Wikipedia talk:WikiProject Color/Archive 2

From Wikipedia, the free encyclopedia
Jump to: navigation, search

Archive 1 | Archive 2 | Archive 3

Split into subcategories

Category: Colors is getting large and will get larger as more colors are added, so I thought it would be a nice idea to split the category into subcategories to make some navigation easier.

Here is what I propose:

So, what do you think? Andros 1337 10:23, 25 Jun 2005 (UTC)

I think Red (color) should include genreral discussion of all shades of red. Pink is a special case, because it is psychologically distinct -- but all shades of blue, light or dark, should be discussed and sampled on Blue (color).
You have the correct general length of the list quite correct: 11 (or so) general color pages, each of which should display a few swatches representative of that color. But there is no compelling reason to multiply pages. Nothing can ever be said about Teal or Aquamarine sufficient to warrant its own page. A large number of redirects might be entertained, such as Aquamarine (color)Blue (color). But there is no advantage to dozens, even hundreds, of individual color pages. — Xiongtalk* 21:03, 2005 August 7 (UTC)

Religion

I've noticed that several of the colour articles have a line describing the meaning of the colour in 'religion' but the religion in question invariably seems to be Christianity. This could do with some rewording, but since I don't have time at the moment I thought I'd mention it here in case some enthusiastic soul felt like taking on this rewarding task. — Moilleadóir 23:00, 13 July 2005 (UTC)

Islam is strongly identified with green -- a particular shade of green if I'm not mistaken. Sikhs wear light blue turbans as a banner for Sikh independence. Buddhists wear saffron; Easterners of all kinds wear white in mourning (which is probably already mentioned). That's all that floats to the top of my addled egg. — Xiongtalk* 21:07, 2005 August 7 (UTC)

Discussion of CMYK

Why the CMYK function given above? This is simple, but gives totally unrealisic colors (see CMYK). I agree with the assertion that CMYK does not belong in the message box because it is not well defined. For that matter, nor is "RGB". See absolute color space. I hope this isn't taken as a personal attack, but I do think the color boxes might have been conceived and perpetuated by people who didn't actually understand color science.

I have seen the proposal that CMYK be dropped from other people than me, but nothing seems to happen (in terms of reply or discussion). How can we actually get this debated. What use is this information supposed to have? Notinasnaid 10:33, 30 July 2005 (UTC)

Walk into any bar frequented by pressmen and graphic designers and shout "CMYK conversion!" Then get a beer and go stand in a safe corner, watch the fireworks. On second thought, order the beer first and stand near the door, ready for hasty departure. — Xiongtalk* 23:12, 2005 August 7 (UTC)

As one of the above maligned groups I would argue that the CMYK values are useless in their current form. For example of you print 100,100,0,0 as defined in Blue prints out a very dark navy. This is not to say that CMYK is useless, just that the current conversion equation is way out. This is because 0,255,255 is cyan, NOT process cyan. Process cyan is quite a different colour, and is the colour of the real printing ink. While print colours will always vary, they don't vary nearly as drastically as the colour conversions in wiki. My proposed solution is to change the colour conversion macros for cmyk, and fix the pages affected. The macro should be based on assumed profiles of sRGB (as endorsed by amongst others W3C) and SWOP (SWOP being more arbitrary, but still the most common), then rounded to the nearest 5%. CMYK needs to (I vote) be fixed, or deleted, if we're to cover process cyan et al. --Bb3cxv 10:46, 31 October 2005 (UTC)

What is Color?

This debate never ends. Color is about 25 times more complex a subject than reasonably intelligent laymen ever guess it could be. Let me see if I can hint at it in a short space:

There are five general ways to influence the perception of color: Emission, addition, transmission, refraction, and reflection. Texture is important too, but beyond our scope here.

Emission

A substance is excited in some way; that is, it is stimulated so as to release photons. Photons may be of almost any given energy level; the higher the energy of photon, the shorter the corresponding wavelength. (See: Quantum Mechanics) If these photons fall into an extremely narrow range of energies, and strike the eye's retina, light is perceived.

