy about exactly what I wrote originally, but something much more aligned with the spirit of the handprint article (http://www.handprint.com/HP/WCL/color6.html) that debunks the common notion and alludes to the complexity of color appearance/perception. Something like:

Primary colors refer to a small set (generally, but not necessarily, three or four) pigments or lights that can be "mixed" to produce a larger gamut of colors. The specific pigments or wavelengths of light that are selected as primaries depend on the color application (e.g., computer monitor, inkjet printing or oil painting) and associated constraints (materials, energy cost etc.). Mixing of colors can happen in a physical sense (mixing or overlaying pigments) or in a optical sense (e.g., at typical normal viewing distances the eye cannot resolve the colored elements of small individual pixels, so the red green and blue light mix "in the eye" to produce a large gamut of colors).

It is commonly assumed there is a precise universal triad of primary colors that is red, green and either yellow or blue (e.g. consider the text of this color wheel widely sold in art stores in the USA and many artistic textbooks), but this is incorrect. There are no small set of "ideal" primary pigments or lights that can mix all colors seen by the human eye (as measured in color matching experiments), but common sets of primaries used in computer monitors/televisions, color printers and paints have gamuts that are sufficient for most purposes. Other sources suggest that the three primary colors correspond to the singular activations of each of the three types of opsin that are the basis for color perception, that is, every color that can be perceived is a result of stimulating the three opsin types in some relative combination. In principle the colors that could be seen by stimulating each of the three opsin types singularly would be a sound notion of primary color but since no wavelength of light singularly activates one type of opsin, the psychophysical response to seeing such a "true" primary stimulation is unknown. In summary there are no "true" primary colors, they are either imperfect (can mix a subset of all possible colors the eye can see) or imaginary (there is no way we can physically see them). Color is ultimately a complicated psychophysical response to light hitting the retina that depends not just on the light of the region of color being viewed, but on the colors of neighboring region, perceived illuminance etc. thus it is often not meaningful to ask what primary color sets produce "better" gamuts without careful, specific perceptual testing.

How do third opinions feel about this general feel for the article as opposed to the current version. I feel it is more precise and correct. Important problems with existing text include:

• painters often don't paint with just three pigments, not always RYB (many suggest CMYK...see handprint).
• hard to understand what, if any, difference there is between "primary color" and "psychological primary", the obvious question is if primary-ness is related to the number of opsin types, why are there 6 primaries in NCS?
• Same for the 4 primaries in CMYK. I think it is unfair to suggest that the K isn't primary, you typically can't mix as black a black with CMY pigments.
• mention of "negative amounts of color" doesn't have any real meaning outside of transformations applied to data from perceptual experiments

Maneesh (talk) 07:38, 7 January 2016 (UTC)

Response to third opinion request:

I like the first paragraph that Maneesh has written as a summary of the current lede, but I don't like the second, even ignoring all the statements that scream out for a citation. Fundamentally, I don't think that belongs in the lede, but instead as a subsection devoted to criticisms. I think the lede as it is does a good job of summarising the rest of the page, and agree with Dicklyon. The lede should not be used as a place to make a substantive argument. Smith(talk) 14:54, 7 January 2016 (UTC)

I agree that the newly proposed first paragraph is better than the one we were arguing over, but it still has the fatal flaw of starting as "Primary colors refers to..." rather than "Primary colors are...", and it mentions "larger gamut" with nothing to compare to. This is not better than the existing "Primary colors are sets of colors that can be combined to make a useful range of colors." For a subsection, "criticisms" isn't exactly where this is going, but maybe "Misconceptions" would be OK; it should provide authoritative secondary sources to back up any statement that a particular misconception is common, or something like that; it's not for us to judge, and I don't think a random web site is adequate either. Dicklyon (talk) 15:54, 7 January 2016 (UTC)

I've never thought carefully about "refers to" vs. "is/are"...a quick look suggests that "refers to" is useful when the term is divergent (primary colors can refer to lights or pigment...and then different subsets of hues). "Refers to" seems correct to me. We can replace "larger gamut" with "mixed to produce much larger sets of colors with additional hues". Moving a reworked version of the second para (I didn't include references in my draft, but all of my assertions are easily supported) to a "misconceptions" section makes sense to me. To be clear, handprint isn't a "random website", it is one of the best sources of color science information on the web (with some, quite insightful, original content); I can't think of any assertions on handprint that wouldn't be easily supported by textbooks. The point about imaginary primaries is not controversial.Maneesh (talk) 18:21, 7 January 2016 (UTC)

Maneesh, see Use–mention distinction. When you say "Topic refers to" you are mentioning the name of the topic. Better to just use it. Yes, what you say about imaginary primaries is not controversial. Some of handprint's position, however, are, and might be hard to find good book sources for. Dicklyon (talk) 21:00, 7 January 2016 (UTC)

In particular, ...actual lights or paints that cannot mix all possible colors, which means they aren't really "primary" is assuming an incorrect and nonstandard strawman definition of "primary". Let's don't do that. Dicklyon (talk) 21:09, 7 January 2016 (UTC)

Why do you feel the [edit: assumed] handprint definition is a nonstandard definition of primary? Consider Oxford dictionary, google. As yet another example of the endemic misconception, look at Merriam-Webster.Maneesh (talk) 21:31, 7 January 2016 (UTC)

I'm sorry, it seems I wasn't clear enough. What I meant was that despite liking that first paragraph, I prefer the lede as currently written over any change. I think it adequately summarises the page, and any further shortening would just lose information. Smith(talk) 22:08, 7 January 2016 (UTC)

Maneesh, OK, I'll grant you that "any of a group of colors from which all other colors can be obtained by mixing" is a screwed-up definition that appears exactly thus all over the web. But not in any book, except those that clarify that by "mixing" they mean allowing negative amounts, or mixing one or two primaries with the color to be matched, as in color-matching experiments with additive primaries. Most decent sources are quick to recognize the limited gamut that one gets by mixing primaries. If we stick with decent book sources, we'll be in better shape. Handprint is reacting to this bad web meme; WP should be written around what's right, not what's wrong. Dicklyon (talk) 03:19, 8 January 2016 (UTC)

Dicklyon, to be fair, it isn't fact that is yours to grant. This misconception didn't originate from the web nor is it limited to the web. This is an endemic misconception that can be found in many books on painting and children's art "education" materials. I recall confusion around the notion of primary colors as a child, perhaps we grew up in different school systems. Confusion about primary colors is the rule, especially amongst artists (from my experience). I hope it is clear that this misconception needs to be dealt with explicitly on this page. My last important quibble, for the lede at least, is the mention of "negative amounts of a color ". This is just confusing and shouldn't be there, look at the confusion around "primary" you can't just say "negative amounts of color" and expect people who aren't familiar with color science to understand.Maneesh (talk) 04:51, 8 January 2016 (UTC)

Maneesh, I understand your point. My issue is about how to present it. It's better to start with a valid definition, accepted in color science, than to start with a wrong definition and work on debunking it. Dicklyon (talk) 05:16, 8 January 2016 (UTC)

Did my best to implement what was discussed here. Provides the common definition of primary color (a set of pigmented media or lights) and the definition from the color science perspective in a digestible way. The misconceptions section leads into the biology section. I suggest that improvements be made from here, rather than reverting; if this sticks it will be worth making the rest of the article consistent with this flow. I'll leave this in other contributors hands from here.Maneesh (talk) 22:13, 8 January 2016 (UTC)

"Common misconceptions" section

This new section cannot remain as it stands, as it quotes no sources and appears to be "original research". Who has stated, in what reliable published sources, either that the definitions given are "misconceptions" or that the misconceptions are common?: Noyster (talk), 23:32, 8 January 2016 (UTC)

