Talk:Subtractive color

It's not "Subtractive color space" It's just Subtractive color--[[Dkroll2]] 08:38, Dec 24, 2004 (UTC)

Where is a description of the attention needed? Notinasnaid 16:34, 26 Dec 2004 (UTC)

@Dkroll2 Can you explain to me why RYB even works if it should be CYM? Seems like R & Y (R light + G light) would cancel everything except Red and yield Red, not Orange. And Red & Blue would cancel all light and make black, not purple.

Subtractive?
It's logically multiplicative colour, isn't it? After all, the proportion of light that one filter ir pigment layer lets through is multiplied by the proportion of light that another filter lets through. &mdash;Ashley Y 23:40, 2005 Jun 2 (UTC)


 * "Subtractive" because the object that you view only subtracts light; the opposite, additive color, is used where the object that you view adds light. Notinasnaid 12:41, 3 Jun 2005 (UTC)

From the article: "except the imperfect system resulting from mixing real pigments, something that is very confusing nowadays." Could this be explained or a link provided that does explain it? How is this confusing and how is it neither additive nor subtractive? Threepounds 06:44, 8 December 2005 (UTC)


 * I have to agree. I've deleted it. If this really is outside additive/subtractive color that is both interesting and useful, but it isn't really either unless it is explained. Notinasnaid 09:44, 8 December 2005 (UTC)


 * It seems to me that mixing oil paints is mathematically equivalent to adding the RGB values in the proportions mixed. So if I start with green paint and add white, it gets brighter, but if I add black it gets darker, so is this additive or subtractive? Neodymion 04:50, 20 September 2006 (UTC) Actually, it seems the Kubelka-Munk Theory is the actual science for this link title but doesn't appear to apply to translucent things like inks. Neodymion 10:57, 1 October 2006 (UTC)

It seemed necessary for me to extend this article with "For those who want to know more ..." section and an image of layers of coloured glass. As I work on Faculty of Graphic Arts, University of Zagreb, I may be able to get expert opinions on the subject from Department chiefs. I currently work on a project in enhancing of teaching material in this area (reproduction of photography department). Somewhere it should be said about use of actual pigments. Although I might agree that my wording on the subject may not be the luckiest.

I have an additional Image:Color-additive-mixing.png picture that is three-dimensional and may catch the eye, for other pages. This subject is IMHO the oprotunity to visually enhance Wikipedia, isn't it? If you have an idea of simple animation showing better how this color system works, I may be able to do it in POV-Ray, as this could fit into my project. Mtodorov 69 22:17, 24 February 2006 (UTC)

Silver and Gold aren't "colors" per se.
Ideas For your point you are trying to make, use monitor green (vivid chartruse you see on a computer monitor.) And make 2 images that domonstrate how the green turns to deep evergreen in CMYK. Or read color printing and add to that.

Lack of information?
I'm curious why no one has yet to mention the fact that colors are multiplied. If you want to get all upity (not a real word, I know) you could say that it's "subtractive color" because there is less light than when you started, and I'll buy that for the title of the page, however there is no good reason why we shouldn't explain how light is multiplied, not subtracted.

For those of you who don't understand it, there is one big problem with subtractive color (other than the fact that you can visibly see that it's being multiplied, not subtracted), and that is negative color. There is no such thing as negative color, yet if you subtracted one color from another it seems like you could produce negative color very easily.

Light is measured in values between 0 and 1, including both 0 and 1. 0 is black and 1 is white. When you multiply two decimal numbers less than 1 you get a smaller number, which is why when you multiply two pigments together you get a darker color.

For example, if you combine magenta (RGB decimal values 1,0,1) and yellow (RGB decimal values 1,1,0) then you multiply each of their values. $$1 * 1 = 1$$, so the R decimal value is 1. $$0 * 1 = 0$$, so the G decimal value is 0. $$1 * 0 = 0$$, so the B decimal value is 0. Thus, when you multiply magenta and yellow you get Red (RGB decimal value 1,0,0).

This needs to be explained in detail by someone at some point. —The preceding unsigned comment was added by Vjasper (talk • contribs).