White light is a psychological perception of the absence of color, conditioned by exposure to natural daylight. (Daylight itself is a complex thing, about which more will be said later.) In general, a "basket" or rich mixture of photons of various energies will be seen as white. There is no such actual color, in the physical sense.

Blackness is the absence of all light, rarely enjoyed in pure form.

Examples: Direct sunlight is more or less white; burning organic matter such as wood or cotton generates light with a yellowish cast; but these are blends. Much more specific colors can be seen when certain elements are heated to incandescence. Sodium is famous for its yellow color, so bright that slight contamination of other samples often masks their true colors. Hydrogen is known for its blue color. In both cases, however, photons of more than one energy are generated. Lasers are the only known sources of pure monochromatic light.

Emission is the genesis of all color experience, for it is how all light is produced. The color of emitted light is dependent upon the emitted photon's energy, which in turn depends on the amount of energy lost as electrons fall to lower energy states. Thus color has been key to understanding of atomic structure and is the gateway to quantum mechanics. (Albert Einstein won the Nobel Prize not for General Relativity, but for his explanation of the Photoelectric Effect)

Addition

Many sources of light produce photons of more than one energy; thus the color perceived is the sum of all. Of course, there is nothing to stop one from setting up two light sources, each a generator of a single color, and shining them into the same eye at the same time. This will result in the perception of a third color; however, this is an illusion.

The standard additions are:

  • Red + Green = Yellow
  • Red + Blue = Magenta
  • Green + Blue = Cyan
  • Red + Green + Blue = White
No violet photons were used in the making of this image.

Remember, however, these are purely illusions. Red and green lights do not interact in any physical way; they do not combine to form yellow light. It is merely the nature of the eye that we misperceive a blend of red and green as yellow. In the figure, squares are filled with a red/blue checkerboard. Except for the white border, no pixels are any other color; the perception of magenta or purple is a mere illusion.

Already, we see that our perception differs from reality. Color is not an exact science. Anyone purporting to make definitive statements about color runs the risk of confusing color as seen by the human eye and color as it is exists in the natural world. It is known that many animals do not see color at all, or see it differently from Man; indeed, not all humans see color in the same way.

Transmission

Light of pure color -- prior to the creation of the CRT -- was very rarely viewed directly. The next step in removal from manipulation of the light source itself is to interpose a substance between source and eye.

Transmission-optics.png

While any substance will emit light if sufficiently heated, many substances do not pass light well; most do so without altering it in color. For a material to alter the color of passing light, the molecular bonds must be strained. The distorted electron field selectively absorbs light of certain wavelengths, but passes others.

Obviously, color by transmission is only apparent when the emitting source is an addition of several colors. If the source emits, say, only blue light; and the filter absorbs blue light; then no light will pass. If the source emits only red light; and the filter absorbs only blue; then it will have no effect. But if the source emits white light (a mixture of many colors); then the filter will pass light with the blue photons absorbed -- which will appear, generally, yellow. Such a filter will generally be called "yellow", although it acts only on blue light.


Refraction

144px|right The seminal Pink Floyd album Dark Side of the Moon's cover features a rather idealized illustration of a classic experiment. White light passes through a glass prism and is refracted into a rainbow -- a band of colored light.

This is because light of different colors (energies) is refracted at different angles. The colors thus seen are real, not imaginary.

Refraction-optics.png

Perhaps a simple demonstration will illuminate. Cast a yellow beam through a prism; the beam will be bent as it passes through, but will remain a tight beam. If red and green beams are added together, they will appear yellow to the eye; but the prism will separate them.

This underlines the subjective nature of color. Well-chosen red and green lights may be skillfully blended so as to exactly simulate yellow to a given eye -- but yellow light has not been so created!

Indeed, materials absorb light of one wavelength (color) and emit another -- another true color -- but these are effects are rarely seen. Usually, the material must be illuminated with high-energy (ultraviolet) light in order to fluoresce in the visible range; most commonly, visible light is absorbed and the energy emitted as heat -- infrared light.

In most cases, when it appears that colors have been created through some combination of light sources and pigments, it is merely that the eye has been fooled.


Reflection

Last in our discussion, but most common of all, is the illusion of color produced by reflection. As in transmission, some materials absorb photons disproportionately in a certain range of energy levels (= wavelengths of light = colors); again, it is due to strained molecular bonds and distorted electron shells.