Please look just at the discussion just above with "This is an endemic misconception that can be found in many books on painting and children's art "education" materials.". The page, as it stands does quote Merriam Webster. I could select almost any popular painting book, I haven't added the right citations since I'm not quite sure how to show something is endemic.Maneesh (talk) 00:04, 9 January 2016 (UTC)

New try

Maneesh, "a small set (generally, but not necessarily, three or four) of pigmented media or lights, each of a single unique color" is still very confusing and wrongish. First, it's almost always three. Second, if you mean "lights or pigmented media", say it in that order to avoid the ambiguity of what "pigmented" refers to; but why "media"? Third, what does "single unique color" mean? I don't understand either "single" or "unique" here. And primary colors are not really lights or media; they are colors. That's as far as I got. We can review proposals here if you prefer not getting reverted on the article. Dicklyon (talk) 05:09, 9 January 2016 (UTC)

And to claim that "Many books" use a wrong definition, you really need a secondary source that's commenting on those many sources; we don't synthesize, per WP:SYNTH. Dicklyon (talk) 05:13, 9 January 2016 (UTC)

Actually, the History section already covers the artists' RYB system. Maybe we can just add a bit there instead of a few section? Dicklyon (talk) 05:16, 9 January 2016 (UTC)

The sentence you cite makes complete sense to me. It isn't "almost always three", CMYK is 4, NCS is 6, the section about Recent Developments talks about televisions with 4 or 5. Read a definitive summary from Fairchild. Key quotes illustrating the points that don't seem to be coming across:

...Red, yellow, and blue, of course! Sorry, not so fast. Even though many of us are taught that red, yellow, and blue are the primary colors in our school art classes, we have actually been misled a little bit.

Technically speaking, primary colors are defined as any set of three (or more) colors for which no one of the colors can be made by mixing any of the others from the set.

I can't imagine how I could bring these facts across more clearly.

His definitions are completely in line with mine, it is correct to refer to pigmented media or lights as colors. The small set of pigmented lights or lights you use to make a picture with paint or on a TV screen are the primary colors you used. Handprint does a good job breaking down material color vs. radiant color vs. visual color vs conceptual color here, but this is not standard AFAIK. Throughout wikipedia and most common language, we don't differentiate between what we precisely mean by "color", therefore the use is correct here.

Pigments can be distinct from dyes, but"pigmented" just means colored so "pigmented media" covers things like inks, colored filterpaper, colored solutions etc.

As for the endemic misconception, I'm sure Fairchild's quote above from "Why Is Color" should suffice.Maneesh (talk) 05:54, 9 January 2016 (UTC)

I don't understand what you're citing Mark Fairchild for. He says "Red, yellow, and blue are one common set of primary colors that are often used in painting pictures. However, they are not the only set of primary colors." This is fine, not at odds with anything in the article, not saying that everything we're taught is wrong. And please repair your links. Dicklyon (talk) 06:08, 9 January 2016 (UTC)

The Fairchild quote above is a synthesis supporting the idea that the misconception that red yellow and blue are the (only) primary colors is something that is commonly taught in schools (it does not provide evidence that the idea that "primary colors mix all colors" is a common misonception, but given what I've shown so far I think it will be easy to find). He goes to the trouble, in a kids book, to specify "three or more" in his definition of primary colors (which doesn't line up easily with linking the common notion of primary colors to trichromacy)Maneesh (talk) 06:44, 9 January 2016 (UTC)

Of course it does "line up". He says "any set of three (or more)" and that three comes from the 3D nature of human color perception; from trichromacy. Using more primary colors doesn't hurt; but it also doesn't help much, and is not so common. And he doesn't say we're taught that RYB are the "only" primary colors. That might happen sometimes, and if you find a source that says so that's worth a mention, but is hardly central to explaining primary colors. Dicklyon (talk) 07:52, 9 January 2016 (UTC)

I am well aware of color science, but when the word "primary" is used as small set of pigments it is very common to use 4 pigments look here. You are making unsubstantiated claims about the degree to which added primaries help (I know that some painters like using 6 hues...how do you know how much using those hues helps or hinders their work?). 4 is a common number in paint and mixing that concept with trichromacy (3 doesn't line up with 4) is very confusing. It seems I haven't been able to help improve the article much, I'll leave it to other users to try.Maneesh (talk) 09:02, 9 January 2016 (UTC)

The four-color set of paints recommended by handprint.com is a great idea. But their term "artists' primaries" is much more widely used for the three-primary RYB set. Introduce the concept of using more than 3 colors as an enhancement to expand the gamut; at the start, it's just a distraction from the basic idea of 3 (or more) primaries being able to create a 3D gamut of colors. Then the "confusing" relation can be explained, after the basics. Dicklyon (talk) 18:12, 9 January 2016 (UTC)

Article Quality Remains Poor

The article is quickly being turned into drivel with recent edits. Instead of converging to a critical understanding of primary colors using correct vocabulary from color science there are now:

• references to subjective/marketing-based color names (cyan is, apparently, a bright "aqua"...never seen anyone in vision science call cyan something other than cyan). Using sources like Web colors isn't terribly salient to understanding the concept of primary colors.
• unsupported statements about what painters do (I suspect written by people who don't actually paint...you can't do much on a palette without white).
• incorrect statements about CMYK vs. RGB (without even using a chromaticity diagram).
• incorrect equivocation between color and chromaticity.
• incorrect equivocation of number of primary pigments and number of photoreceptors.

There are many other incorrect statements in the article. The imbalance between how such edits are tolerated while sound edits based on color science seem to be quickly reverted by editors is troubling.Maneesh (talk) 20:28, 20 September 2017 (UTC)

I'm not familiar with the past reverts but I think we can use your comments to improve what's there now. I agree that the article needs to draw much more material from color science.

• references to "aqua" have been removed
• the sentence about painters helps avoid giving the false impression that those who consider RYB to be the subtractive primaries are therefore limited to an RYB gamut. The phrase "three chromatic pigments" was meant to exclude white or black as these are achromatic. Can you propose an alternative or do you favor simply dropping it?
• please be more specific about the incorrect statements and incorrect equivocations, preferably providing correct statements (or pointing to correct but reverted statements in earlier revisions?)

QuoJar (talk) 21:38, 20 September 2017 (UTC)

The use of "fuschia" is no better. Primary colors are always arbitrary, that should be made clear near the top of the article.

..but some sets of primaries can yield a far wider range (gamut) than others...

This is just vacuous.

Although red, yellow, and blue (RYB) are a well-known traditional set of subtractive primaries, blue can be produced by mixing cyan with magenta, while red can be produced by mixing magenta with yellow, and therefore CMY[K] can produce most of the colors that RYB can, whereas RYB primaries alone cannot produce some intense greens and oranges, nor colors near cyan or magenta themselves for example. However, in painting for example, artists often use more than three chromatic pigments and so are not limited to a colorspace such as RYB even if they think of these three as as their primary colors in an abstract or conceptual sense, for example when constructing a color wheel.

What does the K in brackets mean? Does the paragraph mean chromaticities where it says color? Which pigments? What does "most of the colors" mean? Wouldn't it make sense to support this claim with a number? What styles of paintings is the paragraph referring to (nice and notable paintings can be made with one or two pigments)?

Good additive primaries are often darker than their related subtractive counterparts (for example at least blue darker than cyan and magenta, red darker than magenta and yellow, and green darker than cyan and yellow), because they will be able to produce brighter colors (including the subtractive primaries themselves) when they sum. If subtractive primaries were that dark, they would not be able to produce such bright colors, except by allowing more of an illuminating light or white surface to show through and thus losing saturation as the color becomes more pastel.

This makes no sense. Dark colors are made with dim light, how is that "good" or bad? I suspect whoever wrote this doesn't understand the difference between chromaticity and color. Red, blue magenta etc. are all hues, colors with those hues can be darker or lighter than each other.

