Wikipedia talk:WikiProject Color/Archive 7

Green
Green
Spectral coordinates
Wavelength	520–570 nm
Frequency	~575–525 THz
Common connotations
	nature, growth, hope, youth, sickness, health, Islam, spring, Saint Patrick's Day, money (US), and envy
Color coordinates
Hex triplet	#008000 (HTML/CSS); #00FF00 (X11)
sRGBB (r, g, b)	(Lua error in Module:Color at line 24: Invalid hexadecimal color 008000 (HTML/CSS)<br />#00FF00 (X11).)
HSV (h, s, v)	(Lua error in Module:Color at line 24: Invalid hexadecimal color 008000 (HTML/CSS)<br />#00FF00 (X11).)
CIELChuv (L, C, h)	(Lua error in Module:Color at line 24: Invalid hexadecimal color 008000 (HTML/CSS)<br />#00FF00 (X11).)
Source	HTML/CSS; X11 color names
	B: Normalized to [0–255] (byte);

This is an archive of past discussions about Wikipedia:WikiProject Color. Do not edit the contents of this page. If you wish to start a new discussion or revive an old one, please do so on the current talk page.

Archive 1

←

Green info box looks blue.

The green chosen for the info box looks very blue to me, at least on my calibrated displays. I saw the PDF linked but am wondering why we should use that color for the main sample? To be honest, I still would like to move the color info boxes away from being about a specific color coordinate to being about the abstract color. PaleAqua (talk) 04:55, 17 May 2010 (UTC)

Well, we should get *some* kind of reasonable evidence before putting up any kind of hex coordinates or specific colors. A better thing might be to track down a copy of the 1976 book Color Universal Color Language and Dictionary of Names put out by the US National Bureau of Standards, and go by the most saturated in-gamut approximation at the same lightness/hue as whatever they define for each color term. Alternately, we could base the colors chosen on linguistic/anthropological research – there might be something useful on Paul Kay’s page. Instead, I used the CIECAM02 hue angle defined by that standard to be "unique green", along with a somewhat arbitrary lightness (I don't think there’s necessarily any reason to priviledge one lightness or another, but we could have a discussion about that). Either way, #00ff00 and friends cannot be acceptably labeled "green", etc., in my opinion. The sRGB specification is about making a useful color gamut for a high definition television standard, and has pretty much nothing to do with human color names, and the X11 color names were designed for displays with rather different characteristics than sRGB, by computer scientists making things up as they went along. The fact that these names are used on the web does not make them relevant to typical human color definitions. –jacobolus (t) 05:19, 17 May 2010 (UTC)

By the way, whether your display is characterized or not won’t make much difference, because web browsers do no color management of html/css colors; everyone is going to pretty much see a different color, even if all their displays are properly characterized. You need to take a screenshot and apply the sRGB color profile to it, to see what the color looks like in sRGB. –jacobolus (t) 05:21, 17 May 2010 (UTC)

Actually wasn't looking at the green in the browser but in a colorspace tool set to an sRGB profile, though it looks similarly bluish in my browsers as well. Just like #00FF00 is a bit yellowish to me. I think that the choice of any specific coordinate is wrong and other then giving approximate frequency / wavelength ranges we shouldn't go with any coordinates in the info boxes. I don't see how using the hue angle from CIECAM02 to create RGB coordinates is any better than using X11 colors. I'd rather actually add an entry with the reference 164.25 hue angle etc... PaleAqua (talk) 07:20, 17 May 2010 (UTC)

Okay; Part of that probably has to do with the quite yellow-green colors shown immediately below influencing the color appearance. Another part could be that there's extreme observer-to-observer variation in choice of "unique green": it’s unique for each person, but each person has his/her own. I'm going to try to write a bit of summary of what research I’ve read about that at unique hues, but it hasn’t happened yet. –jacobolus (t) 23:47, 19 May 2010 (UTC)

Have you ever taken a prism and focused a beam of sunlight through it? The color in the green part of the spectrum looks exactly like X11 Green. Also, if you check the color coordinates on the CIE chromaticity diagram on the outer edge just inside the horseshoe shape in the central part of the green part of the spectrum with your computer's digital color meter, you will see that it registers as the color #00FF00. The color you are attempting to foist upon us as "green" is really a medium saturated bluish green similar to the color emerald. Keraunos (talk) 05:47, 17 May 2010 (UTC)

I again changed red and blue back to their normal coordinates. I'm leaving green unchanged for now because this is where the discussion is going on. Keraunos (talk) 06:44, 17 May 2010 (UTC)

Yes, I have split white light through prisms before. You’ll notice that every hue between blue-violet and orangish-red appears somewhere in the resulting "spectrum".... there’s nothing particularly special about one yellowish-green hue as compared to another green hue. –jacobolus (t) 01:53, 19 May 2010 (UTC)

Just because it appears bluish doesn't mean that it's not _a_ green. I'm actually opposed to using the X11 colors as definitive / normal colors as well. PaleAqua (talk) 07:20, 17 May 2010 (UTC)

Yes, it is a green just slightly more than it is a blue, but almost any average Wikipedia user viewing the article will not recognize it as "green". When I view it out of focus, it actually looks more blue than green (I believe that would be the case even if I viewed it on the three different model CRT monitors on my workstation or on the color corrected monitor on my Mac instead of just on my laptop). Whatever value is used, it should at least appear to be distinctly green. It would be better to not display any color instead of displaying the current value. The pink that was suggested for red was also so far from red that I'd also support displaying nothing instead. The blue suggested for the blue article might be okay if it was darker. I agree the actual hex values need not be the X11 values. VMS Mosaic (talk) 09:06, 17 May 2010 (UTC)

What about using the HTML/CSS hex value at least for the infobox title? It neither appears too yellow (which I agree the X11 value does) nor too blue. VMS Mosaic (talk) 20:58, 17 May 2010 (UTC)

That was one we used last time. Actually we had both the colors in the title as a compromise. PaleAqua (talk) 23:19, 17 May 2010 (UTC)

It’s definitely too yellow. –jacobolus (t) 00:46, 19 May 2010 (UTC)

X11 green is NOT too yellow. It is just right. X11 green is a representation of the spectrum color halfway between the spectrum colors Chartreuse green and spring green. HTML/CSS green and CIECAM02 green are too dark to be spectrum colors as you can see if you focus a beam of sunlight through a prism. Keraunos (talk)18:50, 20 May 2010 (UTC)

From a human visual perception standpoint, it is too yellow as displayed on LCD and CRT RGB displays. Humans do not perceive colors based on precise mathematical hex values. The original web developers at least realized that is particulary true for green (probably has something to do with the human eye being more sensitive to green). That's why the web and Wikipedia internally do not use the X11 value for green, but instead use the HTML/CSS value. That is the value which a web designer/user expects to see if he specifies green in his HTML. VMS Mosaic (talk) 21:44, 20 May 2010 (UTC)

You can see that the green light in the photograph at right of the additive colors showing the combination of red, green, and blue lights is synonomous with X11 green. Keraunos (talk) 21:08, 20 May 2010 (UTC)

For a long time--for more than a year until two and a half weeks ago--the title area at the top of the green color box was left blank (white) as a compromise because people had kept changing the color back and forth from X11 green to HTML/CSS green. The color squares colored various tones of green in the image within the color box were used to represent green. The values for both X11 green and HTML/CSS green were displayed under the image in the color box. Keraunos (talk) 04:25, 18 May 2010 (UTC)

I suggest that we go back to this original color box as a compromise. Keraunos (talk) 04:42, 18 May 2010 (UTC)

If you’re going to include the box that way, I suggest removing the “hex triplet”. –jacobolus (t) 00:46, 19 May 2010 (UTC)

World color survey

To see proper color in this image, if you are using Safari & possibly recent Firefox, just open up the full-width image; otherwise, download it and open it in a program such as Photoshop which does proper color management. Wikipedia’s thumbnail renderer strips color profiles from images.

Okay, here above I’ve plotted the brightest possible colors within the sRGB gamut for each of 40 munsell system hues, and integer values between 2 and 9, like those used in the world color survey. I put white dots where there are dots in this plot taken from Sturges & Whitfield 1995, and black dots near where the X11/CSS color names put red/yellow/green/blue. As can clearly be seen, the X11/CSS color "red" is not too far from the hue of "unique red" or "focal red", but the X11 color blue is too purple, the X11 color green is too yellow, and the X11 color yellow is slightly too green. I have no particular opinion on the proper lightness at which to display each color; that should probably depend on doing some more research. I haven’t read too many papers/books/etc. about color naming, and so am admittedly no particular expert at it. I don’t think we can legitimately leave the X11/CSS colors as the representatives for these color names though. Their sources are "reliable sources" only as concerns the web, but not as concerns color naming (they were computer engineers just making something up quickly, off the top of their heads). –jacobolus (t) 01:41, 19 May 2010 (UTC)

Interesting chart. I agree that the X11 color names need to be de-emphasized. PaleAqua (talk) 03:00, 19 May 2010 (UTC)

Where is HTML green on the chart? Does the image have an ICC tag? If so, then Firefox with the Color Management extension can display the image using a default sRGB profile (it can use a monitor specific profile if one is available). It is true the X11 colors were the work of mostly clueless grad students, but a little more thought went into the HTML/CSS colors which is why HTML green is better than X11 green. I agree that the web color definitions are not the best, but we also must avoid WP:OR. What options do we have? VMS Mosaic (talk) 03:37, 19 May 2010 (UTC)

The HTML green is directly below the X11 green; it’s still completely arbitrary. I’ll try to expand this chart and make it a bit finer grained (100 hues maybe, and half-integer value steps), so it's easier to pinpoint just where different definitional colors are, and then I’ll try to highlight where various colors are (additive and subtractive primaries, "focal" colors, etc.). Basically though the X11/CSS colors are based on the chromaticities of the additive "primaries" used in computer displays, which, though they are named "red"/"green"/"blue", are not designed as representatives of those color names from a perceptual or linguistic perspective. Sure, whoever was making the names for HTML colors decided that #00ff00 was too light to be considered "green", but that doesn’t mean any linguistic or psychophysical research went into choosing #008000 instead: –jacobolus (t) 23:43, 19 May 2010 (UTC)

RGB color system used in Wikipedia

The whole color system on which the colors displayed in Wikipedia are based is really simple and elegant. The color system used in Wikipedia to display colors is based on the RGB color wheel which is the physical expression of RGB color space. There are 12 major colors of the RGB color wheel at intervals of 30 degrees. I have listed these 12 major colors below. Any color can be made by combining one of these 12 colors with each other or with an achromatic color such as black, gray, or white. Keraunos (talk) 18:39, 20 May 2010 (UTC)

On the color sphere, by analogy with the globe of Earth, the color white is at the north pole, the color black is at the south pole, the color grey is at the center of Earth, and the 12 major colors of the color wheel (the 12 basic hues) are spaced equally around the equator. Thus, all other colors can be made from mixtures of these 15 colors.

Keraunos (talk) 18:39, 20 May 2010 (UTC)

In an RGB device, by definition, all in-gamut colors can be produced by the additive mixture of three colors. No need for the other nine. –jacobolus (t) 18:22, 21 May 2010 (UTC)

Of course, I know that only three colors are needed. The other colors are there for aesthetic purposes and to define the relationships between the primary three and the other nine. Keraunos (talk) 03:52, 24 May 2010 (UTC)

The 12 major colors of the color wheel

The 12 major colors of the color wheel, at 30 degree intervals on the HSV color wheel (RGB color wheel) are the following: red (Color #FF0000 (0 degrees or 360 degrees), orange (Color #FF7F00) (30 degrees), yellow (Color#FFFF00) (60 degrees), chartreuse green (Color#7FFF00) (90 degrees), green (Color#00FF00) (120 degrees), spring green (Color#00FF7F) (150 degrees), cyan (Color#00FFFF) (180 degrees), azure (Color#007FFF) (210 degrees), blue (Color#0000FF) (240 degrees), violet (Color#7F00FF) (270 degrees), magenta (Color#FF00FF) (300 degrees), and rose (Color#FF007F) (330 degrees). This constitutes the complete set of primary, secondary, and tertiary color names. Keraunos (talk) 18:39, 20 May 2010 (UTC)

The hsv values of the colors displayed on Wikipedia are based on this system, notably the h (hue) value of each color is based on its position on the RGB color wheel or its place within or on the surface of the color sphere in relation to the 12 major colors on the equator of the color sphere (the h (hue) value of an achromatic color on the axis of the color sphere connecting the north and south poles is left blank because achromatic colors have no hue).