A certain material may absorb blue light and reflect all others; such matter will appear black in blue light, red in red light, and yellow in white light. From a physics standpoint, it is itself always "black" -- it emits no light of its own.

Paint is a material that is designed specifically to absorb certain colors and diffuse the rest; so is ink and so is dye. These pigmented materials are labeled and sold according to the illusion of color so produced; so the pigment used as an example above is called "yellow".

Pigments may be mixed together and examined under white light. This is called subtractive mixing and is complementary to the additive mixing mentioned earlier. The standard mixtures are:

  • Cyan + Magenta = Blue
  • Cyan + Yellow = Green
  • Magenta + Yellow = Red

Wise readers will anticipate CMYK with the demand: "Well, what is the K for then? Why use black ink?" They will also grumble: "When you mix Cyan + Magenta + Yellow, you don't get Black; you get a sort of muddy brown." These questions answer one another.

Remember that ordinary sunlight has a broad and continuous spectrum. Nearly every imaginable visible wavelength of light falls on the Earth; indeed, that is why it is the visible spectrum; evolution has found it most useful. Other common sources of general illumination are made to resemble sunlight to a certain degree, for human comfort. Thus a piece of "white" paper reflects a broad range of colors, which together appear white.

To simulate the broadest possible gamut, or range of possible colors, the standard C, M, and Y four color process inks are chosen to be rather light and absorb only a rather narrow range of colors. Mixed all together, they are still rather light in tone, a mere muddy brown. For the appearance of dark color, K (black) is added -- a pigment designed to absorb as much light as possible, equally of every wavelength.

This is an extremely clever system, since by its means, a wide gamut may be simulated using only four inks. However, it is still quite limited. One reason for the tremendous mass appeal of CRTs -- televisions and computer monitors -- is that they offer a much wider gamut than ordinary process printing. To the uninitiated, they appear unusually vibrant and life-like, compared to printed photographs.

Those who must produce high-quality printed matter frequently resort to colors other than the standard process inks. These are able to simulate colors outside the process gamut, and simulate with greater fidelity those within; they are tailor-made to the occasion. However, printing with spot inks grows rapidly expensive; 8 colors means 8 printing plates, and 8 trips through the press. Most commercial presses are able to print no more than perhaps 6 colors, one after another; it is not considered cost-effective to reload the printed pages and take them through the press a second time for 6 more colors. While it is common for manufacturers to insist on one or perhaps two spot colors, chosen to match the company logotype, it is very rare to see a mass-market item printed with any more.

Illumination

We close the circle by returning to the source of "white" light under which some pigmented matter is examined. We know by now that there is no such thing as truely white light; all light thus named is merely a blend of other lights. It is possible to define a standard blend as "white" and indeed this is done -- again and again, each such standard differing from the next.

I should dearly love to illustrate "white point", but this is completely impossible within our context. I have no idea what "white" means to your browser, on your monitor, or how any given swatch will actually appear to you. The article on Color temperature explains the matter in some detail, but be cautioned that the black-body spectrum is a mere illustration -- essentially a cartoon.

The "kind" of white light used to illuminate a pigment alters the perception of color. Thus, "warmer" light (more reddish, actually cooler in a physics sense) makes pigments appear redder; bluer "white" light makes pigments appear bluer.

There is great concern in the graphics arts community that white point be correctly determined for a given application. Thus, two entirely different pigments may be chosen to display a company's logo on a billboard and in a magazine intended to be read at home. The billboard, viewed by daylight, and the magazine ad, viewed by incandescent light, may both appear to the eye to present the same color -- but if you were to lay one on top of the other in any one light, you would see they were quite different. (!)

Halftones

Halftone-demo.png
Color-halftone-demo.png

No discussion of printing would be complete without mention of halftones. This is seen clearly in commmon, "black and white" newspaper photos. Obviously, the ink is itself black; the paper white. How to illustrate grays? The printing plate is so made as to lay down the black ink in a pattern of tiny dots. Bigger dots mean a darker gray.