For real (dye/pigment/light source) primary colors (as opposed to abstract like XYZ), the number of color receptors becomes the minimum number of primaries needed in order to produce more than a tiny sliver of the perceptible color gamut. Thus for trichromats like humans, we use three (or more) primaries for most general purposes. Two primaries would be unable to produce even some of the most common among the named colors. Adding a reasonable choice of third primary can drastically increase the available gamut, while adding a fourth or fifth may increase the gamut but typically not by as much.

Again, nonsensical. Are we talking about paints or lights? Which "purposes" is the article referring to? Many pictures we see are enjoyed in greyscale. Isn't the reason XYZ has three dimensions is because it is a linear transform of three dimensional LMS space? How big is "drastic"?

Demonstrating improved spectral discrimination in any animal is difficult due to the underlying neural complexity of the process.

We understand color despite "neural complexity" because is relatively easy to get humans to do color matching and respond verbally. Doing this with non-human animals is generally difficult since they don't talk. Neural complexity isn't the main reason it is difficult to understand what a mantis shrimp sees, it's difficult because mantis shrimp don't easily obey commands and describes what it sees.

The entire section on "additive/subtractive color mixing" is fundamentally incorrect. Paints don't mix purely subtractively, read [1].Maneesh (talk) 00:39, 21 September 2017 (UTC)

Is it fair to say that many sets of RGB additive primaries can produce a wider gamut than, say, certain sets of barely-distinguishable shades of red as primaries? (As an extreme case, two reds, or three reds aligned on a short, straight, line segment on a chromaticity diagram so that one of them can be produced by the other two.) I agree that either set could potentially be used as valid primaries, but if the gamut of additive RGB includes the gamut of another set of three primaries as an area (or line rather than area in the extreme case) enclosed in its interior, in what sense is it *not* true that RGB has a (much) wider gamut that this "R1,R2,R3"? Thus, "some sets of primaries can yield a far wider range (gamut) than others" -- or is it merely the word "far" that you object to? Certainly one could show in even a roughly-perceptually-uniform space that the gamut of some particular set of primaries span a tiny fraction of the perceptual volume of some other particular set of (say, RGB) primaries.

QuoJar (talk) 13:52, 21 September 2017 (UTC)

How many nearly neutral R, G, B primaries are there (super tiny triangles around the whitepoint) and how many barely distinguishable reds are there with merely tiny gamuts? What is the ratio of those two numbers?Maneesh (talk) 16:11, 21 September 2017 (UTC)

I didn't say that *any* three RGB primaries (such as three clustered near the whitepoint) provide a wider gamut than some set of barely-distinguishable reds. Just that there exist *some* set(s) of RGB primaries that have a demonstrably-larger gamut than *some* sets of three barely distinguishable reds (because the latter gamut is completely inside the former), which is enough to demonstrate that indeed *some* sets of primaries can yield a (far) wider gamut than *some* other sets, which is what is being claimed in the article. No specific numeric claims are being made (and not all valid statements in an article need to be numeric) although I would of course be happy to see any such available numerically-quantified and notable facts added to the article. QuoJar (talk) 16:47, 21 September 2017 (UTC)

The sentence in the article that I think we are discussing is:

Any small set of realizable primary colors are "imperfect" in that they cannot generate all perceptible colors, but some sets of primaries can yield a far wider range (gamut) than others.

It is a vacuous statement. A chromaticity gamut is just a convex shape in the 2D xyY plane. Some convex shapes are larger than others, so what? Pick a set of 3 primaries, I can add one more primary to get a "far wider range". How is this sentence informative? Perhaps you mean to say that some sets of 3 primaries have larger gamuts than some other sets of 3 primaries, that is of course true, but a simple consequence of geometry. I'm not sure why this needs to be stated explicitly to the reader.Maneesh (talk) 18:03, 21 September 2017 (UTC)

The fact that this true statement can be derived from the geometry of color solids, or of two-dimensional projections of them such as an xy chromaticity plane, certainly does not make the statement vacuous, self-evident, redundant, or obvious to most readers of Primary color (or of its first paragraph). Is it important that readers not be given this information, and if so, why? QuoJar (talk) 13:48, 22 September 2017 (UTC)

It's vacuous because it is self evident looking at an xyY diagram. If we were to say "some animals are bigger than others" while looking at a picture book of animals it wouldn't mean much. A statement about relative sizes would presumably lead to some sort of salient information about animals (say, trophic levels). What I presume is intended is an explanation as to why R,G and B hues are used in place of, say, orange, cyan and magenta for most applications. Something like "many color applications, like electronic displays, use three additive primaries for realistic image reproduction. The hues of those primaries are invariably red green and blue since those hues inscribe the largest triangular chromaticity gamuts within the spectral locus" would make more sense.Maneesh (talk) 20:00, 22 September 2017 (UTC)

The statement in question is in the first paragraph of the lead of the article -- the typical reader does not know about triangular chromaticity gamuts or spectral locii, so this will make no sense at all to them. Statements at the beginning of the lead obviously can't go into that level of detail or technicality, but I do see what you're getting at, at least maybe moving down or elaborating a little later. But not even the first paragraph should give the impression that the choice of primararies is arbitrary, as this would falsely suggest that there are no reasons for choosing one set of primaries over others. QuoJar (talk) 22:45, 22 September 2017 (UTC)

Without studying these arguments in much depth, I've made what I think are improvements to the lead. Comments welcome. QuoJar, I'd advise you as I did Maneesh, to proceed more slowly and incrementally. It's much harder to discuss and converge on such widespread changes. So I did one edit, just in the lead, to gauge a few things, before going further. I believe we all know enough about color science to contribute here, and that we can resolve our differences in how best to express, or what to emphasize. Certainly we don't want to leave anything that is an error. And proposing changes is a lot better than just saying it's not great. Dicklyon (talk) 23:00, 22 September 2017 (UTC)

And I did a second edit to add the most common example of abstract mathematical primaries to the lead. Since we started with three types, having all three in the lead this way seems like a good idea, and addresses one of Maneesh's points. Dicklyon (talk) 23:11, 22 September 2017 (UTC)

Gotcha. I tweaked that last paragraph of lead (abstract colorspace) -- feel free to undo or change it further if you don't think it's an improvement. But really I'm not certain we need to go into explaining abstract colorspaces to that level in the lead itself, or at the very least I don't think we need to specifically describe CIE XYZ this early. We mention abstract colorspace/primaries in the first paragraph just for completeness, but it is a somewhat "advanced" topic for someone who might only read the lead and might be happy just to know what an "ordinary" colorspace is. QuoJar (talk) 19:49, 23 September 2017 (UTC)

is an artist's "palette of colors" similar to a color scientist's "set of primaries"?

In the sense of color science, a set of primaries is seen as something like

some set of colors, which can be combined to generate a corresponding gamut

But isn't this roughly what an artist calls their palette of colors (perhaps allowing me to fudge the achromatics like black and white for the moment), without *necessarily* needing to use a term like primary (or secondary)?

An artist encounters terms like primary and secondary and tertiary color in the quite-different color theory sense, for example appearing on a color wheel, which may give them some conceptual framework to help them figure out how to choose or mix the colors of their palette, but might not rigorously describe or predict the gamut of precise colors that can actually be generated from other colors (even if it *claims* that it can describe or predict such).

So maybe we should mention that there are these two meanings of "primary color", or maybe have a disambiguation page that directs them here for one sense and to the color theory article (or to a new Primary color (color wheel) or something like that) as opposed to the present article?
QuoJar (talk) 21:07, 23 September 2017 (UTC)

I'm not aware of any sources that treat the color theory and artist interpretations as different meanings. Dicklyon (talk) 18:47, 24 September 2017 (UTC)

I'm trying to distinguish between color science, for example colorimetry, versus the conceptions found in the visual arts, as exemplified in the articles color wheel and color theory. Looking at the latter article,

In the visual arts, color theory or colour theory is a body of practical guidance to color mixing and the visual effects of a specific color combination. There are also definitions (or categories) of colors based on the color wheel: primary color, secondary color and tertiary color. ..... From there it developed as an independent artistic tradition with only superficial reference to colorimetry and vision science.

color wheel begins

A color wheel or colour circle is an abstract illustrative organization of color hues around a circle, which shows the relationships between primary colors, secondary colors, tertiary colors etc.