Color systems other than the RGB color system

If Jacobolus wants to set up an alternate color system, then he can create a Wikipedia article about it and include the chart below within the article and give us an explanation of how his alternate color system would work and his sources. But there is no need to disturb the present Wikipedia color system outlined above which is so simple, elegant, and serviceable. Keraunos (talk) 18:39, 20 May 2010 (UTC)

There are apparently already Wikipedia articles on the alternate color system Jacobolus is talking about--unique hues and Natural Color System. This system is based on what are called the four psychological primaries (red, yellow, green, and blue) rather than the three additive primaries, (red, green, and blue) of RGB color space color system that is presently used by Wikipedia, is the basis of the web colors, and which is displayed on the RGB color wheel above which I downloaded. All Jacobolus has to do is to download the above chart into one of those two articles (unique hues probably would be best, since that is a short article that needs expansion) and elaborate on his new color system, more fully explaining it and adding his sources. Keraunos (talk) 19:01, 20 May 2010 (UTC)

But trying to mix the RGB color system, the psychological primaries system ("Natural color system") or the old RYB color model together into a single system is like comparing apples, oranges and cherries. Keraunos (talk) 19:09, 20 May 2010 (UTC)

RYGB color system (psychological primaries)

The problem, Keraunos, is that human color terms are not based on additive color mixing. The red end of the spectrum is not fully “red”, but a sort-of orangish red, and the blue end is quite a bit past blue, into “blue-violet” or “violet”. The first additive color models used orangish-red, blue, and yellow lights, but at some point in the 19th century (I’ll try to track down and quote some sources sometime soonish) it was figured out that orangish-red, yellowish-green, and blue-violet light sources produced a wider color gamut. In the 1922 Optical Society of America Report on Colorimetry, these three additive primaries are labeled “red”, “green”, and “violet”, but many of the scientists working on colorimetry called them “red”, “green”, and “blue” instead.... it was never intended that these additive primary colors would embody/represent those color names: the names were just convenient simple labels for the light sources. Making up an RGB model with fancy stars and explicitly named secondary colors and so forth is fine, as long as it is remembered that the names are merely suggestive, rather than definitive. Unfortunately, Wikipedia currently implies the latter, when it shouldn’t. If any coordinates, in sRGB space or otherwise, are to be given, they should be based on some kind of authoritative color name standards (for example, the ISCC–NBS system), rather than on X11 or CSS colors, which were just made up on the spot for the convenience of computer users. –jacobolus (t) 16:03, 21 May 2010 (UTC)

FWIW, the current green (00BA85) hex triplet's closest match in the ISCC–NBS system is Brilliant Bluish Green. There is nothing that is even remotely close to the X11 triplet. The HTML/CSS triplet is reasonably close to Vivid Green. VMS Mosaic (talk) 21:12, 21 May 2010 (UTC)

Where are you getting your ISCC–NBS system values from? Is it a reliable conversion? Do you have the Munsell System coordinates for those? According to this copy of Agoston’s 1987 Color Theory and its Application in Art and Design, the ISCC–NBS color “vivid green” is No. 139, with hue range (by the Munsell renotations) 3G–9G, value range 0–10, and chroma 11+. For what it’s worth, The X11 “green” is at hue is between 9GY and 9.5GY, which is substantially yellower. The best ISCC-NBS name (again, according to my quick skim of the table in Agoston) is No. 129, “Vivid Yellowish-Green”. –jacobolus (t) 23:51, 21 May 2010 (UTC)

From link 1 (includes original and improved values) and link 2. I'd agree that 129 is the closest to the X11 color, but it is not yellow enough. VMS Mosaic (talk) 00:44, 22 May 2010 (UTC)

Your link 1 is broken for me, and your link 2 does not cite its source for the ISCC-NBS values (I believe the original 1955 ISCC-NBS coordinates were relative to the 1929 Munsell Book of Color, rather than to the 1943 OSA renotations, but I’m not completely sure of that), and does not explain its methodology in converting Munsell System coordinates to RGB hex triplets. I wouldn’t necessarily trust it. (The renotations are specified relative to illuminant C, while sRGB uses illuminant D₆₅ as its white point; some kind of chromatic adaptation is required, and the current state of the art seems to be CAT02 (“chromatic adaptation transform”). It’s not clear whether any of your sources do chromatic adaptation, or what RGB primary chromaticities they use, or what RGB white point, or what gamma function, if any. )–jacobolus (t) 01:03, 22 May 2010 (UTC)

I’m going to actually read the chapter of Agoston about the ISCC–NBS system, and report back. :-) If you’re interested, I can try to scan it and email you a copy. (Also feel free to email me for other sources; I have access to a number of books and papers at hand in this library.) –jacobolus (t) 01:12, 22 May 2010 (UTC)

Sorry, I forgot to test them; I put an 'l' on the end of the first link. Fixed. That link includes a fair amount of info about the origins of the data. VMS Mosaic (talk) 01:20, 22 May 2010 (UTC)

Okay, the “Mundie RGB” numbers clearly did no chromatic adaptation, and then furthermore are based on an RGB different than sRGB. As for the Gretag program converted colors, it’s hard to tell, since the assumptions behind the conversion aren’t described. I’m going to try to convert these from Munsell renotation Å® sRGB coordinates myself – I’ll carefully document my methodology – and see how well my versions agree with those. –jacobolus (t) 02:21, 22 May 2010 (UTC)

By the way, the Kelly/Judd NBS publication, Color: Universal Language and Dictionary of Names which is the current authoritative source for the ISCC–NBS system is available online here. –jacobolus (t) 02:10, 22 May 2010 (UTC)

Discussion about the color boxes

Over the coming weeks, I’m going to try to build and put online a javascript color conversion tool that will hopefully provide a useful interface for transforming colors coordinates from one system to another; I think it should be possible to stick in values in whatever space, and get out values in CIELAB (D₅₀), CIECAM02, sRGB, Munsell renotation coordinates, and maybe ISCC–NBS name. Maybe we could even work up a better color infobox template, and set it up to spit out mediawiki code... I’ve long been unhappy with the current infoboxes. No promises, &c., but does anyone have thoughts about the idea? –jacobolus (t) 03:00, 22 May 2010 (UTC)

I've long thought about replacing the current infobox with something that is not so web color focused. I've set up a sandbox for the infobox but don't have much free time right now. Basically I'd like to see the infobox focus more on the color term and less on specific colors. Ideas for sections that I could see included:

Titlebar ( Would list the name of the color/article, but would not be colored - I think White background with Black text )
Picture ( either a bunch of objects associated with the color, or the color grids we use currently )
Basic colors, ( e.g. Teal might have "Green and Blue" listed as it's basic colors )
Spectral coordinates ( only for basic colors - though I guess complementary wavelengths / frequencies make sense )
Color space map ( show the the CIE color space or something similar with the approximate area of the color marked )
Standardized coordinates ( Only for use for colors such as UNECE Amber, would basically be a mostly free-form section that could hold however the color was specified )
Origin of the term ( perhaps with a link to wiktionary )
Dyes and pigments ( would link to articles on dyes and pigments associated with the color)
webcolors ( Samples of the various webcolors related to the color term, my opinion is that the web colors should be shown without coordinates and a link to the list of web colors for those interested in seeing them. )
Non-basic colors ( similar to the web colors, show some samples related to common color terms again sans coordinates )

Finally I've thought about convert some of the "In human culture stuff" ( though I'm not a fan of such trivia ) into subsections. We added symbolism a while back for the infoboxes with it they have grown into overwelming and disorganized list. Things like "In nature" (very short list of plants or animals noted for the color), major sports teams notable for using the color, or religious / liturgical symbolism, etc. PaleAqua (talk) 06:05, 22 May 2010 (UTC)

That sounds pretty decent. There’s no hurry. Where’d you put your sandbox (assuming there's anything there yet)? –jacobolus (t) 00:25, 23 May 2010 (UTC)

It's at Template:Infobox color/sandbox, but the only thing there right now is pretty much a copy of the current info box with some tweaks that I was playing with to see if I could get the template to show what the layout would look like live when previewing changes to the template page itself. See also Template:Infobox color/testcases. PaleAqua (talk) 01:42, 23 May 2010 (UTC)

I disagree completely. There is no use having a color box if it doesn't show the color. That is the whole point of having a color box. Keraunos (talk) 06:19, 22 May 2010 (UTC) I think that the area showing the color should be larger as it is in the Spanish Wikipedia Spanish Wikipedia color box: or the French Wikipedia French Wikipedia color box:. Keraunos (talk) 06:19, 22 May 2010 (UTC) I think the color boxes should look like the French Wikipedia color boxes so there is a larger area in which to display the color and then you can get more of a sense of what the color looks like. Keraunos (talk) 06:38, 22 May 2010 (UTC) I think the color box should look like the upper version in this Infobox testcases page, i.e., with a larger area to display the color: Infobox color test case with larger area in the color box for displaying the color:. This would bring the color boxes in line with the larger color display areas of the Spanish and French Wikipedias. Keraunos (talk) 03:52, 24 May 2010 (UTC) It's best not to make the color box too complicated as you have outlined above. It's better to keep it simple like it is now and put additional information in the article. Keraunos (talk) 06:19, 22 May 2010 (UTC)

As far as major infobox changes, perhaps the discussion should be on the project's talk page or at least have a pointer to the discussion here? VMS Mosaic (talk) 09:54, 22 May 2010 (UTC)

There’d definitely have to be a discussion at the project’s talk page before any planned changes went forward. I’m just floating ideas. :) –jacobolus (t) 00:32, 23 May 2010 (UTC)

Yes, a discussion at the project page would be a good idea. Was just waiting on starting one myself til I had more time. I noticed that the French box uses a grey title bar, which seems like a good choice also. Basically we would leave the title bar neutral and use the samples and pictures to represent the color. PaleAqua (talk) 01:42, 23 May 2010 (UTC)

That is absurd. There has to be a space to display the color itself on the title bar and a place below that to put its hex code and hsv values. Keraunos (talk) 02:43, 24 May 2010 (UTC)

What is absurd? Why the dismissive tone? –jacobolus (t) 04:21, 24 May 2010 (UTC)

The purpose of having a color box is in order to display the color. There is no use having a color box if it doesn't display the color. Keraunos (talk) 04:39, 24 May 2010 (UTC)

No the purpose of an infobox is to give highlights or summary about the subject of an article. And yes for colors an infobox should display a representation of the color as part of that summary, and if you see what I purposed not only does that, but it does it in multiple ways. For most generic color terms giving hex codes and HSV values are misleading (and easy a point of contention) as is cleared demonstrated by the various opinions on what color IS green and NPOV as it gives undo weight to the HTML/X11 colors. PaleAqua (talk) 06:53, 24 May 2010 (UTC)

As long as the infobox displays a representation of the color, then I am happy! All the other representations are icing on the cake. However, as I mentioned above, I think the color display area should be larger as in the top example here: Infobox color test case with larger area in the color box for displaying the color:.

What is the difference between CIECAM02 colors and psychological primary colors?

Question: Jacobolus, why is it that the psychological primary color and the CIECAM02 color is the same in the case of yellow, but the psychological primary color and the CIECAM02 colors are different in the cases of green, red, and blue?

Green: Psychological Primary/NCS vs. CIECAM02
Yellow:Psychological Primary/NCS vs. CIECAM02
Red: Psychological Primary/NCS vs. CIECAM02
Blue: Psychological Primary/NCS vs. CIECAM02

I was just wondering about this. Keraunos (talk) 06:22, 22 May 2010 (UTC)

[I hope you don’t mind, Keraunos, that I put smaller color swatches in your comment and removed the section breaks, to make it easier to follow the discussion.] I chose the lightness somewhat arbitrarily, not knowing what is the proper lightness to choose; I just wanted to get the hue right. It’s quite possible that darker red/green/blue would be appropriate. –jacobolus (t) 00:24, 23 May 2010 (UTC)

The Psychological Primary/NCS colors look more correct than the CIECAM02 colors. Darker might make a difference, but wikipedia editors selecting a proper lightness would be WP:OR. We need to either pick a pre-defined set of hex value/color names to display in the infobox titles or display no color at all. I believe we need to display a color in some manner; for example, a user going to the Baby blue page would be dumbfounded to not see the color. VMS Mosaic (talk) 08:57, 23 May 2010 (UTC)

I'm also not convinced that (in sRGB) the “psychological primaries” actually reflect the colors of NCS – according to the image they came from, they were picked by eye-balling a monitor and a reference color at the same time, and twiddling RGB until the two looked the same: any difference in lighting from controlled conditions, or any divergence of display from sRGB or any number of software effects could render that method substantially inaccurate. I wish I knew where to find some kind of better colorimetric specification for NCS colors. –jacobolus (t) 04:15, 24 May 2010 (UTC)

Alright everyone, I spent some time figuring out how to reasonably interpolate Munsell colors, and made up a bigger image showing the color at max chroma for each hue and value (my code is horribly un-optimized, so it took my computer about a half hour to render.. eep). Hopefully this helps demonstrate why I wasn’t happy with using X11/CSS values for these color names:

The “cross-cultural foci” are from MacLaury 1997, and were based on color chips at 40 hues, so they're a pretty rough guess (that’s 25 pixels, horizontally). The location of the NCS unique hues is taken from the Agoston book, and were to the nearest 1 Munsell hue unit (10 pixels). The other points are accurate to within a couple pixels. –jacobolus (t) 00:43, 23 May 2010 (UTC)

The main issue I see is you used the ISCC-NBS "strongs" instead of the "vivids". The “cross-cultural foci” green and red are respectively too blue and too violet. The ISCC-NBS green box contains nothing that isn't a blue-green. If I had to pick, it would be the additive RGB (sRGB) colors except for green which would be the HTML/CSS color, although red would still be too orange (using Vivid Red might provide a more suitable red). Are we trying to pick colors that look correct on color managed (gamma, etc.) displays or on the typical user's display? VMS Mosaic (talk) 08:57, 23 May 2010 (UTC)

For each Munsell hue and value, the color with chroma 6 if if fits in the sRGB gamut, or else the in-gamut color with maximum chroma.