 * I agree with the principle of this, although the details are a bit off. Light is not measured in values in the range [0,1] since there is no upper limit on brightness, but it's a good model for a filter. It might be more useful to think of a grey filter that reduces the light by 0.5. Two of these will reduce it to 0.25 and so on. A pure cyan filter stops nearly all the red light so it's harder to see the difference between using one or two of them. It's less easy to see with ink since there's a limit to how much ink you can put on one spot. Neodymion 04:34, 20 September 2006 (UTC)
 * Now I have to disagree. When talking about color, RGB (light) values are indeed measured in the range 0 to 1. See RGB color space and other articles about RGB. I understand what you're saying about an upper limit on brightness, but that hasn't stopped color scientists. Notinasnaid 08:06, 20 September 2006 (UTC)
 * But you agree with the principle? It seems rather important to get the science right in an encyclopedia. I would support creating a new article 'Color Mixing', especially since everything else I've seen on the net claims subtraction (and half of them still use RYB instead of CMY).Neodymion 06:43, 30 September 2006 (UTC)

Article too technical
This article spends too much time focusing on printing, and does not spend enough time really fleshing out the details of subtractive color. This article needs to be rewritten for a more general audience, and quite a bit of the content here needs to be moved to a more appropriate article. This article is about subtractive color, and nothing else. Stack 00:00, 15 March 2007 (UTC)


 * What kind of details are you looking for? The article’s content seems fairly reasonable to me, though much of it could be reorganized and maybe fleshed out a bit more.  I think it would be good to have more (even very condensed) historical discussion of the uses of pigments and dyes, with some discussion of the associated changes in color theory, for instance changing the colors considered "primary" from red, yellow, and blue, to magenta, yellow, and cyan, as improved cyan and magenta pigments were created through technological advance.  Not to mention a bit more discussion of color photography. It would also maybe be good to have some images showing spectral power distributions, and demonstrating the "subtraction" obtained from mixing of multiple pigments which absorb different spectra.  But I don't see anything here which should obviously be removed. --jacobolus (t) 08:04, 16 May 2007 (UTC)


 * Most "too technical" complaints can be dealt with by proper use of the lead section. This article, like most, does not have one. The lead section should be a summary of the entire article, with no unique information. That is to say, removing the lead section would leave a complete technical article, and the lead section alone would tell a less technical summary. Notinasnaid 12:20, 16 May 2007 (UTC)


 * I agree, it focuses a lot on printing and its beingness of a color model as inks. It would be better if we make it focus on the CMY concept and its principles and technicalities. Jaspergeli (talk) 16:42, 27 January 2019 (UTC)

Article Grading wildly innaccurate
''Much of the data for subtractive color is wrong. first of all, the CMYK scheme is for additive color. The true primaries of subtractive color and of the world of art are, red, yellow, and blue ...''

CMYK is not additive color. Black would never appear in an additive color system. Both CMYK and RYB are correctly referred to as subtractive color schemes (setting aside that multiplication is the actual arithmetic operation).

There are no "true primaries of subtractive color," to say that there are reflects a profound misunderstanding of the science of color.

It's possible to transform color coordinates between CMYK, RGB, and RYB. The fact that someone might use RGB colors as the source for CMYK, or vice versa, does not mean that they both belong to the same (additive or subtractive) color scheme.

(the previous unsigned comments were left by 69.134.236.134)


 * Um, I'm not sure what you mean here… of course there are no "true primaries" (does the article say that there are? if so it should be amended), but CMYK and RGB are definitely different types of color models. In one, colors of pigment (which absorb particular wavelengths) mix to near black, while in the other, colored lights (which emit particular wavelengths) mix to white.  One of these is "subtractive," and the other is "additive."  That you can transform colors from one space to another (approximately; process printing and RGB displays have very different gamuts), doesn't imply anything about the color models. --jacobolus (t) 05:36, 16 May 2007 (UTC)


 * I do agree that the article is far from great though, and I downgraded it to “start” class. --jacobolus (t) 00:24, 29 July 2007 (UTC)

article too focused on cmyk color model
We already have an article for the CMYK color model, which is easily linked from this article (and is recently improved enough that I feel comfortable doing that :) ). But there are many other subtractive models, including those used in photographic processes, painting, etc.  This article should be more of an overview of what subtractive color means, and leave describing CMYK in detail to the relevant article. --jacobolus (t) 00:26, 29 July 2007 (UTC)