Whereas colorimetry begins:

Colorimetry is "the science and technology used to quantify and describe physically the human color perception.

These do not sound the same to me, and it is especially implausible that colorimetry would accept the discussion of primary, secondary, etc., colors in color wheel/color theory or vice-versa.

So again, isn't the the term "primary color" being used differently in the visual arts conceptions of color wheel and color theory, as opposed to "some set of colors, which can be combined to generate a corresponding gamut"?

I don't think we can pretend that the term is being used in exactly the same way, or that one way is invalid and not notable enough to appear in wikipedia because it is not as rigorous or scientific as the way in which color science or colorimetry is using the term.

Anyway I think the concept from the visual arts that most closely resembles "some set of colors (any number of them), which can be combined to generate a corresponding color gamut" would not be the three primary colors that appear on a traditional color wheel, but rather would be an artist's palette of colors, which are quite literally the colors that are being combined to generate some range of possible colors. (There is some subtlety here with respect to colors versus hues but I think this is secondary, no pun intended.)

QuoJar (talk) 20:46, 24 September 2017 (UTC)

Sources like this[2] don't seem to have trouble synthesizing the notion of primary colors in terms of practical painting and vision science. Perhaps it is no surprise since these two domains are obviously related by the common perception of color. Disambiguating wouldn't make sense to me. I agree that information on the topic is notoriously inconsistent and untrustworthy making it difficult to come any sort of understanding. I also understand that there is a place to *describe* archaic (and generally incorrect) ideas about primary colors, but this needs to be clearly contrasted against a modern, correct and consistent perspective that involves what we understand about color vision. Maneesh (talk) 23:58, 24 September 2017 (UTC)

I agree, the artists and scientists are not using difference concepts, though their color boundaries and names do confuse things a bit. The typical artist's blue is on the green side of blue, while the typical additive primary blue is on the violet side. The typical artist's red is a bit toward magenta, while the additive primary red is a bit toward orange. They're trying to do the same things, which is to make a bunch of colors from a few; the artists describe it more experientially by how their paints and pigments combine, while color scientists try to explain it mathematically using spectra and sensitivity curves. Sometimes the definitions do get screwed up, like by sources that claim a primary color is one that can't be made by mixing other colors – that's just lame. See color wheel for many historical attempts to organize colors by boundaries, names, etc. And things like RYB are better characterized as traditional rather than archaic, since it's still widely used. Dicklyon (talk) 02:17, 25 September 2017 (UTC)

To Dicklyon, I understand that we are discussing informally here but a problem with the article and your paragraph is that the term "artist" is used without any qualification. I have no idea what a "typical artist's red/blue" is. What kind of artist? A red pigment that artists use might be an iron oxide or cadmium sulfide or pyrrole-based. A blue pigment could be ivory black, ultramarine or cerulean. The precise mass tones of paints made with those pigments will vary considerably. Most artists that I paint with understand hue, value and chroma and will use palettes like "Zorn's" (knowing that is was not his invention) but are well aware that the primaries they've selected are arbitrary. Who is using RYB and for what purpose other than to sell color wheels with incorrect information?Maneesh (talk) 07:22, 25 September 2017 (UTC)

To QuoJar, I don't find it "implausible that colorimetry would accept the discussion of primary, secondary, etc., colors in color wheel/color theory or vice-versa". Maybe you can provide an example or two of a concept in one of those areas that's not accepted in another. Dicklyon (talk) 02:23, 25 September 2017 (UTC)

In a color wheel, there are always exactly three primaries, almost always chosen from the blue-cyan range, the red-magenta range, and the yellow-green range, in order to span a large range of colors and their complements (even if the primaries can't literally be mixed to generate all the colors of interest); [added as an edit:] if the artist is using eight colors of paint on their palette these aren't considered to be primaries even though they are in fact the colors that are to be combined to form the gamut available to the artist. In the present article there could potentially be any number of primaries, of literally any colors whatsoever, some of which might happen to to be similar to the primaries on a color wheel, while others might happen to be similar to secondaries, tertiaries, or just arbitrary positions (or no position at all?) on a color wheel. A color wheel doesn't change shape depending on what primaries are used. In the present article, N additive primaries would represent the N vertices of an N-gon shaped gamut in a chromaticity diagram.

What evidence do you have that when eight colors paints are used they aren't considered to be primaries? Here is a thread where self-professed artists are talking about at least 6 primaries[3] and a class where "8 primaries" are being described[4]. I've never heard such a "rule of eight".Maneesh (talk) 18:15, 25 September 2017 (UTC)

Okay, but my main point is that there exist at least two meanings of "primary color" (not that all professional artists are compelled to use one of these definitions exclusively). One meaning is often used in a traditional color theory and color wheel, and whose distinction from secondary and tertiary colors is emphasized, and which almost always has three (non-neutral) primaries (and perhaps usually can at least produce all hue angles, as opposed to all saturations or all values). The other meaning is that of potentialy any combination of any number of colors or colorants whatsoever (eight and 11 were merely my arbitrary examples, not rules) that are combined to produce other colors forming a gamut. (I think this distinction is consistent with some distinctions made in the treatment on Handprint.) If only one of these two meanings is considered valid, or only one is the topic of the present article, then let's say so explicitly. From the current content of this article I would assume it is the second definition -- the one in which all the (non-neutral?) paints on your palette are termed primaries and none are instead described as secondaries etc. But in this case, let's start by removing the scare-quotes from "primaries" in the Recent Developments section because, e.g. LED displays with five or six "primaries" would meet the full meaning of primary and need not be quoted, correct? QuoJar (talk) 19:07, 25 September 2017 (UTC)

(BTW, I'm not even sure the treatments in your links are following "our" definition of primaries consistently (as being the entire set of colors used to mix other colors of the gamut). In the first one they define "Primary Colors" as "Three colors when mixed in equal or unequal amounts can produce a variety of colors. Traditional primary colors are: Red, Yellow, Blue" and in the second one they talk about a "12-color primary/secondary/tertiary palette" rather than merely a palette of about 10 or 12 primaries, just for example. QuoJar (talk) 19:20, 25 September 2017 (UTC))

I'm sure the last two links I provided are not consistent (almost no one is on primary colors), they were the simply the first two I got from google. They still show how people who claim to paint don't mind using "large" sets of primaries and calling them as such. Handprint[5], the authoritative site I keep referring to, puts primaries in scare quotes throughout its pages. I don't think the description of Sharp's four color TV should really be kept in the Recent Developments section. It appears like a defunct product to me, I can't find any news about it past 2010.Maneesh (talk) 20:41, 25 September 2017 (UTC)

As I mentioned, Handprint does acknowledge and try to explain multiple meanings or usages of the term "primary color", and actually uses the term in more than one of these ways as appropriate, which may be why they use scare quotes when using the term without sufficient qualification as to which precise definition. In contrast, the present article focuses mainly on the "any set of colors producing a gamut" meaning, and so, unlike Handprint, never even mentions that there might be other ways of using the term, in the context of which things like color wheels, secondary colors, or complementary colors might become quite important. Do you agree that the term "primary color" can or sometimes is used in such a way that three of the colorants being used are described as primary while three others are described as secondary, as Handprint does? For example at one point they describe a six-colorant set (excluding black and white) consisting of:

• primary light yellow : benzimidazolone yellow (PY154) or hansa yellow medium (PY97)
• secondary red orange : pyrrole orange (PO73) or cadmium scarlet (PR108)
• primary magenta : quinacridone magenta (PR122) or quinacridone rose ("permanent rose", PV19)
• secondary blue violet : ultramarine blue (PB29) or cobalt blue deep (PB73)
• primary cyan : phthalocyanine blue GS (PB15) or phthalocyanine cyan (PB17)
• secondary blue green : phthalocyanine green BS (PG7) or phthalocyanine green YS (PG36).