The “vivids” are defined to be any color with chroma above the boundary for strong, at any lightness whatsoever. Most “vivids” are outside of the sRGB gamut, and they have no obvious way to define a “center”. The “vivids” fall into the same hue ranges as the “strongs” here. As for “containing nothing that isn’t a blue-green”... One possibility is that your color definition has been dramatically biased by your familiarity with X11/HTML/CSS color names, or by looking at 00ff00 (etc.) on computer screens; another possibility is that your “unique green” is different from mine: there is extensive experimental evidence that definitions of unique hues are highly variable from one observer to another, though for each observer they remain consistent across measurements/over time. The NCS definitions were based on the average of many observers. Part of the problem may be that middle-green is not possible to produce with very high chroma in the sRGB gamut (RGB devices are in general have quite a limited gamut in greens, blue-greens, and yellows, while producing vivid colors for hues in the yellow-green–green-yellow range, and the whole range from orange-red to blue-violet). Perhaps a more useful image here would be one which instead of showing the maximally-chromatic in-gamut color at each point instead limited its highest chroma to 6 or 8 (such as the image to the right). –jacobolus (t) 19:44, 23 May 2010 (UTC)

All I have to go on is what my LCD and CRTs display versus what I see in the world around me. While out riding and shopping today, I tried to pay attention to color. I was actually a little surprised to find that most consumer packaging and products like pens, plastic cups, etc. use very nearly the same hue of blue which also happens to be what I consider blue (perhaps due to cultural/consumer training?). On the other hand, most of the red hues I saw appear a little too orange for my taste (perhaps a little orange makes them more eye catching?). The greens were all over the place (blue-green to yellow-green), but when I look out my window at the glass and trees, I see nothing that falls inside the ISCC-NBS green box as viewed on my displays. I see only green and yellow-green hues. I tend to agree that it may be a limited gamut issue, but that brings me back to asking what the goal is. Is it to display something which the average wikipedia user will see as being reasonably correct when viewed on a typical display? VMS Mosaic (talk) 00:20, 24 May 2010 (UTC)

6G 5 / 7.61 (#008B66)
#008B66

What I meant w/ the gamut issue was that perhaps the nearby very colorful yellowish-green was making the much grayer green appear blue by comparison; do those colors still seem blue when viewed in isolation, or when they’re placed next to similarly-chromatic yellow-greens, as in the duller picture to the right? I just added a black square around the ISCC–NBS "green" in this image too. Actually, here’s just a green at Munsell 6G 5 / 7.61 (according to Agoston the hue of NCS unique green), by itself, to the right. You sure that doesn’t look “green”? –jacobolus (t) 00:40, 24 May 2010 (UTC)

On three CRTs, it did appear to be a green hue (although toward the blue side of the green hue range), however, on an LCD at another location under different lighting, it appears as a blueish-green. In other words, it does not appear to be near the center of the green hue range. I tried the slit technique (i.e., sliding a small gap between two sheets of paper both vertically and horizontally across the screen) on the two images with black boxes. What I perceive as "actual" green occurs on a short diagonal running from lower left to upper right in both images. In the second image, the upper end of the diagonal stops at the lower left corner of the black box, so a small part of the corner does look green. BTW, the slit technique is harder to do on an LCD, because any pressure on the screen will disrupt the colors. VMS Mosaic (talk) 03:26, 24 May 2010 (UTC)

(a) is your display properly (or at least roughly) characterized? and (b) did you open up these images in a program which does color management? It’s possible or even likely that your display gamut & gamma are not precisely sRGB-shaped, which would potentially make quite a difference to fine hue judgements. If yes to both of those, you shouldn't be seeing any hue changes along verticals (that is, shouldn't see constant hue along a diagonal), as the colors in each vertical strip are (with quite a bit of "normal color vision observer" evidence/data to back this up) of identical or nearly identical perceived hue. It’s possible that you're conflating chroma differences with hue differences or something, but if you're really seeing a hue difference it's quite likely that your monitor either has somewhat different primaries than sRGB, or is set to use 1.8 gamma (Mac OS X before 10.6, for instance), or something of that sort. –jacobolus (t) 04:12, 24 May 2010 (UTC)

I had to impose on someone in another department with an iMac running OS X 10.6, in order to get a fully color managed environment. In Safari, I saw pretty much what I see elsewhere even using an sRGB profile in place of the default iMac profile. Yes, there are hue changes across the diagonals, but I also see hue changes on the verticals. X11 green and HTML green are on nearly the same vertical, but are not the same visually perceptible hue. The HTML green, being slightly to the left (if you marked it correctly), should be slightly more yellow, but appears less yellow than the X11 green. Admittedly, I only had a couple minutes access to the iMac. When I look at any of the first image's additive RGB primaries on my CRTs, a constant hue line with value light to dark drawn thru them using a slit is not vertical. And no, I don't believe it is a chroma difference I am seeing. Was the observer evidence based on reflected (with what type of lighting?) or direct light? Were the observers actually given a wide enough range to pick from (i.e., print gamuts can be very limited)? Etc.?

The iMac owner also agreed the green in the title bar of the green article is a blue-green. So, once more, I ask what is the goal? Is it to display colors which the average user on his average PC perceives as being reasonably close to the color names? VMS Mosaic (talk) 22:37, 24 May 2010 (UTC)

HTML/CSS color, sadly unable to use color management in current browsers.
#008B66

Sorry, what I meant was that, since web browsers do absolutely no color management of CSS/HTML colors, you have to look at these colors either as images (and not wikipedia thumbnails because color profiles are stripped by mediawiki's thumbnail generator) in a browser that color-manages images, or else open them up in an image editor that does color management. So the picture to the right here might look substantially different than the HTML in the page, but exactly how it looks different will depend on the user’s display. The image, assuming the display is properly characterized and the browser does color management (of images with attached profiles), should be approximately “correct”, while the pure HTML version would be “off”. –jacobolus (t) 02:13, 25 May 2010 (UTC)

These two look alike on my browsers and on Safari. I believe the point is that they should look different on Safari (Mac OS X 10.6)? Since the iMac wasn't mine, I only spent a few seconds, along with the owner, looking at them, so perhaps there is some very minor difference we didn't see. VMS Mosaic (talk) 02:31, 26 May 2010 (UTC)

Alright. They look quite different in my Safari (Mac OS X 10.5), because prior to 10.6, OS X used a 1.8 default display gamma. The display I’m looking at now (a 13" MacBook Pro) also has a gamut slightly different than sRGB, so the swatch under color management looks a bit yellower than the other one. –jacobolus (t) 22:49, 26 May 2010 (UTC)

Okay, I spent a bit more time than I intended to, putting more ISCC–NBS categories on this chart:

–jacobolus (t) 03:21, 25 May 2010 (UTC)

I thought this quotation was cute: “Once upon a time, someone chose 16 full and half coordinate RGB primaries as the default color-map for early personal computer video cards, perhaps planning to implement full color using a halftone-like method. Instead, these were used as pure colors, often resulting in garish displays. This bad practice was made worse by using common color names to identify them; worse still by putting these in the Windows color-dictionary; and worse still by codifying these common color names (with RGB primaries) in HTML-3.2 and HTML-4.0, two of the most widely used standards in the world.” [...] “The development of HTML has embraced an ignorance of color science. Will the pervasiveness of HTML change our color lexicon” from <http://people.csail.mit.edu/jaffer/Color/Dictionaries>. –jacobolus (t) 20:20, 23 May 2010 (UTC)

I don't find this comment so cute. And this gives a bad perception of what is HTML or CSS: they clearly state that the named colors are just aliases to colors expressed within the sRGB color gammut, which has a very precise definition. So web standards are clearly defining these colors, independantly of the subjective "perceptieve model". Yes, the browsers still have problems to properly render the sRGB model (because many of them (not all) still don't honor the CIE profiles, but this is not the case of all of them (e.g. Chrome and IE8 are respecting the CIE color model and its profiles). Note that the native support of the sRGB model is mandatory now for all digital monitors, HDTV, MPEG video formats, even if they may also support other CIE-based color models (notably the PAL and NTSC color models that are algorithmically computable from/to the sRGB model.

Note that different color models have their own limits if their values are cropped in a (minimum-maximum) "valid" value range : ONLY this cropping is defining a "color gammut". Unfortunately, there's still no standard for defining how to crop colors that are converted from one color model to the other, but that don't fit in the restrictions of a color gammut (which is defined separately). We still lack a formal definition of how to "approximate" the color properly: some implementations are cropping the R,G,B components independantly of each other, may are trying to best approximate the Ligthness before approximating the Hue value and then the Saturation. This applies to HTML and CSS because HTML and CSS colors have a strictly defined color gammut with strict ranges (an error in my opinion, that existed in the legacy hex format for sRGB colors, but that was dropped in the "rgb(r,g,b)" notation which can be used for wider gammuts and for better precisions.

For colors that fall WITHIN the supported color gammut, there is also wide interpretation on how to fit an exact color to sample components that are not defined as continuous functions: most implmentations are just truncating the extra digits of precision, others are considering the nearest integer, others are trying to distribute the cumulative error onto nearby pixels (creating patterns that will increase the number of colors that are visually recognized). Here again, we lack a standard to specify the intent for the color approximation. verdy_p (talk) 10:38, 6 August 2010 (UTC)

Two Questions

1. After you (Jacobolus) replaced my color swatches, the color swatches you inserted do not register on my computer. Jacobolus, you still did not answer my question. Why is the yellow CIECAM02 color the SAME as the psychological primary color and the green, red, and blue CIECAM02 colors DIFFERENT (darker) than their CIECAM02 colors?Keraunos (talk) 02:43, 24 May 2010 (UTC)

2. My other question, Jacobolus, is what is the difference between the CIECAM02 colors and the four psychological primary colors? What are the CIECAM02 colors representing that is not already represented in the psychological primary colors? Keraunos (talk) 02:43, 24 May 2010 (UTC)

(1) Wait, what do you mean they don’t register on your computer? Those templates don’t work? They should.. it's just divs with background-color set. If there’s something broken with the colorsample template, we should fix it straightaway, since that's used all over wikipedia. (2) Well first those “psychological primaries” are not actually based on any source or measurement, but were one Wikipedian's eyeball match to some samples in a book, sometime in 2006 (I believe), but second, I only picked the hue to be correct, as I mentioned, and am not sure what would quite be the best lightness to use for these colors. The “psychological primary” swatches there are of roughly similar hue, but lower lightness. It’s probably reasonable to use a lower lightness than the swatches I put in; I’m not suggesting that those swatches were ideal; only that they were better (in hue anyway) than the ones there before, and that we should start a real discussion about the proper colors to use. Did you take a look at the images I’ve been sticking in here, and do you have any responses/thoughts about them? –jacobolus (t) 04:06, 24 May 2010 (UTC)

Different versions of the psychological primary green

You now have THREE different shades of green: Green (psychological primary), Green (NCS), and Green (CIACAM02). As far as I can tell, they are three different attempts to represent the psychological primary green in the RYGB color system (the four primary color psychological primary color system based on the opponent process theory of vision).

(Please do not change these color swatches since the other method of entering the color swatches you were using doesn’t work.)