I agree to this jacobolus(t) and I will try to improve this page Uddhav9 (talk) 11:29, 5 January 2019 (UTC)

Assessment comment
Substituted at 07:13, 30 April 2016 (UTC)

RYB
My changes to delete the RYB section of this article were reverted. Where, in the paragraph of this article, are RYB described as a subtractive mixing model or an example of using RYB primaries described? As I said in the edit summary, the section mentions painting but it is widely understood that subtractive models do not predict the appearance of paint color accurately. What is the justification of the revert then? The edit summary of the revert claims "Historically, RYB was and is important in art and was described as subtractive."....where has RYB been described as subtractive?Maneesh (talk) 18:55, 4 March 2019 (UTC)


 * I was the editor who reverted; thanks for bringing it to the talk page. That the RYB primary colors are used for paint mixing and that they form a subtractive color system is elementary art school 101 level of common knowledge, at least in the US. Look at any color wheel in an art instruction book--almost all of them will be based on RYB and paint mixing, it the source goes into that depth, will be described as subtractive. A quick google search yields book sources like and  that discuss the idea. Here are some not quite RS that discuss the RYB vs CMY issue: ,.  Sure, CMYK is superior to RYB in many ways as a subtractive color system and should be described as such in the article. But Wikipedia is descriptive, not prescriptive, and we would be remiss in not describing RYB as an important and probably dominant primary subtractive color system in the art world. --  20:25, 4 March 2019 (UTC)
 * I don't think there is anything authoritative about those sources which are sort of standard commercial, confused, low quality art books that aren't notable and don't really have any underlying verifiable sources. Can you find a notable source that is closer to (proto) color science, art history or art?  This isn't an argument about superiority, this is about asking what RYB really means.  I think RYB has been appropriately trimmed down a little to only say things that have some verifiable support.  This article is making claims about RYB as a subtractive system for painting.  That doesn't make any sense to me because subtractive color isn't used to predict the color of paint mixes (since they, ostensibly, do not work like color filter layers).Maneesh (talk) 21:00, 4 March 2019 (UTC)
 * Some notable sources that define "subtractive color" (perhaps the original definition?) here, here. here and Ponyton's FAQ. All of these source confirm that 'subtractive color' refers to the model that multiplies transmission/absorption across layers.  What does that have to do with "RYB" (the use of red, yellow and blue) beyond perhaps mentioning a specific notable application where red, yellow and blue filters were used?  It's easy to see why subtractive models don't predict what happens in paint, Kubelka-Munk (mostly) do. RYB doesn't need a section in this article.Maneesh (talk) 23:35, 5 March 2019 (UTC)
 * Those are good sources for the scientific approach to subtractive color mixing, but they are irrelevant to the assertion that RYB subtractive color mixing conceptual model is widely used in art and art education. Similarly, whether it makes sense to you is also irrelevant to the assertion. The RYB subtractive approach is usually taught orally in elementary schools, so I can't point you to textbooks for that. But art instruction books that cover color and paint mixing often discuss it. I gave two book refs above that are fine as refs for the assertion. Here are some more. Let's see, from my personal library, The Artist's Handbook of Materials and Techniques, a standard handbook in the field, talks about subtractive RYB a bit on page 161. The book Blue and Yellow Don't Make Green, near the start of the book, talks about RYB subtractive mixing in detail and why it is not the best approach to paint mixing. The book The Watercolor Book: Materials and Techniques for Today's Artist pages 78-79, talks about subtractive color mixing. I could find many more sources; this is ubiquitous knowledge among artists. I am not opposed to giving RYB due weight; CMY and CMYK should have more content devoted to them as the dominant approaches in the printing and photography world. But deleting the entire RYB section, as you did, was not due weight to the RYB subtractive approach used in the art world. --  04:48, 6 March 2019 (UTC)
 * I have some familiarity with painting; no one I've ever learned from suggested that RYB are subtractive colors. I am familiar with these sorts of art books, they don't have a lot of credibility when it comes to color science and terms like "subtractive color". Surely you agree a subtractive color model is one that predicts spectral power distributions by multiplying by transmittance/absorption profiles of stacked filters, yes? You must also agree that such a model does not do a good job predicting the appearance of paint mixtures, yes?Maneesh (talk) 07:47, 6 March 2019 (UTC)
 * What must be some of the earliest uses of "subtractive color" found via google ngram: from Light and Color, from The Photographic Times-bulletin. The idea that subtractive color is related to sequential filters is really as clear as day there.Maneesh (talk) 20:23, 6 March 2019 (UTC)
 * At least some of your sources are really dubious. I browsed through "Blue and Yellow Don't Make Green" on amazon and I could immediately spot fallacies that looked familiar but I had never read the book.  Then I remembered handprint debunking the author's claims here. I am very concerned with this sort of pseudo reasoning being represented uncritically in WP.Maneesh (talk) 22:45, 6 March 2019 (UTC)
 * In summary, subtractive color clearly refers to predicting the spectral distribution of light passing through sequential filters by multiplying transmission/absorption profiles. This is quite tangible here on Marc Levoy's cs178 page at Stanford, where you can play with curves and see results calculated by a program.  Note the lament about how "subtractive color" is really misleading since it involves multiplication. More support from The Photographic Times-bulletin (1903), from Light and Color (1931) and a more modern colorimetry book here.  A very nice comprehensive explanation is found at handprint that also describes the confusion around the term. The sources that Mark viking cites are not reliable, they represent confused thinking often found in art books. Indeed handprint debunks  Michael Wilcox's ideas and reveals them to be quite crankish (he doesn't know why yellow paint looks yellow). It isn't at all clear what "RYB subtractive mixing" is. Is it a model? What does it predict? Is there a program I an use assess those predictions?  If subtractive color does have to do with sequential color filters, then how could it possibly apply to paint? I can't imagine devoting any more than a sentence to RYB mixing in this article that said something like "Some books on painting refer to red, yellow and blue as 'subtractive primaries' or to an 'RYB subtractive mixing' model. These claims are not consistent or coherent with long established color science/colorimetry.  Subtractive color models are not applied to predicting the color of paint mixtures." There is a claim in the above thread that RYB subtractive mixing has something to do with an "oral tradition".  I have some familiarity with painting and have never heard anyone tell me about RYB subtractive mixing nor have I applied it as a matter of practice.  I *have* made paintings with only red, yellow, blue (black actually) and white pigments but there is nothing special about mixing those pigments vs. any others.Maneesh (talk) 20:00, 8 March 2019 (UTC)