Where might this type of usage of the term "primary" fit into the present article, or should any usage of "primary" that describes one of the colors producing a gamut as "secondary" be summarily excluded from present article? QuoJar (talk) 22:24, 25 September 2017 (UTC)

I don't understand what use of primary you are referring to. This article should cover imaginary primaries like XYZ (it should have LMS as well), sets of pigmented media (paints, inks) and lights. These are the same sets of primaries that are described in handprint. Describing a pigment's position in a palette as "secondary" or "primary" in the way handprint is doing from where you quote[6] isn't really meaningful; you can see all the text that surrounds your quote emphasizes that the pigment choices are "arbitrary" and the notion of a primary triad is "obsolete". I don't think this specific way handprint is using the term merits any sort addressing in the article.Maneesh (talk) 00:44, 26 September 2017 (UTC)

The concepts of secondary colors, the traditional color wheel, and traditional color theory have articles in WP because they are notable either for their current or past usage, regardless of whether they might be considered "obsolete" and "meaningless" within modern color science or colorimetry. Do you propose that those articles be marked for deletion because they don't fit well into modern color science? If secondary color is a valid topic, then certainly its "parent" concept of primary colors in the sense from which secondaries are defined in a color wheel, which it fundamentally depends on, must be valid as a topic, whether that topic should be addressed in the present article or merely mentioned as something that is *not* covered by the current article, with a link to an article where it is described. Maybe we need to begin this article with one of those "This article is about the concept of a primary in color science and colorimetry. For the traditional color wheel concept of a primary in the visual arts, on which secondary colors are based, see color theory and color wheel."QuoJar (talk) 02:10, 26 September 2017 (UTC)

I'd rather see traditional primary colors treated as a section (maybe with a main link) than as something different. They're not different concepts. The current section Primary color#Limited palettes in visual art seems to be written to deny the traditional RYB primaries rather than to describe them, so that should be fixed. Dicklyon (talk) 05:44, 26 September 2017 (UTC)

They certainly aren't the same concept, as key parts of each is clearly excluded or deprecated by the other, and burying these mutually-exclusive aspects of the two concepts (concept variations if you prefer) because some of us would like them to be thought of as the same concept is doing a disservice to the casual reader.
Also do we all agree that an artist's palette of paints (maybe or maybe not excluding neutrals black & white) can form a set of primaries in the sense of colorimetry? If so then we should point this out in the lead, as many casual readers will be more familiar with a term like "palette of paints" or pigments than with calling them only a [type of] set of [realizable subtractive] primaries.QuoJar (talk) 13:03, 26 September 2017 (UTC)

"Pigments" is correct. I can mix colors with ball point pens, colored pencils, wells of water color on the canvas (without any sort of "palette"). [added as an edit:] "pigmented media" is even more correct since a "pigment" is a solid and a "dye" is a solution.Maneesh (talk) 18:16, 26 September 2017 (UTC)

Looks like Maneesh had trashed the RYB section in this edit, and added the replacement crap here. This should be fixed. He's too intent on denying the traditional primaries or proving them inferior or bogus or archaic or something. That's not a useful approach. Dicklyon (talk) 05:55, 26 September 2017 (UTC)

No one is in denial, the idea that RYB doesn't really mean anything is an underlying message of handprint. No one is saying using hues of red, yellow and blue is "inferior". I have no idea what the claims of "traditional" primaries are and you've claimed above that RYB is in widespread use and I've only politely asked you for evidence. Who is using it and for what? I know people that use red, yellow and blue pigments (along with white) for realism. Does that mean they are using RYB?Maneesh (talk) 06:20, 26 September 2017 (UTC)

We can't remove notable and widely-discussed concepts or points of view from wikipedia just because there are some authors (even a lot of reliable sources or an entire field) who view those concepts as meaningless in the context of a "superior" or more "more modern" approach, regardless of whether we as editors agree with this. The fact that there is much arbitrariness in the choice of secondaries or primaries, or that when an artist uses some shade of RYB it is not clear whether these are being used "as" traditional triad primaries does not mean that the traditional colorwheel concept of secondaries and primaries should be banished from wikipedia. So long as the articles secondary color, color wheel, and color theory are valid topics, then the concept of primary as they use it is a valid topic (or one of the major aspects of one), and it is perplexing to the casual reader (who is likely more familiar with that concept of primary than the one emphasized here) when those other articles discuss or refer to a conception of primary color which will, at best, be buried in a subsection of primary color. QuoJar (talk) 13:03, 26 September 2017 (UTC)

I've explicitly said that I understand that RYB should be described in terms of history. Again, handprint goes into great detail of this type of history like here[7] and here[8] in far greater rigor than what is on WP. By all means, describe the history using the primary sources that handprint cites and anything else you can find. You'll see that RYB, as best it can be defined, wasn't a tool used by painters/artists it was a conception of people who were thinking about what color was (proto-color scientists). Claims about how systems like RYB are used today should be backed up by evidence, I've asked for it several times here and none has been produced. I've never met a painter who says they use "RYB", most painters skilled in realism understand a few simple heuristics (and are aware they are mere heuristics) of things like opposing hues and are aware of things like "substance uncertainty". The problem isn't that RYB is evil or that it even exists, the problem is that when claims are made about its use when there is no evidence for that use. In this light it doesn't make sense to describe the subtle difference between primaries as sets of physical pigments and the (doing my best here) "concepts" of R Y and B in RYB. I can see that RYB is making a claim about CMY which makes no sense ( a high chroma RYB yields a larger gamut than a low chroma CMY).Maneesh (talk) 18:16, 26 September 2017 (UTC)

I think you're arguing about RYB with Dicklyon, not me. I'm not as concerned about which specific primaries are being used in somebody's color theory/wheel, but rather the actual meaning of the term "primary color" described by this article. Article currently only talks about the sense in which it is any set of colors that combine to form a gamut, as opposed to the traditional meaning of primaries where it is three (chromatic) primaries (what ever they may be, which could vary according to your color theory!) along with three secondaries, etc., as part of a color wheel or traditional color theory that might provide (non-rigorous!) guidance on paint selection and mixing. In the latter concept of primary color, the artist's palette of pigments might (as Handprint does at one point) or might not have been constructed such that some of them approximated some "primaries" and others were chosen to approximate some corresponding "secondaries" (but perhaps with higher chroma or saturation, at least for some values or lightnesses, than the secondaries literally made by mixing the primaries, as a particular strategy for increasing the gamut).
These are clearly two different variants of the concept of "primary color" since both the primaries and secondaries of the traditional meaning are all called primaries in the "palette of colorants or lights generating the gamut" meaning! We can't simply choose to ignore (or just bury down in a subsection with no meantion in the lead or first section) the meaning of primary that relates to a concept called secondary color.QuoJar (talk) 19:57, 26 September 2017 (UTC)

I believe you are saying that we have color wheels (true) and those wheels are models that are used to make inferences on mixing paints (true). Certain hues on those wheels are labeled as "primary" or "secondary" (true). In general the model asks us to map physical pigments to the labels on the color wheel. In this context the correspondence between the physical paint and the labelled hue on the color wheel are so direct it seems odd to spend much time differentiating between the pigment and the labelled hue on a color wheel as different meanings of "primary". In the case of handprint's use of "primary" in terms of watercolors, I think it is explained well here[9] under "Why Use Primary Colors?".Maneesh (talk) 22:18, 26 September 2017 (UTC)