Green (Psychological Primary) (Green (NCS))(#009246)
#009246

Green (NCS)(6G 5 / 7.61) (#008B66)
#008B66

CIECAM02 Green (#00BA85)
#00BA85

You still haven’t explained what CIECAM02 IS. WHAT IS IT? What does it represent? What does it represent that the Natural Color System or the darker Psychological Primary in the Wikipedia color chart doesn’t represent? Keraunos (talk) 04:28, 24 May 2010 (UTC)

NCS S 1080-R (#C40233)
#C40233

NCS S 0580-Y (#FFD300)
#FFD300

NCS S 1565-G (#00A368 )
#00A368

NCS S 1565-B (#0088BF)
#0088BF

Hrm? CIECAM02 is a color model defined by the International Commission on Illumination as the latest international standard for specifying color appearance correlates. I believe that its definitions of red, yellow, blue, green are based on the NCS, but I’m not positive about that. Only the hues are defined, because the goal is for a color to be expressible as a combination of hue, lightness, and chroma. Again, I should figure out what the best lightness is for approximating the defining colors of the NCS. Secondly, really, this template doesn’t work? (That should appear mid gray in sRGB). –jacobolus (t) 02:22, 25 May 2010 (UTC)

It still doesn't work. All I see is a vertical black bar. Keraunos (talk) 06:36, 26 May 2010 (UTC)

Okay!! That “psychological primary” green is somewhat wrong (which is unsurprising given the method by which it was created), according to the official "NCS Navigator" available from the NCS Website (click NCS navigator in the menu at the left, and then click “Start NCS Navigator”, even though it doesn’t look like a link at first glance). Here to the right are big swatches of the brightest red/green/yellow/blue available from among the choices in that official applet, and the sRGB coordinates given for them. –jacobolus (t) 22:22, 25 May 2010 (UTC)

(Note that these colors still aren’t quite ideal, since that NCS tool doesn’t have colors with whiteness=blackness and maximum in-sRGB-gamut chromaticness all choose-able from within its interface: that red is s10, w10, c80, the yellow is s5, w15, c80, the green is s15, w20, c65, and the blue is s15, w20, c65. Also, I’m not convinced that these actually fall within the sRGB gamut; they might just be approximations. Still, it’s an “official” source, so it should still be much better than the pretty sparse definitions I’ve been able to track down elsewhere, and better than the eyeballed guesses that the original maker of that opponent colors picture came up with in 2006. –jacobolus (t) 00:43, 26 May 2010 (UTC))

I think that the four NCS primary colors displayed in those four swatches above are an EXCELLENT representation of the four psychological primary colors! Keraunos (talk) 05:34, 26 May 2010 (UTC)

You read my mind! That is exactly what I was going to ask you next, to provide a representation of the four NCS primary colors. Keraunos (talk) 06:36, 26 May 2010 (UTC)

Jacobolus, you made an error when you changed the former renditions of the psychological primaries in the opponent process color chart. You mistakenly entered the code for red twice, but did not enter the code for yellow. I edited the image on Wikimedia Commons and corrected that error. New Opponent Colors (Psychological Primaries) color chart with correct coding for yellow: Keraunos (talk) 07:18, 26 May 2010 (UTC)

By the way, I'm the one who edited several months ago the Primary colors article to point out the fact that these four particular shades of red, yellow, green, and blue that we are fine tuning are called the psychological primaries. Keraunos (talk) 07:27, 26 May 2010 (UTC)

Oops. Thanks! –jacobolus (t) 22:54, 26 May 2010 (UTC)

NCS S 1080-R (#C40233)
#C40233

NCS S 1080-Y (#ECBD00 )
#ECBD00

NCS S 2060-G (#009F6B )
#009F6B

NCS S 2060-B (#0087BD)
#0087BD

Actually, I realized: maybe these, shown to the right would be even a slightly better representation, since they all have whiteness=blackness, and green, yellow, and blue are still apparently outside the sRGB gamut, given that the R, G, and B coordinates given by the NCS tool are at the extreme edge of the gamut (and this is the same red as above). –jacobolus (t) 22:54, 26 May 2010 (UTC)

Jacobolus, when you again refined the hex codes for the Psychological Primary Colors:, your forgot to change the rgb values to reflect the hex code changes. I just edited the image description and did that for you. Best wishes, Keraunos (talk) 06:05, 1 June 2010 (UTC)

NCS S 0580-Y (#FFD300)
#FFD300

I think the original yellow, at right, is better. I think the new yellow you changed the old yellow to is too dark. But I think the other three you changed are fine. Keraunos (talk) 06:32, 1 June 2010 (UTC)

Is there any chance we can go back to the old yellow (#FFD300)?Keraunos (talk) 06:37, 1 June 2010 (UTC)

Oh, probably. I think the main problem w/ my change here is that the lightness of yellow is too close to the lightness of the background, and the reduced contrast doesn’t look as good. I don’t really care too terribly much though. I’m not quite sure how the NCS people did gamut mapping, so I’m not sure how light their “aim yellow” really is (I don’t have access to a physical NCS swatch book, or to the specification). I just figured that picking the color with whiteness=blackness would probably be closer than the lighter alternative with whiteness 15, blackness 5. –jacobolus (t) 21:42, 3 June 2010 (UTC)

Did that help any? –jacobolus (t) 21:54, 3 June 2010 (UTC)

Kuenhi on variability in unique hue

VMS (and everyone else): part of our issue w/r/t differences about green may be differences in monitor set-up, etc., but a decent part might also just be variation in our personal understandings of unique green. Here’s what Kuehni has to say about it in his 2003 book Color Space and its Divisions:

“An interesting issue in connection with hue scaling is the psychophysical identity of the four unique hues, given an achromatic surround. Historically, these have been determined in the spectrum, using optical equipment. Individual results have varied to a surprising degree (e.g. see Ayama, Nakatsue, and Kaiser, 1987). For example, the wavelength of unique green has been determined in various experiments to be anywhere from 488 to 561 nm. Similar results have more recently been reported by Webster and co-workers (2000). Unique hues can also be determined, with considerably reduced variability, using object color samples (Hård, Sivik, and Tonnquist, 1997; Kuehni 2001a). The mean choices of 40 observers that directly selected their unique hues from arrays of Munsell hues were found to be 3.0R, 3.5Y, 2.5BG, 2.75PB. Indow, based on his principle hue scaling method, obtained the following locations: 3.75R, 5Y, 6G, 3PB (Indow, 1999b). Note the discrepancy in the green unique hue. In a recent, as yet unpublished, study using Munsell chip arrays with 75 observers average unique green was found to be 7.0G. Webster’s group (Webster et al., 2000) found green to have the largest individual variation among unique hues. This was also the case for Kuehni’s original experiment as well as the experiment with 75 observers. It is evident from results with Munsell chips that the range of dominant wavelengths in such experiments is much narrower than the reported range of spectral wavelengths for determinations using optical equipment. Recently it was shown that there is no relationship, as previously surmised, between the ratio of L and M cones in an individual’s retinas and the perception of unique yellow (Yamauchi et al., 2002). There is currently no model with good explanatory power for the relationship between stimulus and perceived unique hues.”

So anyway, it’s very possible that my unique green and yours differ substantially, and any color we pick is going to seem too yellow to one of us or too blue to another, even under identical viewing conditions. (7.0G though, which Kuehni refers to as the latest best estimate as of 2003, is even bluer than the color samples I’ve been showing above, which are at 6.0G, and 2.5 BG as estimated in one study is quite a bit bluer still.) –jacobolus (t) 16:08, 25 May 2010 (UTC)

I went back to the iMac today and took a look at the images as non-thumbs in Safari and with several external viewers. Nothing changed.

Could it be a color temperature issue? On the one CRT I have access to right now, I can make the blue-green hues appear nearly green, if I go from 9300 kelvin to 5000 kelvin while keeping the video driver set to 9300 kelvin. Putting the driver to 5000 makes the hues closer to "unique" green. The video driver on my Fedora Linux laptop does not have a way to change the LCD's color temperature, but I'd guess the default is around 9000 kelvin. As far as the studies above, it sounds like they were done under reflected light. I suspect it is likely that the temperature of that light was much lower than 9000 kelvin. VMS Mosaic (talk) 03:16, 26 May 2010 (UTC)

I found the Standard illuminant for sRGB is D65 (6504K), but it appears that CIECAM02 doesn't have a specified illuminant although experiments using it appear to usually use D65 for a D illuminant. VMS Mosaic (talk) 09:03, 26 May 2010 (UTC)

All of these RGB coordinates are in sRGB space (white = D65). If your display is well characterized, programs which do color management (e.g. Photoshop, or most of the system on OS X; I don’t know about Windows/Linux) should perform some chromatic adaptation to the white point and primaries of your display, which should make the color appearance stay roughly consistent, since your eyes will adapt to the color of the display. –jacobolus (t) 22:33, 26 May 2010 (UTC)

The definition of Munsell space is for standard illuminant C, which is somewhat similar to D65, but a bit purpler, and yes, is based on painted squares illuminated obliquely. Assuming that the CIE standard observer is actually a reasonable description of human metamerism, the appearance of colors on a screen should be describable in the same terms though. The chromatic adaptation using CAT02 (a linear transformation between XYZ points under one illuminant to XYZ points under a different illuminant), which I used for transforming the Munsell renotation data to be relative to illuminant D65 for sRGB, should do a pretty decent job of preserving color appearance, especially since illuminants C and D65 are fairly similar in chromaticity (should do a decent job at least from what I’ve read; I’m not really an expert). –jacobolus (t) 22:40, 26 May 2010 (UTC)

I doubt that the laptop has a CCT of 9000K; I think most laptops are in the 6000–7000 range, but I’m not positive about that. My Apple laptop claims to have a white point of D65 (I haven’t used a hardware colorimeter/spectrophotometer on it). –jacobolus (t) 23:21, 26 May 2010 (UTC)

Almost all CRTs default to around 9300K. LCD monitors default to either 9300K or 6500K. Some newer ones default to 6500K (but can be set to 9300K), while older ones were mostly 9300K. The backlighting in laptops appear to mainly use 6500K, but when I put current model Thinkpad laptops (running either Linux or Windows XP (with default video settings)) next to a 9300K CRT being driven by the laptop external video connector displaying the exact same desktop, the color temperature appears to be the same (9000K+) on both. The same is true when I place my laptop next to a non-color managed workstation with three 9300K CRTs (i.e., the laptop appears to have a color temperature of 9300K). I suspect the Linux (the Windows laptop is not currently available to me to check it's settings) laptop's video driver/card is defaulting to displaying 9300K on what it knows to be a 6500K LCD, and it assumes the external display is 9300K. The images on this page look pretty much the same regardless of where I look at them, except when I force a CRT monitor to 6500K.

As far as eye adjustment, I buy my own fluorescent light bulbs for my office at my job because my "eyes" don't like the blue-green cast everything has otherwise. My eyes would never adjust to the colors on a 6500K CRT/video driver; I find the pinkish/orangish white to be off putting. I spent more than a month finding a combination of fluorescent light bulbs (different CRIs) which gave decent light in my kitchen (a bit of a hassle to keep several bulb types for spares). No, it is simply false that everyone's eyes automatically adjust, however, I do believe it is true that 99.99+% of people simply don't notice the color differences, particularly since I have never once walked into anyone's home with a CRT TV where faces weren't either radically too red or green. I purchased my own color bar generator to adjust my sets, but it provided only a slight improvement over doing it by eye. My point here is that my eyes, at least, are not capable of ignoring the fact that the current "green" in the green infobox title will appear too blue on the average Wikipedia user's display. Yes, if users manage to get their video drivers and displays set to properly display at 6500K, then it should look acceptable, but I'd wager that I could check all 50+ computer displays at my IT employer (including the state of the art iMac user's) without finding even one which displays it as green instead of blue-green.

It's obvious that we aren't seeing the same thing on our displays (I have my doubts that it is a simple matter of having very different ideas of what green is), if not due to color temperature, then for some other reason. So, what is the goal? To display a color which looks correct on a few displays, or one that looks correct on the vast majority of displays? My contention is that the current green looks too blue on the vast majority (at least several hundred million) displays currently in use. Maybe at some future time, the situation will change.

At this point I propose we return to no color in the green infobox title given that there is no consensus that the current color is appropriate. VMS Mosaic (talk) 05:44, 27 May 2010 (UTC)

After several hours of fooling around with multiple LCD and CRT monitors at all settable color temperatures dual-headed on my laptop under multiple lighting conditions (along with an LCD on a Windows Vista system), about all I'm sure of is that I can't make the G part of the images appear closer to green than a bluish green, and that the apparent color temperature is fairly dependent on the screen brightness setting. With some combination of temperature, gamma and brightness settings, it is probably possible to make the G section appear green, but I didn't find one. I got several more opinions about the images including on their own PC displays, and so far no one sees green where marked. I'm out of ideas. VMS Mosaic (talk) 23:28, 27 May 2010 (UTC)

NCS S 2060-G (#009F6B)
#009F6B

Well, supposedly we’re assuming that displays are sRGB; that’s the standard for the web. If displays are not in fact sRGB, and the colors in browsers aren’t properly color managed, there is close to nothing we can do about it. So the color sample to the right, taken from the NCS conversion of its "green" to sRGB, looks more “blue” or "blue-green" than “green” to you? To me, on four different monitors of different types and 3 operating systems, it looks pretty "green". It's hard to say what accounts for our difference on this. –jacobolus (t) 03:54, 28 May 2010 (UTC)

I'd call it a shade of green, but on the bluish side. If I stand six feet back from the laptop and look about 15 to 20 degrees below perpendicular, it does appear to be close to "actual" green, but up close, I have to get far below perpendicular to see it as "actual" green. If I get anywhere above perpendicular, it appears very bluish. This is why I prefer CRTs; color and brightness do not depend on viewing angle. To use a big word, the shade is not monolexemic (i.e., to many observers, it is bluish green or blue green instead of green). We need a monolexexmic shade (i.e., one which almost everyone calls "green"). VMS Mosaic (talk) 05:48, 28 May 2010 (UTC)

Yeah, laptop LCDs’ color shifts with vertical viewing angle are tremendously annoying to me as a photographer, often trying to do color work on them sitting in a café or library; the top of the screen and the bottom end up with slightly different colors. Anyway, I dunno. As for monolexemes: the “green” primary of an sRGB monitor is just about “yellowish green”; it hasn’t quite reached “yellow-green” or “chartreuse”, but it’s well on its way. Here’s an image to the right showing some various greens, this time at a consistent single lightness and chroma, for ease of comparison. –jacobolus (t) 19:45, 29 May 2010 (UTC)

Green (or the closest thing to it) is approximately under the 's' in ncs. VMS Mosaic (talk) 07:39, 31 May 2010 (UTC)

Okay. For me, on both a CRT and a laptop LCD, that’s on the yellow side, and “green” falls pretty close to the NCS definition, or perhaps a bit to the right, say under the second “o” in tool. –jacobolus (t) 22:41, 31 May 2010 (UTC)

Note that all the existing CIE and XYZ color models will soon be deprecated as well (including sRGB). Why? Simply because they are ONLY defined as trichromatic color models. Newer displays (and models used by printers that need FOUR basic pigments at least) are appearting that use 4 pigments, but you should also know that the color model transform that occurs on those newer monitor (notably those initially made by Yamaha and promoted now by various other HDTV manufacturers) are all proprietary:

These transforms are highly correlated to the subpixel geometry used, and to the color pigments chosen (for example when choosing the fourth subpixel as white or as yellow, or if adding another pigment within the greens, that are the most precisely defined colors where we perceive most differences). The pixel geometry, and the physical properties of the diffusion mask (that helps hiding the black separation between subpixels) are playing a big role in the effectively perceived color (because of how much the very thin "black" borders of subpixels are covering the space). There are also other technologies that DON'T define subpixels, but instead use half-transparent cristals, so that the subpixel geometry no longer plays any effect).