 * Let me address some of your points. References to art books are great for describing artistic practices, which is what was needed here. The stacked, multiplicative, selective filter approach to subtractive color mixing is one model of subtractive color mixing, but not the only one. Stacked filters are a fine model for describing color filters in front of a camera lens or photographic film. CMY halftone printing, however, which is considered a subtractive process, is not a stack of filters. It is color dots of varying size, that may or may not overlap, with the color of a region given by a mixture model of pigment/dye and paper absorption spectra. Mixing of paints usually involves the mixing of opaque pigment particles, so stacked transparent filters is a terrible model there. As much as some would wish it, there isn't a single True Model of subtractive color mixing; rather there are several notable approaches that take a subtractive approach. The RYB subtractive conceptual model taught to artists is one of those models, one that is widely known. We should use reliable sources to summarize strengths and weaknesses of the different approaches, to compare and contrast them, and to give due weight. Agreed that the RYB subtractive model doesn't merit the level of coverage of a CMY or CMYK halftone approach, or a stacked, radiative transfer approach. But as a widely known subtractive model, it merits some coverage. Suppressing other notable approaches to emphasize one favored model violates neutrality; neutrality is one of our pillars and should always be taken seriously. With regard to the handprint website, I am not sure that it is a reliable source. It looks like a one-person operation and is effectively a personal website. Is there any independent editorial oversight? I think we have had a good discussion, but by now it is becoming clear that our dispute is more due to differences in our views of what Wikipedia should be than in particular disputed facts that could be resolved. Perhaps it is time to bring in 3rd party opinions? If you concur, I could contact Wikiproject Color. -- 22:21, 8 March 2019 (UTC)
 * Pigment particles are (perhaps except for carbon blacks) *not* opaque, take a look at a picture of them in a microscope sometime (you might be surprised to look at a rutile crystal that titanium white comes from). Subtractive models are still not used to predict the appearance of paint mixtures though, the handprint link I've provided twice explains why clearly.  Subtractive color is only defined in one clear way: multiplying spectral distributions to predict the spectrum of light passing through sequential filters (I think the earliest use of "subtractive" is from Louis Ducos du Hauron). Please, just tell me what these artists learn when they learn the RYB subtractive model? What do they predict with this 'model' that doesn't seem to be clearly defined anywhere?  I think you, and the books you cite, are just thinking of the vague idea of red, yellow and blue as conceptual colors that in some way can be used to mix other colors.  I can see why one might associate the term "subtractive", since white light is partially absorbed by paint to reflect color in the eye, but *all* paint is subtractive in that sense (not just red, yellow and blue ones). This seems to simply be a widely promoted misconception since it is quite clear (I hope) that paint piles are more complicated than sequential filters. The sources I've provided are quite concrete, there are clear prescriptions (programs even) for subtractive models and exactly what they predict (spectral distributions). Those sources also explain why subtractive color doesn't explain the appearance of paint. The comparison and contrast here is precisely that when art books talk about "RYB subtractive mixing" they seem to be very confused and it cannot be aligned with the use of "subtractive" in color science; we do not use subtractive models to predict the reflectance curves of paint mixtures, they are simply do not do that accurately. By all means, invite anyone you wish to assess my claims.  EDIT: Wrt to handprint, it's an excellent reliable site; but don't rely on my word, you can check the claims that come from there quite easily and the quality of sources it cites. Compare handprint to what you are reading from Wilcox. Maneesh (talk) 23:20, 8 March 2019 (UTC)