Again, this is a pure illusion. The ink is always black, no matter how distributed on the page. The imperfection of the human eye -- the inability to perceive tiny objects -- is turned into a virtue. Provided one does not put one's nose directly on the paper, one's eye is fooled into imagining shades of gray.

Color-no-halftone-demo.png

Halftoning is extended to the printing of color, and for the same reasons. A solid block of magenta might appear very bright, but actually it is quite a ways towards black -- it has absorbed much light. A color photo printed without halftoning would appear very dark.

To further extend the gamut, halftoning allows the press to put down only a fraction of the ink of a solid blob. The process is complicated by the fact that now 4 inks are being put on the paper at one time. It has been found that the best appearance is achieved when the dot patterns are set at precise angles to one another, so that the overlap of dots of different colors is tightly controlled.

You see, by now, that color in mass market publications is a double illusion: carefully-blended process inks and tiny halftone dots combine to fool the viewer into believing he sees far more than he does.

Screens

Taking as a whole the class of color display screens -- televisions, desktop computer monitors, laptop and cell phone screens -- there is simply no way to say what users are actually seeing.

Each device has a standard, which may be followed more or less faithfully during manufacturing, and which differs from the next model; each item in use ages and is calibrated (or otherwise meddled with) by the user. Each device will display color a different way. Also, the ambient light will affect color perception.

Taking the last two points together, it is a simple absurdity to equate the display of some color on all screens with any display on all printed matter. Considerable money and effort is spent to match colors among specific screens and printing technologies, under controlled and assumed illumination. These are known by the technical users themselves to be poor approximations, of limited validity even under carefully-controlled conditions -- and of almost no value whatever when that control is lost.

Language and Human Nature

Up to now, we've been pretty technical, and we've already called into question the meaning of the word "yellow". But when we extend ourselves, and walk away from the optics lab and the printshop, we see color is even more ambiguous.

Spanish -- if I understand correctly -- has no word for Orange (color). There are only reds and yellows. Some languages use the same word to mean both blue and green; even gray may be included.

The Spanish language has the word "naranja" which means "orange" (the colour and the fruit). --surueña 17:05:57, 2005-08-08 (UTC)
Well, I don't know everything -- I'm especially weak on foreign languages, like most Americans. I once had a Mexican girlfriend who made such a statement, but she'd been in El Norte a long time. — Xiongtalk* 04:14, 2005 August 9 (UTC)

In the American language, there are a great number of words for colors. Some of them have existed since antiquity; others are more recent. A vast number of color names are modern coinages (or repurposings); such are driven by the marketing demands of paint manufacturers and the egos of interior designers. These tend to be entirely relative to the palette of the given manufacturer. For instance, Ebony is a wood, and like all natural materials appears in a range of colors and textures; but it has come to be used as a name for one shade or another of black -- a pure marketing contrivance, for nobody can say, with authority, how Ebony (color) differs from solid black. Many such names are attempts to put a decent face on industrial sterility, such as a "bone" that is merely a dingy white. Crayola, infamously, introduced Flesh in 1949 as a synonym for "skin tone of the ideal 'white' person"; in 1962 they renamed this Peach, which it resembles even less.

Almost all color names are somewhat contrived. Primitive man's visual system developed to identify the items in his environment that were critical to his survival. These were:

  • Sky (that way is up)
  • Dirt (that way is down)
  • Shit (somebody has been here, possible food)
  • Water (for drinking, but don't drown)
  • Forest (safety and some food)
  • Plains (danger and more food)
  • Blood (more danger but possible food)
  • Human genitals (sex)

These give rise to the only important colors:

  • Light Blue
  • Brown
  • Blue-green
  • Green
  • Amber
  • Red
  • Pink

Additionally, of course, many things lack color, being white, black, or gray. Other colors are not found in nature, or found rarely, in small amounts. I suggest primitive man is far more concerned with a plethora of shades of brown, marking different beasts, than in the entire range of purples.

Colors in general are emotional subjects. One might say we are hard-wired to emotional reaction at the sight of certain colors; other colors and blends evoke complex responses because of the combination of stimuli. The Lüscher color test analyzes personalities based solely on indicated preferences among swatches of standardized colors. Different people and different cultures assign different values to various colors, but for all, color is a hot topic.