Still, the most important differences between displays is not there: it's in how these displays are illuminated (because most flat panels are LCD and DO NOT emit their own light, but are illuminated by a "white" emitting tube, whose light will then be absorbed or will pass through the LCD mask: the nature of this white illuminating light is effectively what differenciates the most the various flat displays. For most notebooks, the white emitting illumination tube uses a 6300K white, only because these tubes require less energy (so this better saves the batteries on notebooks). Flat panels for desktops or HDTV still use the reference white light around 9000K because energy saving is not a problem for them (so they can use more energetic hallogen tubes, instead of basic, and cheap, neons that give a yellowish "white" that you frequently see on most notebooks or on smartphones ; high-end notebooks are using white LEDs that give a more energetic white, around 9600K, but that looks a bit "blueish").

New HDTV fixed displays (with proprietary quadrichromatic subpixels and proprietary conversion from sRGB to their own specific color model) use hallogens for their white: their power consumption (and emitted heat) is much higher, only because of this illumination, but not because of their proprietary subpixels geometry or because of their diffusion mask.

New notebooks using the proprietary quadrichromatic geometries will appear soon, and their white will also still depend on how they are illuminated (same tradeoffs will apply because of energy used by this illumination).

Active displays (emitting their own light, for example with LEDs) are much more expensive than LCD.

OLED displays are within them, but the bad thing is that the pigments of this displays are migrating, so the colors tend to be less vivid over time (due to chemical migrations), becoming more grey... So even if they still maintain their white temperature, their color gammut will constantly shrink over time and it's impossible to maintain their calibration.

Finally, note that the diffusion mask (between subpixels and the display surface), which is useful to avoid seeing color artefacts on vertical thin lines or contrasting vertical borders, and to avoid seeing the very thin black/non transparent borders around subpixels (the electric conductors and the isolating NPNP interfaces) , take a big role, because they also absorb part of the light, and change the white temperature, even if the liquid cristals are fully transparent. The nature of the semi-transparent liquid cristal (whose transparency is controlled by the voltage applied to reorient it) also plays a minor role because it also absorbs a part of the spectrum of the illuminating white.

The CIE is currently working (in a joint "technical commitee" with ISO) on a newer standard model for quadrichromatic displays and printers (and possibly with more base pigments, notably because now many printers support up to 6 colors plus black), with exact formulas to map existing CIE color models (including sRGB and XYZ) onto these models, with new standardized color profile definitions. This will solve the interoperability problems with all devices that are not natively using sRGB or a trichomatic native rendering system.

For now the relevant standard is ISO 11664 (last edition in 2008) which is the same as the CIE Standard 014 (last edition in 2006): it is based on a standard tri-stimulus observer only, ignoring the differences that exist in humane vision on tha nature of these three stimulus (notably, there's at least TWO distinct green pigments in the humane vision, where people have one or the other). A quadristimulus model would better match with most peoples (notably within males, because they are those the most affected, by the shift of sensibility of the "standard" green). The tri-stimulus model does not work correctly.

Other stimulus are still ignored by the fact that the curent model is based on the total energy and a "standard" emission/absorbtion profile for each pigment (this profile is a curve of sensibility depending on wavelength, and the three stimulus is effectively the result of the integration of effective light spectrums multiplied by these sensibility curves; these curves are exactly defining each pigment, and contains all other effects than just absorbtion/emission, but also reflection, and photoluminescence which is the effect of reemitting a part the light energy that has been absorbed but at a different wavelength).

verdy_p (talk) 11:15, 6 August 2010 (UTC)

xkcd Color Survey

Randall Munroe performed a color survey and published the results at http://blog.xkcd.com/2010/05/03/color-survey-results/ and at http://xkcd.com/color/rgb/. Is this data suitable to be used for citations? --Humanist Geek (talk) 21:43, 30 May 2010 (UTC)

I believe the results for the set of 48 colors are spot on. I'd have no problem using the green RGB value. VMS Mosaic (talk) 07:32, 31 May 2010 (UTC)

Randall is great – I love his comics, and I’ve exchanged some emails with him about his color survey – but given the vagaries of internet surveys, the variations in display and lighting context from one observer to the next, and the nature of his methods in calculating color representatives (in particular, measuring distances relative to the perceptually non-uniform distances in RGB of unspecified primaries, white point, and gamma), they can hardly be described as “authoritative”. I suppose we could describe them as “relative to the random mishmash of conditions of a large internet color survey” – as opposed to relative to sRGB – but that makes it somewhat difficult for anyone to interpret specifically what they mean. –jacobolus (t) 18:26, 31 May 2010 (UTC)

In particular, the locations are heavily biased by the differing chroma available for different hues, which makes Randall’s representative for “red” much too orange, for “green” too yellow, for “yellow” too green, and for “blue” too red. This is the same problem facing the X11/CSS definitions. –jacobolus (t) 18:34, 31 May 2010 (UTC)

I'm afraid we will have to agree to disagree. On a 6500K LCD (apparently the standard display today), his green is actually very slightly bluish, the yellow is a little reddish, and there is not the slightest hint of red in the blue. The red is closer to actual red than any of the NCS and CIECAM02 reds on this page (they are all close to being magenta). Some of the very slight variations I see come and go on several CRT monitors depending on color temp, so given monitor variation, I think they are about as close as one is going to get. Show me a valid observational study done on the average PC/laptop LCD under typical office/home lighting conditions which says otherwise and I'll shut up and go away.

There really is little more to discuss unless others have opinions. At this point I am starting to lean strongly toward displaying nothing at all for the basic colors; doing so would be much better than displaying colors (both infobox titles and icons) which look very wrong to the average web user on his average 6500K LCD.

Another option might be to display multiple colors bars in the box with none in the title. Green would have X11, HTML, CIECAM02, NCS, etc., each labeled with source and coordinates. Then the vast majority of users would see at least one color they'd call green. VMS Mosaic (talk) 04:25, 1 June 2010 (UTC)

Okay, well I’d tend to go with every scientific color term survey, color model based on psychophysical experiment, and official color designation system created by experts that I’ve ever seen, and you can “agree to disagree” with them all, I suppose. What would your standard for evidence be that either your display is showing you something odd or your definitions are idiosyncratic (or at least different from average; still well within the range of typical inter-observer variation)? As far as I can tell, there’s no possible way I could convince you. –jacobolus (t) 19:05, 1 June 2010 (UTC)

I simply don't believe studies done on paper are relevant, given the colors being suggested by the studies for use here. This entire discussion started with PaleAqua stating that the green title bar color "looks very blue to me." I agreed completely before she said anything, but I didn't want to get into the exact discussion we are having. One person I was showing the colors and images to actually suggested you were putting me on (aka yanking my chain) and someone else in the room immediately agreed that had to be the case; I didn't believe that then and I still don't. The people who said it was a put-on were all using their own LCDs, so it is not a case of me misadjusting monitors or anything like that. It's not just a matter of convincing me; I know four others who also need convinced that they are very wrong in their color perception. VMS Mosaic (talk) 21:58, 1 June 2010 (UTC)

(I’m searching for unique hue studies done on computer displays; I’ll report back w/ findings hopefully soonish. –jacobolus (t) 19:26, 1 June 2010 (UTC))

I agree with this option. Point out in the articles about primary colors the fact that the definition of a primary color depends on which color system one is using. For the primary colors, having a color box for the color as rendered for each of the major color systems--RGB (Primaries of light)--the X11 color, CMY (CMYK) (Primaries of pigment)--the process color (CMY) or the pigment color (RGB), RYGB (Psychological primaries)--The NCS color and the CIECAM02 color. However, I think the X11 color should be displayed first because RGB color space is the color system that is used to display colors in Wikipedia (and on all computers and television sets) and thus the X11 versions of the primary colors are those colors which the Wikipedia user expects to see displayed first. Keraunos (talk) 04:55, 12 June 2010 (UTC)

Would you accept HTML green in place of X11 green since the HTML value is used by the web and Wikipedia? Would you be willing to remove the icons? I find most of the current icons to be unhelpful. VMS Mosaic (talk) 09:56, 12 June 2010 (UTC)

The article on green should display BOTH the X11 and the HTML green, along with Jacobolus' green psychological primary and the CMYK pigment green composed of equal amounts of cyan and yellow (already in the Variations of green article). They are all different definitions of green according to different color systems. Keraunos (talk) 06:47, 5 July 2010 (UTC)

VMS: The prior set of “icons” showed an extremely limited and RGB-centric depiction, backed by absolutely no reasoning/evidence, just made up by whoever came along. In particular, the depictions of colors like orange and purple were terrible. (In general there was nearly no variability in hue for any of the icons, only in lightness and chroma.) Keraunos: Only the primary lights/pigments themselves are reasonable to include as authoritative depictions of anything. That is, for instance, somewhere in the article about blue we should mention that one of the primaries of an RGB display, called “blue” in that device-centric model, is a violet-blue color of such-and-such appearance. But including any RGB-centric definition of (for instance) brown, pink, or chartreuse would be inappropriate, in my opinion. CMYK primaries should also be depicted and mentioned in the relevant articles, but it should not be implied that they are in any way unique or representative of general color-name categories. The X11 definitions should be banished to the pages specifically about them; elsewhere they are non-authoritative and nearly useless. –jacobolus (t) 01:21, 13 June 2010 (UTC)

My suggestion would be to list:

ISCC–NBS centroids or their closest in-gamut approximations (I just downloaded the paper defining these, and I’ll try to report those definitions here) for a name
best approximation to NCS target colors
perhaps some kind of anthropological/linguistic cross-cultural color loci, if a reasonably clear source can be found
typical CMYK/RGB primaries where applicable, or their nearest in-gamut approximation
representations of common pigments of a color (not sure where to find a good source for this)

And to additionally include some chart or charts representing the typical range of colors known by a name, either with reference to the Dictionary of Color or some other reliable source. Also, we could really use more images for all of these color pages, in particular showing a broader range of significant historical artworks employing each color.

–jacobolus (t) 02:57, 13 June 2010 (UTC)

I'm going to change the yellow in the psychological primary color chart back to the slightly lighter version within a few days since you said you had no objection to that. Then within the next week or so I'm going to create a suggested template for displaying the various versions of the primary colors. However, the RGB (X11) colors have to continue to be the one of the primary primary color sources, along with the CMYK colors and your four psychological primaries. The ISCC-NBS colors are totally useless for primary colors because they are too dark. Keraunos (talk) 06:42, 5 July 2010 (UTC)

“[...] suggested template for displaying the various versions of the primary colors [...]”. Sure, sounds like a good idea. Put your template in here and we can get a discussion started about it. “[...] the RGB (X11) colors have to continue to be the one of the primary primary color sources [...]”. There’s no possible way to use them as a remotely reasonable (much less “reliable”, “expert”, “authoritative”, etc.) source though. These colors were defined by some unknown individual for his own personal 1980s computer display, with hardware specific phosphors and transfer functions, based on his personal (non-expert) notions. “The ISCC-NBS colors are totally useless [...]” Which ISCC–NBS colors are we talking about? –jacobolus (t) 23:44, 6 July 2010 (UTC)

The X11 primary colors red, green, and blue are three spectral color approximations defined in such a way so as, when mixing lights of these three colors in various combinations, to be able to display the maximum possible gamut of colors in RGB color space on a computer or television set. That is the purpose of the RGB color system. Keraunos (talk) 00:10, 7 July 2010 (UTC)

Okay. Those shouldn’t be referred to “X11 primary colors” since they don’t have much to do with X11. Definitely, the articles about red, green, and blue should each include a section called “additive primary” or some such. Every device has a slightly different set of RGB primaries, so the specific chromaticities of one device are less important for those sections than the description of additive light mixture, in general. –jacobolus (t) 13:44, 7 July 2010 (UTC)

Jacobolus, do you want the CIECAM02 colors on the template? If so, can you provide me with four color swatches for the red, yellow, green, and blue CIECAM02 colors and give me a brief summary in three or four sentences explaining what the CIECAM02 colors are for? What is their purpose? Best wishes, Keraunos (talk) 01:25, 20 July 2010 (UTC)