The point isn't whether you or I think it's odd -- the point is that there are notable uses of the term primary color that always have three of them (and three secondaries...) regardless of how many pigments you decide to use or whether they match these primaries or secondaries at all. The color theory does not *require* that you use any particular pigments or that these match the primaries or secondaries. What ever arbitrary pigments you use, you might approximately locate them on the color wheel and and then use this as a guide to mixing them. There need not be such a direct mapping, although there could be. This is simply a widely-used meaning of "primary color" as opposed to ours which would more likely be used to enumerate all the pigments being used and call these the "primary colors" which generates our gamut. This concept of a color theory or wheel having three primaries (regardless of whether ryb or cmy or something similar) and three secondaries (etc.) is a notable concept, period, and it is not the same as the concept we are describing in this article at the moment. Do you agree that we should mention this as one of the common meanings of "primary color", however odd or unscientific we might find it? QuoJar (talk) 22:46, 26 September 2017 (UTC)

I don't think the labels of hues as primary colors on various color wheels is a notable thing. Look at Color wheel, none of the images of color wheels I can see have hues labelled as "primary". If this is a notable concept, why isn't it seen in prominent historical examples of color wheels? We can see it in writing about obsolete color theory, but the concept of primary would be specific to each obsolete theory.Maneesh (talk) 23:41, 26 September 2017 (UTC)

Depends on what you mean by notable, I suppose. Plenty of color wheels do label primaries, like these traditional RYB ones: [10], [11], [12], [13], [14], [15], [16], [17]. Sometimes you'll find them labelled in RGB or CMY systems, but that's less common than with traditional RYB color wheels, I think. Dicklyon (talk) 00:23, 27 September 2017 (UTC)

Right, but it's the concept I'm talking about, not the precise colors designated as primary. "3 Primary and 3 secondary...." colors is a notable concept and it's reasonable to expect it to be mentioned at the start along with other concepts of primary color in an article on primary color, no matter how much disdain we as editors might have for it. QuoJar (talk) 01:36, 27 September 2017 (UTC)

None of Dicklyon's links seem terribly notable. They include a men's style website, Lifehacker and handmade charts on pinterest. Surely if the context we are talking about is specific colors being labelled as "primary" and "secondary" on color wheels (which is the precise case we are talking about here, as has been laid out above by QuoJar), we should be able to find something more notable of authoritative. Quojar your latest message gets away from wheels and now just that "3 primary and 3 secondary" is notable. How do you support that statement? I've seen such statements before in a lot of confused thinking about color but nowhere notable. You can see clearly in my message below that handprint explicitly states there is nothing special about "four colors — or six, or twelve, or twenty" colors. I hope that makes it clear your earlier citations to handprint don't really support the idea that there is something notable about palettes with "3 primaries and 3 secondary" colors (given handprints extensive scare quotes and clearly stated position).Maneesh (talk) 06:20, 27 September 2017 (UTC)

I don't know how many more ways I can say this: I'm not specifically or exclusively talking about a palette of primaries, secondaries, or any other particular category of palette (or specific primaries or secondaries for that matter), even if I (or Handprint) might describe some examples of such.
I'm talking about a well-known conceptual type or category(edit) of color theory or color wheel, often "abstract" in the sense that the primaries do not directly represent any real (or even realizable) colorants, but potentially-useful as a schema or mnemonic for choosing and mixing any set of colorants. This type of color theory or wheel may be considered to be "defined" by having three "primary colors" distinguished (often but not always labeled as such), typically fairly "spread out" in hue around the wheel, and often three secondaries which are considered to be approximately or allegedly midway between them in hue in some sense (often not in any rigorous sense).
I'm talking about this sort of three-"primary" schema for a color theory or wheel(edit), irrespective of the precise primaries (or secondaries) chosen, or in fact whether it chooses precise primaries or secondaries at all -- it might be intended only as schematic, using names of colors or color ranges to represent its primaries or secondaries. But these schemas do not include the idea of being able to designate an arbitrary number of colors on the wheel as co-equal "primaries", as in a palette for example. This is a notable (traditional and still widely-used and described in some form) concept of primary color, though not the only one.
"Our" conception of a primary color (as a potentially-arbitrary number of colors that can be combined to form a gamut) could just as well have been named a color[space] basis or dimension or component, palette member (okay that one isn't good for an abstract CS like XYZ), etc., but as it happened/happens, the term "primary" was chosen and is widely-used in color science for this.
But that doesn't change the fact that the word/term "primary" is also widely used in the "three-primary color theory" or "three-primary color wheel" sense, in which any additional points on the wheel are not described as "primary colors" in the same sense (some might be called secondaries or just arbitrary colors or hues that might or might not be intended to approximate actual or realizable colorants).
"Primary color" is just a term -- there shouldn't be an ideological dispute as to whether we are "allowed" to document its notable uses because some are considered "inferior" in some sense or that some of its uses must necessarily supersede or depracate all other uses of the same word. In fact we are *required* to prominently describe *all* of the most notable usages of of "primary color", either together in the present article or using separate articles, and if in separate articles then they should prominently link to one another because the term "primary color" would then be ambiguous as to which article the reader is looking for. QuoJar (talk) 12:56, 27 September 2017 (UTC)

I suppose this isn't getting anywhere. My last comment is that I think the confusion lies in your the nebulous but "well-known conceptual type of color theory or color wheel". Which theory or wheel precisely, i.e. what is the proper name of this theory/wheel? What notable sources describe it? Who has used it and for what? What are the claims and predictions of this theory? It is necessary to be able to answer these questions in order to judge if "primary" has some sort of distinct meaning in that context. Handprint does describe a family of obsolete color models/theories that are grouped broadly under "material trichromacy", but this must not be what you are talking about since I've brought it up already. I strongly suspect you are referring to the amalgam of misinformed statements about color you will find many places like this very talk page, dictionaries, low quality educational materials etc. (op. cit.'d much earlier). Naturally this amalgam is not a well defined thing but just a jumble of misinformation. I'll abstain from this thread from here, good luck!Maneesh (talk) 17:59, 27 September 2017 (UTC)

Have you read color wheel or color theory? If this is non-notable misinformation, then these articles or much of their content should simply be removed. If so, I suggest bringing that up in those articles' talk pages, not here. We can't let this be the only article linked to by those that does not even mention the concept in the form in which it is being described in those articles. Those articles link to secondary color, whose content is consistent with those articles, but they also link here to primary color where the reader will likely be baffled by the inconsistency with the concept as it was described in those other articles. Simply calling it misinformed and omitting it from this one article (and not the others, presumably because you don't have any other content which you'd like to put in those articles, as opposed to the present one where you do have a concept called primary color that you would like to be here and completely exclude the other meaning) only serves to help make WP an incoherent, inconsistent web of contradictory articles. Traditional three-primary color wheels are still widely taught, published, and used in practice by artists, which makes them notable even if they should not exist at all because they are nothing more than "misinformation" in your opinion (or even if it were in the opinion of all color scientists, but not all visual artists, even if because the latter are "misinformed").
The usual way to deal with such a cross-discipline terminology disagreement is often to have two separate articles, one beginning "In color science..." and the other (like color theory does) beginning "In the visual arts...". That would be preferable to the present one-sided article, but I think it would be better to integrate both concepts into the present article, since the two concepts are closely related in meaning, even though there are aspects of them that are completely inconsistent with one another. And in practice for example when almost all printing uses CMYK inks, and almost all emissive displays use RGB, these particular usages happen to be reasonably consistent with both conceptions of "primary color", so it would be a shame to separate them. QuoJar (talk) 18:28, 27 September 2017 (UTC)

Without getting into a long reply, look at color theory and look at the paragraph on primaries ("Many historical..."). That paragraph seems entirely consistent with what is here. What is the inconsistency you are referring to?Maneesh (talk) 19:15, 27 September 2017 (UTC)