In this light, it seems obscene to kill color, to dissect it out into unrelated bits, to presume to encompass all color by technical definition.

Summary

I only wish I had spent perhaps less time on technical exposition than on humanistic; for that is where my heart lies. "Man is the measure of all things..." and for a subject as integral to the human experience as color this is doubly true. I hope my readers will note that the thrust of the technical exposition is to show that color theory is not even a sandcastle built on the beach; it is a mirage, a shimmering curtain in the sky, a house of mirrors, a wizard's trick, a lover's sigh.

I stand here opposed to the entire direction taken by WikiProject Color. I blame nobody; I admit I, too, began my involvement here by questioning CMYK normalization -- a petty technical point and one subject to endless debate. For a technical person, it seems safe and sane to state, "Red is 100% R, 0% G, and 0% B in the RGB color space." Like so many technical definitions, it is absolute and inarguable, except in the most petty ways. And I admit that I, too, am a technician.

I resolve to overhaul the group of named color pages, as best as I'm able in the weeks to come. I don't promise haste or perfection, and I certainly don't mean to exclude other contributors. But it is time to toss false technical perfection in the trash and replace it with discussions of colors appropriate to readers who are merely human.

Xiongtalk* 03:59, 2005 August 8 (UTC)

Red (color)

Red (color).jpg This is what I mean. — Xiongtalk* 23:20, 2005 August 15 (UTC)

I haven,t really read through the entirety of the above summary, but I think this structure would do great as a basis for the color article. The montage idea sounds fascinating too. Circeus 23:51, August 15, 2005 (UTC)

Every object in that image might be called red by a fair fraction of people. It's not just the image I'm trying to promote; it is the holistic approach to color. I don't think it helps much to crank up the red gun on a reader's monitor, cut off the other two, and say: "This is red -- the red." Going further, to mathematically define it, is very narrow. I admit, it made sense to me at first -- I, too, am a technical person. But Van Gogh would set fire to any book that made such a statement. — Xiongtalk* 08:26, 2005 September 1 (UTC)

I agree: "red" is not a single color. I had been trying to get interest in discussing the proposal to remove CMYK from the color boxes, because it is misleading at best, and only reinforces the wrong idea that a fixed conversion from RGB to CMYK does exist. I have felt this is putting something for the sake of it, rather than understanding the principles. (I have failed completely to engage any discussion. Shall I take silence as indicating I should do it?)
But your discussion makes me feel that does not go far enough. What these color boxes do is try to say that red or cyan is just one color. But indeed, this doesn't reflect the popular view, or reflect any scientific principle. Just an entry in some arbtrary list, for a particular purpose. Someone just picked a color. So I propose the boxes all be removed from all articles. Comments? Perhaps there needs to be, if there already isn't, an article reflecting the popular use of color names in web design, where this stuff (less CMYK of course) belongs. Silence = agreement? Notinasnaid 10:26, 1 September 2005 (UTC)

Patience

Well, silence does not imply assent. The fact, to which I'm trying to be sensitive, is that a lot of editing time has been invested in a certain view of color, which includes a myriad of distinct color articles, each with its single-shade swatch. I started my membership in this WikiProject in basic agreement with this view; my first effort was to refine {{web colors}} -- a similar, formal treatment of color that assumes there is exactly one definition for a certain word.

Indeed, in the case of this template, this view is quite correct -- HTML keyword "red" really does mean "turn on your red gun full and the other two off." It doesn't even matter what color is actually generated onscreen -- to descend to another level of detail. The markup used to generate the mini-swatches constitutes a tautology: style="background:red; " -- that is, the web color keyword "red" is illustrated by whatever the user's browser generates when given that word.

Imagine a Martian whose screen displays -- as a matter of course -- everything from short microwaves up to the ultraviolet (because the thinner Martian atmosphere encouraged evolution to bless him with a visual range that wide). Marvy's browser will undoubtedly render the "red" keyword in the short infrared -- invisible to human eyes, though perhaps warm to the touch. Or imagine the Centaur, who -- living at the bottom of a Jovian sea of thick atmosphere -- has adapted to the almost total lack of light in what we think of as the visible range. His browser might vibrate to the touch of his hoof whenever a word was tagged as red -- and indeed, this template would produce the same response. So it is perfect -- as far as that goes.