Which template are we talking about. I think that listing the best approximations to NCS aim colors that we can is probably sufficient. –jacobolus (t) 02:14, 20 July 2010 (UTC)

The raw data from the colour survey was also provided [84Mb], just in case someone else wants to mine it to see if something other than a web RGB distance is a better fit for determining if two colours are called the same thing. gringer (talk) 01:26, 6 August 2010 (UTC)

It's worth putting a link to the XKCD survey on the relevant wikipedia pages, in the "external links" section. The XKCD blog is not a reliable source for article content, though, and any analysis of the data on Wikipedia’s part would be original research. I recommend that those interested analyze the data and publish results on their own personal websites. –jacobolus (t) 05:46, 6 August 2010 (UTC)

Category:Ultraviolet

FYI, Category:Ultraviolet is up on WP:CFD for renaming, or splitting, or something else (that's the nomination). As UV is a colour that is visible to some animals, I thought I'd let you know. 76.66.192.55 (talk) 04:54, 10 July 2010 (UTC)

Variations of blue deletion discussion

Some other project members might be interested in the discussion over at Wikipedia:Articles for deletion/Variations of blue. –jacobolus (t) 22:45, 24 July 2010 (UTC)

By the way Jacobolus, if you could vote a definite keep for the Variations of blue article that would be helpful as we will need that later. Keraunos (talk) 03:40, 27 July 2010 (UTC)

Munsell colors in the color vision and colorimetry subsections

Jacobolus, I think you have done a wonderful job in the section in the color green on color vision and colorimetry. That is pretty much what I was going to say so the next step is for you to create similar sections for red, yellow, and blue. I’ll let you take care of that. Keraunos (talk) 03:40, 27 July 2010 (UTC)

An example green
Munsell 2.5G 5/9
#088C56

I was wondering, when we put in the Munsell colors, wouldn’t be better to use the 10 or 5 Munsell colors instead of the 2.5G Munsell color you used in the green section on color vision and colorimetry? Does 2.5G Munsell mean it is halfway between 10GY and 5G? Do you have a way to calculate the complements of the 10 or 5 Munsell colors?

The hues of the Munsell color system, at varying values, and maximum chroma to stay in the sRGB gamut.

The 10 or 5 colors would be more to the center of the red, yellow, green, blue, or purple Munsell colors. Do you think it would be better to use the Munsell 5 or the Munsell 10 colors? Keraunos (talk) 03:40, 27 July 2010 (UTC)

Munsell colors are approximate because they have been adjusted to remain within the RGB gamut. Keraunos (talk) 03:40, 27 July 2010 (UTC)

Munsell 5 Colors:

Munsell 5P Purple (#9F00C5)
#9F00C5

Munsell 5B Blue (#0093AF)
#0093AF

Munsell 5G Green (#00A877)
#00A877

Munsell 5Y Yellow (#EFCC00)
#EFCC00

Munsell 5R Red (#F2003C)
#F2003C

Munsell 10 Colors:

Munsell 10P (#0082B2)
#B900A6

Munsell 10B (#0082B2)
#0082B2

Munsell 10G (#00A78A)
#00A78A

Munsell 10Y (#F0EA00)
#F0EA00

Munsell 10R (#EE82EE)
#F85900

The Munsell 5 colors look better to me. Keraunos (talk) 03:50, 27 July 2010 (UTC)

I really don’t understand what you’re asking. I think 2.5G 5/9 is a perfectly fine example color for the green article. As the caption says, it’s just an example. Also, yes, 2.5G is halfway between 10GY and 5G. –jacobolus (t) 18:58, 30 July 2010 (UTC)

I don’t understand why after listing 5 evenly spaced hues (5R, 5Y, 5G, 5B, 5P), you followed up by listing 5 others that are a quarter of the way between pairs in the first bunch you listed. –jacobolus (t) 19:03, 30 July 2010 (UTC)

Template: List of Colors -- why?

Can anyone see the purpose of this template, with its odd assortment of "major colors", and variant definition of "secondary" that conflicts with the linked article? Dicklyon (talk) 05:17, 4 August 2010 (UTC)

I've proposed deleting it, so either way, opinions would be useful at Wikipedia:Templates for discussion#Template:List of Colors. Dicklyon (talk) 05:24, 4 August 2010 (UTC)

Template

Due to some people not understanding the list template. I am all for changing the template into a Color topic template instead. Here's sort of example of what it would look like.

Jhenderson777 (talk) 16:38, 5 August 2010 (UTC)

And if you haven't noticed it already. There is a hidden comment on the Colors list. What do you think? Jhenderson777 (talk) 16:46, 5 August 2010 (UTC)

I have no idea where/what you're referring to. Dicklyon (talk) 18:17, 7 August 2010 (UTC)

The template organization needs to be given more thought. In particular, what sourced rationale will you rely on the limit the colors in the colors list, to keep it from growing to hundreds? Or just leave out that bit? Dicklyon (talk) 18:17, 7 August 2010 (UTC)

External link - project watchlist

Wikiproject Watchlist - WikiProject Color

vs

Special:RecentChangesLinked/Wikipedia:WikiProject Color

The toolserver link catches every single page tagged with the project banner. The Special: link only catches the pages that are mentioned on the project-page itself. -- Quiddity (talk) 21:39, 26 August 2010 (UTC)

Okay, then we just need to add a link to a category of "all wikiproject color pages" to the talk page template, and then use it in a similar way to Special:RecentChangesLinked/Category:High-importance color articles. Since MediaWiki software has this as a built-in feature, and its change browser is nicer than the external link, it makes sense to just use it. –jacobolus (t) 01:37, 27 August 2010 (UTC)

Also, your tool seems to be broken. Gives me a page like “<class '_mysql_exceptions.ProgrammingError'> ...” –jacobolus (t) 01:39, 27 August 2010 (UTC)

It's not "my" tool in any way! I just noticed the disparity between the results it gives, and the results the special page gives. I had seen the same tool link added to a different wikiproject days ago, so was familiar with the idea.

Is it still giving you the same error? It works for me currently... If you still have problems, I'll poke the editor who added the link (User:Wavelength), to see if he/she knows what's up.

Just utilizing the {{WikiProject Color}} talkpage banner would not work, because it would only catch changes made to the talkpages. Hence the invention of this toolserver tool! (afaik)

(The "Wikipedia:Toolserver" is an official Wikimedia endeavor (despite the simple domain name), if that is your concern. It's not just some random person's script...! :) This particular script is apparently coded by User:Tim1357, according to the url)

HTH. -- Quiddity (talk) 03:15, 27 August 2010 (UTC)

Sorry, I didn’t mean "your tool" as in "you specifically created it". Hmm, you’re right about the talk pages. Too bad there’s no way to do “pages that link here/are linked from here & also their corresponding talk/article pages”. The built-in special page should have the ability to do that & also to take a list of pages as a source rather than only a single page. Anyway, it’s fine with me to link the toolserver tool. It still seems like there should be some better solution within mediawiki itself, but apparently there isn't. Alas. –jacobolus (t) 19:11, 28 August 2010 (UTC)

NP :) Reality constantly interferes with my plans for World Domination, too ;) -- Quiddity (talk) 05:31, 29 August 2010 (UTC)

By the way, I think I have addressed all of the problems raised in the thread above, the reason for jacobolus' error was that I was rapidly developing the tool, uploading patches and extensions that frequently brought the tool down. Tim 1357 ^talk 23:34, 10 November 2010 (UTC)

Optic yellow

I was just wondering about this color, there's not much information about it in Wikipedia. It's used for high visibility, most notably in tennis balls. It looks more yellow than green-yellow... I don't know what's the correct place to write about it. If someone's up for writing it, now's a perfect time because the US Open tournament is underway. ;) Wipe (talk) 03:01, 1 September 2010 (UTC)

I've never heard of that. Do you have a source describing it? –jacobolus (t) 14:36, 1 September 2010 (UTC)

Check out the ITF Tennis - Ball history. They also have a section on testing color, [1] but doesn't seem to have much details. PaleAqua (talk) 07:49, 5 September 2010 (UTC)

Color articles have been selected for the Wikipedia 0.8 release

Version 0.8 is a collection of Wikipedia articles selected by the Wikipedia 1.0 team for offline release on USB key, DVD and mobile phone. Articles were selected based on their assessed importance and quality, then article versions (revisionIDs) were chosen for trustworthiness (freedom from vandalism) using an adaptation of the WikiTrust algorithm.

We would like to ask you to review the Color articles and revisionIDs we have chosen. Selected articles are marked with a diamond symbol (♦) to the right of each article, and this symbol links to the selected version of each article. If you believe we have included or excluded articles inappropriately, please contact us at Wikipedia talk:Version 0.8 with the details. You may wish to look at your WikiProject's articles with cleanup tags and try to improve any that need work; if you do, please give us the new revisionID at Wikipedia talk:Version 0.8. We would like to complete this consultation period by midnight UTC on Monday, October 11th.

We have greatly streamlined the process since the Version 0.7 release, so we aim to have the collection ready for distribution by the end of October, 2010. As a result, we are planning to distribute the collection much more widely, while continuing to work with groups such as One Laptop per Child and Wikipedia for Schools to extend the reach of Wikipedia worldwide. Please help us, with your WikiProject's feedback!

For the Wikipedia 1.0 editorial team, SelectionBot 22:16, 19 September 2010 (UTC)

All the articles chosen look fine to me as excellent choices. Keraunos (talk) 00:02, 20 September 2010 (UTC)

New color displays

Here are the new color displays I have prepared, based on the suggestion of VMS Mosaic, for the colors red, yellow, green, and blue to resolve the ongoing dispute since May 2010 regarding which colors to use to represent the four psychological primary colors. I am following VMS Mosaic’s suggestion to blank out the color boxes and provide samples of the various colors in different color systems. This solution should satisfy everyone. This way, each person can decide for themselves which color they choose to regard as red, yellow, green, or blue.

Please leave your comments regarding this new primary colors display as I would like to install these new color displays within the next week. Keraunos (talk) 04:02, 6 October 2010 (UTC)

Link displaying the new color displays: New Color Displays for the colors Red, Yellow, Green, and Blue:

I think they're very ugly; worse even than the strip charts you used to do. I'm reverting those. Dicklyon (talk) 05:49, 28 October 2010 (UTC)

Comments regarding new RYGB color displays

This is a good idea, but the execution takes up a really large amount of space, which makes it hard to compare. The infoboxes for colors are in particular really huge; it should be possible to consolidate all the variants to a much greater extent.

Other notes:

I don’t like the naming. It would be better to just write, for example, “sRGB secondary yellow”, “GRACoL CMYK primary yellow”, “NCS Y”, “Munsell 5Y”. All the parenthetical stuff is excessive, and in particular there’s no reason to mention HTML/CSS/X11.
Some of the material is original research. For example, I’ve never heard the name “pigment green” for the mixture of process C and Y inks.
Some of the phrasing is misleading. e.g. “the color defined as green in the RGB ...” should say instead something like “the green color used as a primary in the RGB model...”
The multiple repetitions of hex triplets are unnecessary. Hex triplets aren’t especially useful even when only written once. Once is enough.
I’d use some clearly specified industry standard CMYK color space (e.g. some GRACoL/FOGRA space), and convert to RGB via some explicitly described process, instead of using an arbitrary source.
It would probably be good to include data coming from one of the big anthropological/linguistic studies of color naming.
The caveats about individual variation in unique hues might be mentioned.

–jacobolus (t) 00:28, 7 October 2010 (UTC)

I believe it would be a major mistake for a web based encyclopedia to not include the web (HTML/CSS) color definitions in the applicable color articles. The HTML/CSS colors are what one sees when a color is specified on the web. If nothing else the reader can see how "wrong" (personally, I think the HTML green is the only one suggested which even comes close to accurately representing green) the web colors are compared to the "expert's" opinions of the "true" colors. VMS Mosaic (talk) 05:51, 7 October 2010 (UTC)

VMS Mosaic, I did include HTML green in the green color display. For red, yellow, and blue, the HTML color is identical to the X11 color. I'll incorporate that fact in the display. Keraunos (talk) 21:57, 7 October 2010 (UTC)

Jacobolus, that is why I put the color swatches at the top of each color display, to make it easier to compare the different definitions of red, yellow, green, and blue. Thank you for your input. I will be incorporating most of your other suggestions. Keraunos (talk) 22:05, 7 October 2010 (UTC)

VMS, when you make comments can you leave the "minor edit" box unchecked? There’s nothing “wrong” about the RGB primaries of a particular computer display (or with "green" for instance, the mixture of the green primary at 20% intensity with the other two primaries off). They’re just arbitrary. –jacobolus (t) 00:31, 8 October 2010 (UTC)

My edits default to minor, but I see that the guidelines state that talk page comments should not be labeled as minor, so I will try to remember to mark them as major. Being "arbitrary" does not change the fact that they are in common usage and therefore need to be represented/discussed. From a statistical standpoint, I don't believe the HTML/CSS colors are significantly more "arbitrary" than any of the other color representations. From a personal standpoint, the HTML/CSS colors are the most accurate representational colors given the other alternatives. For example, IMHO, the green article's infobox icon doesn't contain anything even close to a "pure" shade of green. VMS Mosaic (talk) 11:30, 8 October 2010 (UTC)

By the way, Jacobolus, someone has already changed the green in the color box in the green article to HTML color #008000 from the color code for green (NCS). This is an example of why we need to install these color displays, so as to solve this problem once and for all. Keraunos (talk) 22:16, 7 October 2010 (UTC)

Oh, they did it in a "minor edit". One other thing, K. Can you put the immediately preceding sections here into a subpage of your user page (something like user:Keraunos/primary color displays or similar) instead of inline in this discussion page? –jacobolus (t) 00:34, 8 October 2010 (UTC)

WikiProject cleanup listing

I have created together with Smallman12q a toolserver tool that shows a weekly-updated list of cleanup categories for WikiProjects, that can be used as a replacement for WolterBot and this WikiProject is among those that are already included (because it is a member of Category:WolterBot cleanup listing subscriptions). See the tool's wiki page, this project's listing in one big table or by categories and the index of WikiProjects. Svick (talk) 20:00, 7 November 2010 (UTC)

Invitation to participate!