There is nothing in color theory that would suggest that one could or would ever add an arbitrary number of primaries, or designate an arbitrary set of three (or two) colors as primaries. If an artist uses, say, low-chroma red pigment and a low-chroma purple pigment (+ black & white) because these happen to provide the entire gamut needed for a particular intended painting, these would not be called "primaries" in any sense mentioned in color theory or color wheel. If an artist uses 11 pigments, some higher and some low chroma and saturation, these could not be described as 11 "primary colors" in the way they are described in those articles. On the other hand, in the present article, we would probably call the dull red and purple, or the 11 arbitrary pigments, each a set of primaries. So it's the number of primaries (three) but also the fact that in a color wheel or related color theory they would always be reasonably spread out around the hue dimension (not all in one corner for example), and they would all have reasonably high chroma or saturation. That's a big inconsistency between their concept of primary and ours, but in practice it doesn't become important so long as one sticks to abstract and common "general-purpose" emissive and printing primaries, which are usually reasonable primaries in both senses of the term. Also if we allow for the fact that in some context, some colors could be loosely described as "secondary", but if some version of those were used to generate a gamut, e.g. as palette pigments, then we would be in the awkward position of saying that these are merely "secondary primaries". Of course we could simply never use the term "secondary", but we can't deny that such a term has been widely/notably used. QuoJar (talk) 20:25, 27 September 2017 (UTC)

The sentence you refer to makes a factual statement, but does not prevent them from continuing to talk about three "primary colors" on their wheel, because they can do so without having to specify them completely, and without having to combine them to produce a gamut that they are limited to in some way. They can and do choose and talk about three primary colors anyway (not always the same choice of three colors, but still always three "primary colors"). The fact that these primary colors (as they describe them) can't be used to literally generate all visible (or generatable with a given set of what they would call palette colors) colors does not change this. To them, this is not a reason to add "more primary colors". Of course they can add an unlimited number of colors to their palette (although they might not have had any primaries per se on their palette to begin with) but they don't call these "primary colors" (they might call them palette pigments or colors) -- only we would call them primary colors, because we are using "primary" in a a way that's not consistent with theirs, but they are not required to call them that. QuoJar (talk) 20:57, 27 September 2017 (UTC)

Is this[18] an example of a color wheel or vague color system that you are talking about?Maneesh (talk) 22:36, 27 September 2017 (UTC)

At first glance, sure, that [19] looks like it could be one example. QuoJar (talk) 23:43, 27 September 2017 (UTC)

Ok, this specific wheel is pretty common at the art stores that I frequent. The wheel claims to be a guide for "mixing" (in this context, they mean paint) and claims that R, Y and B are primary and that paints with those hues cannot be mixed from any other colors. This is an obviously false claim, an inkjet printer can mix a decent red with its CMYK. I can't be sure it can mix that red color on the wheel itself but don't know enough about commercial printing to be sure it wasn't printed with a CMYK process. How is this wheel a reliable source of information? I agree that there are notable ideas and groups of people that are identified with things that are easily falsifiable. The group of people that believes R, Y and B are "primary" as described on this color wheel are simply misinformed, not a notable group. I don't understand why one would go to the trouble of making an entry for the definition of primary as described in this wheel since it isn't correct nor is the group of people that subscribe to it a notable or well defined group (unlike, say, people who believe in Flat Earth). [added as an edit:] The company contradicts itself on its own webpage[20] and links to handprint! Do you see what I mean that this type of thinking can't be delineated from simple nonsense? Maneesh (talk) 00:15, 28 September 2017 (UTC)

It looks like they are treating their stated "primary colors" as special in the sense that the wheel shows you the effects of adding each of them to an existing (non-primary) color (by physically turning that wheel to the position of that existing color). It doesn't directly show what happens if you add a blue-green to a red-violet for example. BTW, I guess they can probably generate all hue angles, and of course all lightnesses/values (by adding black or white), but it's the chroma/saturation that ends up being limited by the physical pigments being used (which of course are not limited to their three "primaries" nor to any precise versions of the vague colors mentioned on the wheel) -- but here I may just be speculating. QuoJar (talk) 23:54, 27 September 2017 (UTC)

The wheel says explicitly that the primary colors are ones that can't be mixed from other colors (RYB here and CMY on the company page). This is the amalgam of "popular color theory" that I am try to show does not have enough coherency to write much about other than to point out that some take such positions but they aren't at all coherent. I don't think such positions are held strongly enough by a notable group to be something notable like a Flat Earth, it's ignorance of a complex topic more than anything else IMHO. I think I can't do better than this color wheel to demonstrate my point.Maneesh (talk) 02:48, 28 September 2017 (UTC)

It's still a notable and widely-published and used point of view, even if it is scientifically, demonstrably false. They may believe that "ideal" colorants of these primary colors could exist, and that all real colorants fail to do so because they are "imperfect", which we know from color science is not true, but it also doesn't affect how their theories are actually used in practice, because nothing they actually do can contradict it because there are no "ideal" colorants. We as editors don't have the right to choose points of view based on whether we find them to be true or even logically consistent. What we can do is balance that point of view with others that explain why they are false. The fact is that people read and use these sources when they paint, and they find parts of them useful. Even if their stated definition is a false statement, that doesn't mean that they can't use it to label some points on a useful color wheel. The fact is that much of what else they do with the wheels, and much of the way they are intended to be used (or are used) do not depend on the truth of that claim.
Just as an mediocre analogy, one commonly-stated (for example in textbooks) "definition" of mass is that it represents "the amount of matter". This definition is absurd, meaningless, not directly usable, contradicts many other principles in modern physics (for starters matter has several legitimate definitions and much more useful ways of quantifying its "amount" by baryon number and lepton number), etc. Yet in the lead of mass it says "In Newtonian physics, mass can be generalized as the amount of matter in an object...". This is quite notable, not just historically but even today because Newtonian physics is still useful and widely-used (usually in contexts where it is accurate), and that "definition" is often stated, even in books published by notable physicists.
Also, is it possible that some sets of three colorants (inclusing some "RYB" or "CMY" sets) can be mixed to generate all hue angles, ignoring chroma/saturation? Certainly this is at least true for some ideal subtractive color mixing of dyes having certain absorption spectra, correct? Maybe their mistake is that they don't distinguish between hue in the colorimetric sense, vs. chromaticity or some analogue of it that describes everything about a color except its value/lightness? But even if this is false, their false statements, as well as the way the structure of their wheels is used in practice, is notable.
In any case, we can't exclude their point of view from this article unless perhaps if we begin it with something like "In modern color science, a primary color, or often simply a primary, is....", but in this case we would need one of those thingies at the top that says (a more concise version of!) something like: "This article is about "primary color" in the colorimetric sense as an arbitrary set of an arbitrary number of existing, physically realizable, idealized, or fully-abstract colors that are rigorously specified and capable of generating some calculable range of other colors when combined or mixed in some specified way. For a traditional conception of "three primary colors" in the the visual arts, see color wheel, color theory, and secondary color." Or it *could* even be "..., see primary color (traditional)" if someone wants to write such an article.
(Yes, of course the pedantic detail in my "this article is about" above is facetiously detailed.)
(In my personal opinion it would be better to reply right here in the gap below, rather than "inline" above, or else the thread gets out of order and some replies get "lost" in the mess.) QuoJar (talk) 13:38, 28 September 2017 (UTC)

A resolution that makes sense to me is to begin the article with something like "from the perspective of modern color science", then have something like "Popular Conceptions" as the first subsection (as opposed to "misconceptions") section with links to popular educational materials that say things like "RYB are the primary colors" (and a picture of the very wheel I've linked to, since we agreed that that is the kind of thing your are talking about). The article can then describe the lack of consistency and coherency amongst these popular conceptions. The mass analogy is interesting, I'll have to think about it. FWIW Materials (like the color wheel I linked to) that promote these "popular conceptions" aren't really used by skilled practicing artists, they are generally just pablum used to sell books and give students some nonsense to do.Maneesh (talk) 15:51, 28 September 2017 (UTC)

What would make even more sense to me would be if we mention it (briefly) near the beginning of the lead (and *not* dedicate an entire first section to something less-notable like this) because this is the only way that the reader, who may be a victim of such pablum and nonsense, will know that they have indeed found the relevant/correct article/topic, of which they may have learned a different (and less-notable) point of view, rather than being driven to search for a completely different article describing a more-familar sense of "primary color".
A WP article has to be quickly *recognizable* (even if partly via a debunked and meaningless but popular phrase) as to whether it is indeed the topic the reader is looking for, or else they will likely never get to the first section, and never learn that the point of view of this topic that they previously learned is.... outmoded.
QuoJar (talk) 16:58, 28 September 2017 (UTC)

Makes sense to me.Maneesh (talk) 17:53, 28 September 2017 (UTC)

Awesome. QuoJar (talk) 14:00, 29 September 2017 (UTC)

The following is an older part of the discussion than just above.