I think you understand the basic idea of the RYB subtractive color model. It is a qualitative empirical model, a conceptual model, that is widely used by artists to understand and predict color mixing. I think the thing you are getting hung up on is: if there is a quantitative subtractive model better at predicting color, why bother with such a conceptual model? Our neutrality policy is one answer, but you seem to discount that as well. Even from a blinkered color science POV, however, these empirical models should be covered in the article, because they have their uses. For printing we have Demichel-Neugebauer models, Clapper-Yule models, cellular Neugebauer approaches, Kubelka-Munk theory, etc., all are models which are conceivably useful. But manufacturers reject all that theory in favor of simple empirical color management schemes--not because they are cussedly against sophisticated approaches, but in part because in real life, such models are vastly under-specified. No one knows the particular quantitative properties of a given printer, with given ink cartridges, with paper of unknown properties, on a day with unknown temperature and humidity. The first principles models are mostly useless in the case and the best we can hope for is a prediction of average behavior derived from a printer specific empirical model. Now consider the artist with dabs of paint on their palette. They will be mixing quantitatively unknown amounts of paints, themselves of unknown, trade-secret compositions and applying them to canvas or paper of unknown properties. In this case, sophisticated first principles models, even if they got the physics right, are useless for actual real-time color mixing. Quantitative models are useless. The most useful model for an artist under these circumstances is a qualitative empirical model of color mixing. A basic model from which they can use personal online-learning techniques to learn how to mix their paints. The most widely used such model is a simple subtractive model with an RYB color basis, that has been taught for generations, as described in the sources I gave you. One could use other color bases; a CMY color basis would likely give a better gamut. One could try to come up with a more accurate qualitative model. But at Wikipedia, we describe the world as it is, rather than the world as it should be. And so the widely used RYB subtractive model, described as such in the sources, should be in the article. Regarding handprint, it does sound like a one-person operation. That the person uses good references is encouraging, but reliable sources have fact checking and editorial oversight. This is why journals are considered potentially reliable sources, whereas sources like unpublished Arxiv preprints and personal web sites are usually not, according to the WP:RSSELF guideline. -- 22:00, 9 March 2019 (UTC)
 * I support Mark Viking in this. Maneesh has been all over trying to narrow WP's coverage to a narrow modern colorimetric view, but NPOV requires that we represent all points of view.  And RYB is very often described as a subtractive color model (since pigments absorb light); it is widely used with a "red" a bit on the magenta side, and a "blue" toward cyan; that makes it more useful, not less of a thing. Dicklyon (talk) 04:22, 10 March 2019 (UTC)
 * And here is an example of a book calling RYB "subtractive"; there are lots. Here is another. Dicklyon (talk) 04:32, 10 March 2019 (UTC)
 * It is unfortunate that you cite such unreliable sources. These are typical art books that don't make a whole lot of sense. Look at the definition for "subtractive color" in one of your links: "Color that is created by mixing pigments. The subtractive primary colors are red, yellow and blue pigments". Does that sentence make sense to you?  That definition excludes sequential filters, since pigments are not mixed in that case. I have no interest in narrowing WP's coverage but it really ought to be made clear where definitions can't be aligned with established colorimetry, that is clearly the case from "Foundations of Art and Design" (and many other comparable, and entirely common, book). Maneesh (talk) 08:02, 10 March 2019 (UTC)