Hello! As you may be aware, the Wikimedia Foundation is gearing up for our annual fundraiser. We want to hit our goal, and hit it as soon as possible, so that we can focus on Wikipedia's tenth anniversary (January 15) and on our new project, the Contribution Team.

I'm posting across WikiProjects to engage you, the community, in working to build Wikipedia not only through financial donations, but also through collaboration in building content. You can find more information in Philippe Beaudette's memo to the communities here.

Please visit the Contribution Team page and the Fundraising page to find out how you can help us support and spread free knowledge. ⇒DanRosenthal ^{Wikipedia Contribution Team} 18:05, 15 November 2010 (UTC)

A few great resources at Google Books

Troland et al.’s “Report of Committee on Colorimetry for 1920–21” in JOSA, which is the best comprehensive summary of color science as it existed circa 1920.
Troland’s The Present Status of Visual Science, 1922. Does a nice job laying out the main problems of psychophysics and color vision science (and is quite relevant still ~90 years later).
Nickerson’s “handbook” Color Measurement and its Application to the Grading of Agricultural Products, which provides a good introduction to the state of color science and colorimetry up through 1946.
Kelly & Judd’s Color: Universal Language and Dictionary of Names, 1976.
A collection of many of Judd’s color science papers (750 pages worth) from the mid-20s through the mid-70s and covering an awful lot of the color science of that whole period, many of them co-written with other leading color scientists, Contributions to Color Science, published after his death by the National Bureau of Standards, where Judd had worked for decades.
Building for people: behavioral research approaches and directions, Chapter 13, “Color”. A pretty gentle introduction in this architecture book put out by the NBS in 1980.

PDFs of all of these can be downloaded.

–jacobolus (t) 02:20, 8 December 2010 (UTC)

The Keraunos problem

Our long-time color contributor User:Keraunos continues to march to the beat of a different drummer, having no interest in WP:V, WP:RS, and consensus style issues. He has restored "in culture" and such trivia sections that contain exactly zero supported items (and I'm pretty generous in interpreting "supported", meaning that if there's a link to an article that even mentions the color, I'll usually accept that in lieu of a source, at least temporarily). And he puts in his favorite html "strip charts" that we long ago deprecated and removed from every place we could find them (some may still exist, since they don't use a template that would make them findable). I reverted a bunch of his "restores" today. Any support for continuing to push back on his contributions? Dicklyon (talk) 19:15, 28 December 2010 (UTC)

Let's call a truce. I don't want to cause you extra work. You should have the time you need to make your valuable contributions to the scientific articles and technology articles because you know so much about science and technology. I promise not to put in any more new unsourced items. All the new items I've been putting in that I've been putting in myself have all been sourced for a long time. I just don't see why every single item that someone else puts in about what they saw a character mention about a color in a novel, in a movie, or on TV needs to be sourced. Why would anyone lie about what they saw a character say about a color in a novel, a movie or on TV? It seems to me they only need to source an item if it is a claim about a serious subject such as anthropology, sociology, history, technology, or science. It's just that I'm am extreme inclusionist. I hate to see anyone delete anything. I think that Wikipedia should contain all human knowledge, even including what characters mentioned what about which colors in which movies and TV shows and video games. Most of those items about things they saw in the media are probably contributed by teenagers and young adults. I remember when I was a teenager I used to watch a lot of television. They can get practice in adding to Wikipedia now and then add more scholarly items when they get older. You shouldn't take Wikipeda so seriously! Relax! Calm down! Wikipedia should be something fun and enjoyable. But let's call a truce. Best wishes, and Happy Holidays! Keraunos (talk) 10:09, 29 December 2010 (UTC)

Personal observations of colors in shows and such are exactly what I'm trying to clean up. I don't think they're lies, but they are essentially WP:OR if they are not found in a secondary WP:RS. You should take Wikipedia more seriously; I'm a moderate deletionist. Have a happy new year. Dicklyon (talk) 16:07, 29 December 2010 (UTC)

I read on the Internet that by the year 2020, because of the operation of Moore's Law, it will be possible to have a one page Wikipedia article about everyone in the world! Last year, I got a half a terabyte backup hard drive for my computer. By 2020, the average personal computer hard drive will probably have a petabyte of memory! So there will be plenty of room for a lot of new information. Keraunos (talk) 20:23, 29 December 2010 (UTC)

In the context of a regular encyclopedia like the Encyclopedia Brittanica, it would probably be considered trivial to have like Wikipedia does articles on just about every major film, TV show, comic book, and video game that ever existed. But I think it is wonderful that all of this information is available on Wikipedia. Keraunos (talk) 20:23, 29 December 2010 (UTC)

If it's on the internet, it must be true. So fill up your drive with all these factoids, and let wikipedia to an encyclopedia. Dicklyon (talk) 00:41, 30 December 2010 (UTC)

Actually, according to this chart, if Moore's Law holds we'll only reach hard drive capacities of an average of 100 terabytes by 2020. If Moore's Law holds beyond 2020, it will take until about 2025 to reach an average hard drive capacity of one petabyte. Increase in Hard Drive Capacity Since 1980: Keraunos (talk) 13:43, 30 December 2010 (UTC)

But hard drives have nothing to do with Moore's law, and capacity and cost have nothing to do with wikipedia policy. So get with the program. Dicklyon (talk) 16:35, 30 December 2010 (UTC)

The problem with adding spurious unencyclopedic unsourced information is not that it takes up too much space, but that the Wikipedia project can only be trusted and useful if it applies some standards. It is difficult enough to maintain an encyclopedia even sticking to facts which can be verified: articles still need to be carefully written and tended to be readable and relevant. We should strive to use the best and most authoritative sources, to be careful with our definitions, and so on. At the point where any mention of anything by some character in a novel or TV show is considered relevant, articles balloon quickly out of control, and turn into unmanaged and unmanageable messes. Instead of arguing about which trivialities should be kept and which abandoned, it would be a much better use of everyone's effort to try to abide by the clearly described policies in WP:RS and WP:V and so on, and then do the actual hard research required to write clear and comprehensive articles about each subject. --jacobolus (t) 05:44, 31 December 2010 (UTC)

LAB co-ordinates and tagged RGB?

Alright, I'm late to this discussion. I just came here from "Ecru" (didn't know what it was before now), and I saw something like:

— Colour coordinates —

Hex triplet #C2B280

RGB (r, g, b) (194, 178, 128)

HSV (h, s, v) (39°, 27%, 77%)

Now, an RGB triplet without any reference standard is meaningless; I'm assuming HSV is similar. Equally, web triplets are arbitrary.

An L*a*b* specification doesn't have that issue. An image with an ICC profile attached doesn't have that issue (at least outside the cyan region :-).

Can we at least agree to list which RGB model is being assumed, provide the L*a*b* co-ordinates, and indicate if the color is "out of gamut" on sRGB?

(I know, this doesn't begin to address the question of "Which shade of green gets the "Green" standard label?". I know paint stores have that big book of color samples and shades that cover all the visible color range. Can that be used as a reference standard?)

Keybounce (talk) 02:07, 7 February 2011 (UTC)

RGB on the web is generally assumed to be sRGB, though most of the RGB triplets are nevertheless unsourced, and the sourced ones don't generally specify a color space. Are they sources for L*a*b* coordinates? Ecru has been source to a stamp collector's site that has been rejected as a WP:RS, so it should really just say unsourced. Dicklyon (talk) 02:39, 7 February 2011 (UTC)

Nearly all of the sources for color names given on Wikipedia are suspect, being mainly transformations of transformations by amateurs of something that once might have had some authority. Feel free to assume that the space is sRGB, but don’t expect that knowing the RGB color space will make a tremendous difference in the accuracy of the listed colors. –jacobolus (t) 22:34, 8 March 2011 (UTC)

Citation templates now support more identifiers

Recent changes were made to citations templates (such as {{citation}}, {{cite journal}}, {{cite web}}...). In addition to what was previously supported (bibcode, doi, jstor, isbn, ...), templates now support arXiv, ASIN, JFM, LCCN, MR, OL, OSTI, RFC, SSRN and Zbl. Before, you needed to place |id={{arxiv|0123.4567}} (or worse |url=http://arxiv.org/abs/0123.4567), now you can simply use |arxiv=0123.4567, likewise for |id={{JSTOR|0123456789}} and |url=http://www.jstor.org/stable/0123456789 → |jstor=0123456789.

The full list of supported identifiers is given here (with dummy values):

{{cite journal |author=John Smith |year=2000 |title=How to Put Things into Other Things |journal=Journal of Foobar |volume=1 |issue=2 |pages=3–4 |arxiv=0123456789 |asin=0123456789 |bibcode=0123456789 |doi=0123456789 |jfm=0123456789 |jstor=0123456789 |lccn=0123456789 |isbn=0123456789 |issn=0123456789 |mr=0123456789 |oclc=0123456789 |ol=0123456789 |osti=0123456789 |rfc=0123456789 |pmc=0123456789 |pmid=0123456789 |ssrn=0123456789 |zbl=0123456789 |id={{para|id|____}} }}

Obviously not all citations needs all parameters, but this streamlines the most popular ones and gives both better metadata and better appearances when printed. Headbomb {talk / contribs / physics / books} 02:37, 8 March 2011 (UTC)

AfD for color article

Input needed at Wikipedia:Articles for deletion/Carmine (color). --Noleander (talk) 06:08, 7 July 2011 (UTC)

More AfDs

I have nominated a few colo[u]r articles for deletion. All suggestions welcomed!

bobrayner (talk) 08:04, 9 September 2011 (UTC)

Problems with some colour articles

Hi,
I'm no expert on colour, but articles like Persimmon (color) are quite worrying. I've come across a lot of articles like this, where the source for the name is one catalogue out of many - perhaps just one entry in a sample book - but this source does not actually give RGB/HSV values so the author has got their own values by looking at some photo on the internet. It seems unlikely, to me, that one can get an accurate colour from one photograph out of many, as the exact colour in the photograph will be so dependent on lighting, post-processing, and of course the ripeness of that particular persimmon. Although there are areas where editors seen keen to fill in empty fields with the first value they can get their hands on, I think that an encyclopædia should not give exact values for something if the exact value is not actually known or knowable. What do the colour experts think? bobrayner (talk) 13:46, 9 September 2011 (UTC)

We also have articles like bondi blue which is apparently a colour from a photo of an apple iMac with an interestingly-coloured plastic case, but the source is claimed to be List of Crayola crayon colors. Which contains no such blue. bobrayner (talk) 13:46, 9 September 2011 (UTC)

Another example would be Byzantium (color); the source for the RGB/HSV values would appear to be a photo of a bracelet on the internet which is now a dead link. Various others simply have "unsourced" in the infobox but that hasn't stopped somebody claiming the colour exists and putting values in anyway. bobrayner (talk) 13:51, 9 September 2011 (UTC)

It’s been tried a couple times to get some support for deleting all of these articles, or merging them together and removing all unsourced or poorly sourced parts. It’s been hard to actually get a consensus agreement that that should be done though. Many editors agree with you that something should be done about the color name articles. Feel free to make a proposal. –jacobolus (t) 05:56, 10 September 2011 (UTC)

Well, from a layman's perspective, I don't see why the encyclopædia should contain articles that are poorly sourced (or not at all), often with made-up content; that content should be removed. Many of these colors seem to fall short of the GNG - one entry in a big catalogue, or one shade of Crayola, is not sufficient for notability (if the bar were that low, we'd have articles for every person in the phone book). If an article is otherwise good apart from one section which has been made up, it's probably best to blank that section. However, if this project has different ways of doing things, or different solutions... bobrayner (talk) 11:19, 10 September 2011 (UTC)

Thanks for taking on the tedious job of cleaning these up. I personally feel that it is good to have encyclopaedic cover of colors, in the old-fashioned sense of complete, so long as there is a valid source. In particular, I think there is virtue in having all the colors from a standard color chart (such as Crayola's), even if some individual colors are not so notable. So I would say keep Timberwolf (color), and lose the other animal colors. The main rational for deleting non-notable pages is that they are endless, unverifiable and hard to maintain, but I don't think that this is the case for colors from well-known color charts --- the total number of colors is limited and their values are known. Francis Bond (talk) 10:29, 11 September 2011 (UTC)

Does every shade of crayola and every entry in a colour catalogue pass the GNG in its own right? bobrayner (talk) 13:06, 11 September 2011 (UTC)

On a different note, I've found a very large number of colours which are sourced like this. The editor decided to write something about "Medium Candy Apple Red", browsed through google images, and turned up this picture of a toy car. The toy car is not even labelled with "this is a Medium Candy Apple Red car", it's not from a reliable source, and of course you certainly can't get reliable RGB/HSV values from a photo of a toy car which has been subjected to unknown post-processing. Nonetheless, large volumes of this stuff gets added to the encyclopædia. bobrayner (talk) 13:11, 11 September 2011 (UTC)

It's a mess. The magpie-like collection of "in human culture" trivia leads to huge sections like Orange (colour), and even to contradictions (the Tangerine and Orange articles claiming that a football team played in tangerine or orange shirts respectively). Some of those trivia sections might reasonably go in a dab; some already overlap with the content in existing dab pages. We also have overlapping articles like "Salmon pink" and "Salmon (color)" and so on. bobrayner (talk) 13:48, 11 September 2011 (UTC)

There's yet another problem: If a naïve google search returns two different uses of a name for a colour, such as Mikado yellow, they'll both be brought into the same article. The 1968 Lincoln automobiles painted "Mikado yellow" were a different colour to what's in the article, but hey, accuracy doesn't matter, the mission is to create as many colour articles and swatches as possible.