A color wheel has three secondary colors (and perhaps some tertiary colors), and for each color there is a complementary color. The present article doesn't mention anything about secondary or complementary colors. In the present article one could potentially use 11 primaries, and it isn't clear that defining secondaries between them would serve any purpose. In color science, complementary colors can be discussed in terms of opponent processing, etc., but this isn't necessarily directly related to the 11 primaries I might be using for a particular colorspace (or their "secondaries").

QuoJar (talk) 05:15, 25 September 2017 (UTC)

The color wheel article shows wheels of 7 and 12 and other numbers of colors. The 3/6/12 arrangement has become pretty common since the concepts of mixing primaries to get secondaries, and mixing those to get tertiaries is appealing. In color technology, 3 is most common, 4 is sometimes used for a larger gamut, and I haven't really heard of more than that being called primary, though it's possible. Secondary is a useful concept for mixtures of two adjacent primaries, but if you had a lot of primaries you probably wouldn't bother with that concept. Dicklyon (talk) 05:44, 26 September 2017 (UTC)

Okay, the problem is that the present article does not mention "secondary" at all, and if someone says they use three primaries and three secondaries on their palette, the present article would say no, all six of them are primaries because they are being used to generate a gamut, end of story. I'm concerned about this lack of recognition in the present article of this traditional meaning of having only three (real colorants or idealized/abstract) primaries (what ever they may be -- I personally don't care much whether they are CMY or RYB or some other variant) even if they don't necessarily literally generate every color (even if perhaps they *can* generate every hue angle, ignoring chroma & brightness) that your palette of colorants can generate. QuoJar (talk) 20:59, 26 September 2017 (UTC)

And credible sources do equivocate between "all six of them are primaries because they are being used to generate a gamut" and "three primaries and three secondaries". In handprint[21]: "Finally, the flexibility of the six color or secondary palette appears in the freedom you have to customize the three complementary color pairs — yellow with blue violet, red orange with green blue, and magenta with green — to suit your needs for a particular painting." The labels in this case really refer to the three that give you the biggest useful gamut to begin with and the next three that are used to grab slightly more volume, but again it's all in scare quotes because "primary" doesn't really really mean anything in this context. Maneesh (talk) 23:08, 26 September 2017 (UTC)

True. But the handprint guys also seem to be on a mission, with statements like "...some art school graduates develop a rigid attachment to primary colors and the formulaic approach to color mixing that goes with them. So it seems surprising to ask ... do 'primary' colors exist? Even more surprising to learn that the answer is — no!" So I wouldn't pay too much attention to the definitions and opinions from a cite that denies the existence of what we're supposed to be talking about. Dicklyon (talk) 00:23, 27 September 2017 (UTC)

Handprint guy, Bruce MacEvoy. I agree with essentially all of what handprint says, it is a remarkably thorough overview of color science and watercolor painting with lots of careful writing. I invite others to read through handprint[22] to understand the above comment is a rather blatant misrepresentation of its long and rigorous explanation of primary colors as understood by modern color science. Search for this quote: "The conclusion of this historical excursion is that "primary" colors are only useful fictions. They are either imaginary variables adopted by mathematical models of color vision, or they are imperfect but economical compromises adopted for specific color mixing purposes with lights, paints, dyes or inks." as well as "And if you are not building eyes or modeling color vision responses or running a printing press or designing a computer monitor, and can inexpensively "expand your gamut" with four colors — or six, or twelve, or twenty — on your palette, then "primary" colors are irrelevant to the task before you.". It is quite correct.Maneesh (talk) 06:07, 27 September 2017 (UTC)

The latest part of this section is above, before "The following is an older part of the discussion than just above".

recent changes to "Additive and subtractive color mixing"

While I think a lot of the most recent changes by Maneesh do indeed add some good material, I'm concerned about the loss of some aspects of some of the previous version text, such as:

When combining partially-opaque materials or pigments, a strictly-subtractive model is not accurate, as some of the incident light will strike a particle of one color and may sometimes be reflected without ever encountering a particle of another color, so that the other colors cannot absorb and "subtract" from that reflected light. Thus the result at any given wavelength might not be darker than some of the primaries. The darkest shades can be obtained by adding black as a fourth pigment, as in the cyan, magenta, yellow, and black (CMYK) inks used in printers. Printers generally do not need a white ink because they can allow parts of the white paper to show through instead. With opaque paints, the underlying surface might not show through, in which case white paint can be used to obtain lighter and less-saturated shades such as pastel colors.

In particular,

1. the loss of a term like "opaque" or "opacity",
2. the explanation of a specific reason/example where strictly subtractive isn't accurate: "the incident light will strike a particle of one color and may sometimes be reflected without ever encountering a particle of another color, so that the other colors cannot absorb and "subtract" from that reflected light"
3. Inability to achieve darkest shades in a not-strictly-subtractive mixing as one of the reasons for needing to add black.
4. Paper showing through in places where no dot of ink is placed, as opposed to adding white paint as painters might do

Also a good related point to add (which was and is not covered adequately) might be the fact that printers place dots of ink side-by-side (though some overlap occurs) instead of atop one another, leading to additive averaging as another (idealized) type of mixing, e.g. "Physical mixing of opaque paints involves a substantial component of additive-averaging as well as subtractive mixing. With transparent pigments the process is more purely subtractive." or "Use of minute interspersed colour areas results in primarily additive-averaging colour mixing.", both in http://www.huevaluechroma.com/061.php . I don't have time to add something like this myself at the moment.
QuoJar (talk) 14:35, 3 October 2017 (UTC)

The use of terms like "opaque" or "opaque paint" here can be troublesome since "hiding power" of paint becomes complicated. The particles themselves are generally transparent (except for carbon char based pigments), scattering/reflection, pigment particle concentration are the sorts of things that go into hiding power and makes explanations here complicated. I suggest avoiding the term "opaque" here, paint mixing is complicated because of things like scattering effects not because the particles are opaque. Will go to next point if we can agree on this one.Maneesh (talk) 16:00, 3 October 2017 (UTC)

It would be good to convey more than just how difficult it is (in the most general case) to make (precise) predictions. Partially opaque may be for complex reasons, but isn't it still partially opaque? Let's not get stuck on this one though, rather than moving on. QuoJar (talk) 18:25, 3 October 2017 (UTC)

Consider that rutile is partially transparent (often wine red) but when ground into titanium white, it has outstanding hiding power. Things like the variance in particle size distributions, concentration, impurities and presumably some other reasons we don't know yet etc. all make it difficult to predict what paint mixtures look like . Let's move on. Inkjet printing more correctly does apply additive and subtractive color mixing[23], the current writing is cheating by emphasizing "ideal", next edit will correct this.Maneesh (talk) 20:25, 3 October 2017 (UTC)