For me, the trouble began with the text links in that template, just below the mini-swatches. For technical reasons, the "green" gun in a standard CRT excites phosphors that glow with a color best described (to my way of thinking) as acid green. It is a fairly unattractive color, certainly far more vivid than most green shades found in nature or used as pigments by man. It is a poor representative of the family of greens. But it is -- by definition -- "Green" within the world of cathode ray tube technology. W3C thinks otherwise, and named that shade -- that condition of turning that gun on full -- "lime".

Actually, the matter gets even murkier; the most official standards are the least specific. W3C standards are shot through with "browser-specific implementation" and "exact rendering left to the discretion of the user agent".

Ried La Wantzenau 4-8-2002.jpg

But long before I got that far, I started to question the whole basis of our thinking about color in this Project. "Green" is not a concept defined by a single shade of green, or a standard controlled by a bunch of Belgians. It's an analog world -- and like most real things, "green" has no sharp boundaries, only a locus of values. Near the locus, some shades are widely accepted; further out, there is disagreement and local usage.

The whole business of CMYK conversion is so intensely technical I wonder if we can possibly treat it properly within the bounds of a general reference work. Are we in some sort of competition with Wikipedia:WikiProject Mathematics for the greatest number of most impenetrable articles? I was not joking when I said this topic was one of the fastest ways to start a barfight among publishing professionals. It is even more explosive than orphans, kerning, consecutive hyphens, bleeds, and the proper choices among different sizes of dash.

I'm tempted to assert that all CMYK process colors are frauds. There is no such thing as a process blue -- it simply does not exist. Different processes -- and there are many -- use different algorithms for covering paper and other media with patterns of dots of various process inks, and some of these produce illusions of various shades of blue. But this is more of the pettifogging I'd like to get away from. A guy standing in line at the supermarket says, "Hey, that J-Lo sure does look tasty in that blue bikini" -- and to him, that scrap of fabric is blue, never mind that the flesh and the suit she is not quite wearing are both pure fiction. So, yes, no harm to say this-and-that is a kind of process blue. We professionals know there is no such thing, and if there were, then there would be a dozen different ways to get there from here.

I think we really need to face up to the limitations of our chosen medium: GFDL-licensed wiki. We have to assume, primarily, that our content will be viewed through a web browser on any one of a thousand different makes and models of screen, using four or five major families of technology, and under completely uncontrolled lighting conditions. Beyond that, we need to allow for the possibility that somebody will actually go out and do the impossible, and print the damn thing on actual paper -- what paper, we don't know; what inks, we don't know; and we certainly don't know whether the work will be viewed by candlelight, daylight, or sputtering carbon arc.

So we cannot hope to present exact samples or exact definitions of any color. We can realistically talk about what we mean when, while using a given technology, we define a certain color using a certain algorithm and some specific numbers. But what it all boils down to, in the end, is a numbered swatch out of a swatch book. Nobody but an accountant says, "Our logo should be Chinese Red."

It's a Good Thing that we discuss the many meanings of "green", however much they vary and contradict one another. It's quite reasonable to illustrate colors, showing small portions of the great variety of shades that might be brought to mind by the word "Green". It's certainly a public utility to list commonly-used numeric representations of some standard shades of green; I think we should both expand our range and restrict ourselves to a limited number of practical applications -- for instance, Web colors, or Pantone Process. (And yes, there's more than one of those.)

But I've come slowly around to the thought that it's just plain silly for us to try to define any color here.

It may take some time before others get to the same point. Right? — Xiongtalk* 12:40, 2005 September 2 (UTC)

CMYK + Disclaimer?

I think it is probably nuts not to have *some* kind of CMYK colour listed here. People will come to the colour pages (I did once...) to try and find a colour.

I think that there is a legitimate controversy over CMYK that is well debated here.

So... I suggest that we put in a reasonable value for the major colours, say in the major format (US Web Coated SWOP or whatever the current photoshop default is) with a disclaimer or 'gotcha' linking to a page explaining the problems with the CMYK values as a definitive source.

There must be some more or less agreed values for the major colours somewhere?