 * It's not all that sensible to a color scientist, but makes adequate sense for artists. This "subtractive" model is not identical to your ideal or more general case, but it's still subtractive, that is, based on absorption by pigments.  Feel free to clarify. Dicklyon (talk) 22:23, 10 March 2019 (UTC)
 * That is a rather broad generalization to suggest that this model makes sense to artists. None of the artists I know have ever mentioned that model and are generally aware of colorimetry. Even the Wilcox's book, "Blue and Yellow Don't Make Green" cited by Mark viking specifically denies what this source says about the RYB subtractive model "Y+B=green". Wilcox says a lot of other interesting things if you leaf through the book "virtually everything that has ever been written about color mixing has been inaccurate". Ok then. Do you really believe that this morass of contradictory ideas makes any sense to anyone including people that identify as artists? Is it sensible to try and represent each of these positions in this article?Maneesh (talk) 17:18, 11 March 2019 (UTC)
 * It's a little funny to see someone be concerned about handprint when they are citing such an easily falsifiable source like Wilcox. Handprint URLs have been cited a number of times in the published literature. "It is a qualitative empirical model, a conceptual model, that is widely used by artists to understand and predict color mixing.".  Really? So lets be specific, can you find me a reference that describes the procedure to predict the color of mixing, say, burnt umber and ultramarine blue oil paint using the "RYB subtractive color model"? Other pigments? You seem to be subsuming the entire practice and heuristics of paint mixing under the "RYB subtractive model" (there are many pigments that have red, yellow and blue hues). What tradition or institution of art uses this model? I've never heard of any sort of specific principles using this model.  Here are few "personal online-learning" resources where painters who (broadly) paint realistically describe the process of restricting themselves to RYB oil pigments (not really, because you can't paint much realistically without using tonnes of white paint in oil), but none of them call that "RYB subtractive mixing": Paul Foxton, Mark Carder, Cesar Santos.  The usual term in these circles is "limited palette", you can see the same on the figure painting class descriptions on the Florence Academy of Art. Can you show me something comparable where an artist paints a picture and describes how she is using the RYB subtractive model to predict the appearance of pigment mixtures? There is a very clear separation between color science terms that form a generally consistent vocabulary that is supported in published scientific literature/executable models, and then the mess of popular color theory/art books.  You'll find it is very difficult to draw any sort of box around the latter (past the primary/secondary color relationships) since it is not consistent with color science or even among members within that mess.  E.g. look how at the source  Dicklyon  cites that defines subtractive color as color created from mixed pigments; that excludes mixing via sequential filters. The article needs to make that distinction clear if it is going to represent both the generally clearly defined notion of subtractive color and the "low quality art book" notion at the same time.Maneesh (talk) 00:56, 11 March 2019 (UTC)
 * Nonsense. The definition that "excludes mixing via sequential filters" is not excluding, it is just a definition in a domain where that concept is not what's relevant.  If you're trying to say that RYB is not a subtractive color system, you are making that up, synthesizing a position from a color-science point of view instead of just reporting what sources say. Dicklyon (talk) 05:17, 11 March 2019 (UTC)

A Commons file used on this page or its Wikidata item has been nominated for deletion
The following Wikimedia Commons file used on this page or its Wikidata item has been nominated for deletion: Participate in the deletion discussion at the. —Community Tech bot (talk) 13:22, 25 October 2022 (UTC)
 * Pigment Colours - Classification.jpg