I'm getting increasingly frustrated here, and should stop commenting on this page before you're all bored to death by a litany of colo[u]r problems

bobrayner (talk) 14:14, 11 September 2011 (UTC)

And finally...

A lot of colours seem to be based on crayola. I have no idea how the Crayola colours got interpreted into RGB values - I doubt it's accurate but have been leaving them untouched for now - but quite often these crayola-sourced ones differ from the colour coordinates given by other sources. The Crayola fan has got around this difficulty by simply creating a new colour name in most cases. So, for example, at Cerise (color) we also have a "Deep cerise". The "Deep cerise" name doesn't actually exist in the Crayola list or elsewhere; it's just been invented so that a different shade of crayola can appear in the Cerise article with a different name (the alternative - that the editor omits an infobox for any crayola - seems to have been unthinkable). And yet the article tells readers that there is a notable colour called "deep cerise". I think this is deeply unencyclopædic, but it's happened across countless colour articles. bobrayner (talk) 16:54, 11 September 2011 (UTC)

Nearly finished.

Practically all the unsourced, contradictory, and made-up stuff has been removed from the "variations of...", and related colour articles, apart from those articles which are now at AfD. (Some may be closed as "keep"; I bow to the community's wisdom). Most of the overlap between different lists has been resolved, though some may remain.
I started removing the unsourced colour-blobs from the "shades of..." navboxes but somebody reverted a couple so I stopped that. Obviously we can't expect sources in the templates themselves, but I removed the blobs where there was no sane source for RGB values in the target article. My personal feeling is that we shouldn't show readers stuff which is not demonstrably true; others may disagree. What do y'all think?
Some of the "in human culture..." content actually belongs on dab pages; few of these articles had links to the respective dabs so I've added some hatnotes. Maybe if these articles were better linked, it would help address a couple of problematic tendencies.
It's probably worth keeping an eye on these articles; I believe much of the problematic content was added by an editor who had previously promised to stop adding unsourced stuff. I expect more cruft will be added in future.
I haven't really touched the crayola problem. What's the expert opinion of WikiProject Color? Are those RGB/HSV values for crayons reasonable? Do we even need separate listings for every crayola colour, or is the crayola list article sufficient? Many of the infoboxes in colour articles give different coordinates to those in the list of crayola colours, and they can't both be right. Perhaps they were guessed, by simply looking at a crayon.

Any suggestions / comments? What needs fixing next? bobrayner (talk) 10:20, 14 September 2011 (UTC)

Great, keep up the good work! In my opinion the “shades of” navboxes are worthless and should be scrapped. The RGB/etc. values for crayons are essentially completely arbitrary, even the ones that supposedly come from some Crayola website. –jacobolus (t) 06:58, 15 September 2011 (UTC)

The hexadecimal codes of the colors on the List of Crayola crayon colors are sourced from the hexadecimal codes on Crayola's website, so there is no problem with the sourcing of these colors. Keraunos (talk) 17:20, 14 September 2011 (UTC)
- The reference at the top of the list points to a set of cartoonish pictures of crayons. Do you have some other source which gives real, reliable RGB values? Anyway, why do some crayons have different RGB values on different articles? bobrayner (talk) 17:37, 14 September 2011 (UTC)
  Hold your cursor over the picture of the tip of each crayon in the picture at the bottom of that link and it gives the hex code. The reason some of the hex codes for the Crayola colors in the articles are slightly different from those on the Crayola website is that until about two and a half years ago, the hex codes of the Crayola colors on the List of Crayola crayon colors were listed ad hoc before Crayola posted the hex codes on their website. Then about two and a half years ago, someone immediately went through the list and matched all the colors to the newly entered hex codes on the Crayola website. Even so, there are usually only very small differences between the old Crayola colors and the new ones. It is a tedious task to go back to and change the Crayola colors in the articles that were sourced from the previous list to the new list now based on the colors from the Crayola website and no one has gotten around to doing it. If you want to do this, that would be very helpful, but don't stress yourself--this can be done over the next few weeks if you want to do it. Keraunos (talk)
  None of the “crayola” hex colors are at all reliable, whether from the crayola website or otherwise. None of them was made via any kind of reasonable process, and the colors produced by actual crayons don’t correspond at all closely with any of the colors used on Wikipedia. –jacobolus (t) 06:52, 15 September 2011 (UTC)
  ??? The crayola hex numbers are by definition correct per Crayola LLC. The hex codes are exactly whatever the company says they are per how the company believes their 'named' colors should appear on the web. VMS Mosaic (talk) 11:33, 17 September 2011 (UTC)
  - It's true that some of hex codes given on the Crayola website are not that accurate a representation of the actual color of the crayon. Ideally, someone should take a photograph of all the crayons under a full spectrum fluorescent light in a room with a skylight with north light, download the photo onto their computer, take color samples from each crayon from the photo, and post the results on an Internet website. However, until that happens, we have to work with what Crayola has given us. Keraunos (talk) 23:38, 17 September 2011 (UTC)
    - I suspect the Crayola numbers are what the color looks like on paper and not what the crayon itself looks like, so any such effort would be meaningless. VMS Mosaic (talk) 04:03, 18 September 2011 (UTC)
The unsourced color boxes that you removed were mostly "legacy" colors--colors entered into Wikipedia before November 2006. November 2006 was when an editor changed the color box template to require sources. Before that, no source was required. No one had gotten around to adding sources for these "legacy" colors. Since November 2006, most editors have been careful about sourcing the colors. Most of the unsourced color boxes that you removed have reliable sources from various color lists, so I will re-enter them sourced from these various color lists (the ones that have sources) over the next few months. I am familiar with the content of many different color lists. Keraunos (talk) 18:40, 14 September 2011 (UTC)
- Thank you for your good work in keeping everyone on their toes and improving Wikipedia! Keraunos (talk) 18:40, 14 September 2011 (UTC)
  - I don't think there is that much left to do unless you want to correct the color codes of the Crayola colors in the articles to the color codes in the List of Crayola crayon colors. But like I said, don't stress yourself--if you want to do it, that can be done over the next few weeks. Anyway, only a small percentage of the colors on the List of Crayola crayon colors are displayed in the articles, so if you want to do this, it shouldn't be too hard to do. Keraunos (talk) 18:40, 14 September 2011 (UTC)
Pretty much all of the sources used for color values on Wikipedia color articles are somewhere between unreliable and completely made up. Nearly all of them should be removed, or at least numerical values should be deemphasized, given that the exact values given are nearly meaningless given the poor processes which were used to generate them. –jacobolus (t) 06:56, 15 September 2011 (UTC)
I agree that the exact numbers can be deemphasized, but strongly believe each color article should display a representative color. I'd have no problem with a strong statement in each article that the displayed color is only representative (i.e., not definitive). VMS Mosaic (talk) 11:43, 17 September 2011 (UTC)
Just to be clear, I think the exact numbers should be given for colors which have an official source (e.g. Pantone, X11, HTML, Crayola). Of course the Infobox would have to define the color as Pantone, X11, etc. VMS Mosaic (talk) 04:03, 18 September 2011 (UTC)

I do realize that even agreeing on representative colors may be impossible (i.e., see the completely futile discussion at Green (and above) where editors refused even to consider compromising with the result being that the displayed color of green is randomly changed by whatever editor wanders by). If we can come to a compromise on the displayed colors, then I will defend them just as strongly (and equally) as I do the color/colour issue. VMS Mosaic (talk) 11:58, 17 September 2011 (UTC)
There are plenty of different color lists available to present a sample of each color. If someone doesn't agree with a particular rendition of a color, they are free to download a different sample of the color sourced from another color list. Keraunos (talk) 23:38, 17 September 2011 (UTC)

I think two things can be learned from this discussion

A lot of editors think that there is a lot of cruft on color articles
Several editors would prefer if things were sourced rather than axed outright

Even though BobRayner is technically correct about WP:V, I believe that he has misinterpreted this discussion as a mandate to ax a lot of content. Per policies such as WP:BEFORE, when an AFD or a large BOLD removal of content is done, the nominator/removal should try to see if sources can be found. I believe Bob has not made that attempt, and is too removal-happy. Keraunos and others point out that such sources are pretty easily accessible. So instead of axing content, perhaps Bob should try to source some of that content Purpleback pack 89≈≈≈≈ 19:44, 7 November 2011 (UTC)

A second batch of AfDs

A second batch of colour neologisms/dicdefs has been nominated for deletion:

Thanks for your time; bobrayner (talk) 18:03, 11 September 2011 (UTC)

Indigo ambiguity

Indigo is color #4B0082, but in the Blue footer template Template:Shades of blue it uses #6F00FF for "Indigo". That seems wrong. I dont doubt that 6F00FF is a fine color, but the template should not be using the name "Indigo" (with a link to Indigo) since that causes WP to give two different meanings to one color-name. Thoughts? --Noleander (talk) 18:17, 25 September 2011 (UTC)

This is part of a broader problem. For instance, one of the main sources for these RGB values actually gives multiple examples of colo[u]rs known by any given name. Which of those RGB values ends up being used in templates and infoboxes is just a matter of luck. I'd rather we remove content which is unsourced or which contradicts existing sources.

I tried removing unsourced colo[u]rs from the "Shades of..." footer templates, but somebody reverted. bobrayner (talk) 18:49, 25 September 2011 (UTC)

Okay, I can see how a given word like "Indigo" may have 2 or 3 different hues, even in official color sources. So, the Indigo article needs to mention all the various sourced shades. The X11 color names is an official source, and it has Indigo as #4B0082. I guess the question is where did the color #6F00FF come from? Ah, lo and behold, 6F0FF is Electric indigo, and it does have a source. So: the solution here is just to change the text "Indigo" to "Electric Indigo" in the Blue footer template. --Noleander (talk) 19:03, 25 September 2011 (UTC)

If you ask me, all the “shades of...” footer templates should be scrapped entirely as they are redundant and hard to keep in sync with color articles and the “list of colors”, etc., and the color name articles in general should be consolidated (into maybe 15 or 20 prominent color categories) since most of them are stubs with no potential for further growth. Any of those categories which currently has a “shades of” template might be a decent candidate for a consolidated article. At some point user:Wrad put a great deal of effort into cleaning up Green and made a Variations of Green article as a sort of overflow bin for all the color list trivia. The same should be done with other broad color categories. Independent articles could exist for color names which are actually notable (e.g. International Klein Blue), etc. –jacobolus (t) 23:30, 25 September 2011 (UTC)

Why use phrase "Varitations of ..." instead of "Shades of ..."

I notice that the articles on the shades of blue/red, etc are named Variations of blue, Variations of red, etc. That doesn't seem like the best phrasing: "Shades of blue" etc seems much more natural and concise. I looked in the Talk archives, and all I could find is Wikipedia_talk:WikiProject_Color/Archive_4#Move_to_Variations.2FVarieties_of_... ... which doesn't really say much. Can anyone explain the choice of the word "Variations"? --Noleander (talk) 19:22, 25 September 2011 (UTC)

... also: the footer navbox templates use the word "Shades .." as in Template:Shades of pink, Template:Shades of red, Template:Shades of blue, etc. Shouldn't there be some consistency? --Noleander (talk) 19:23, 25 September 2011 (UTC)

One reason is that “shade” has a variety of meanings related to color which are contradictory and can be confusing, e.g. tints and shades. “Variations” is a neutral term. –jacobolus (t) 23:33, 25 September 2011 (UTC)

I find "variations" rather off-putting: its use in that manner is a musical term. I've never heard "variations" in relation to colors, but maybe that is just me. I understand that "shade" may have some particular meaning within the world of color-cognoscenti, but for article titles the preference is to use common terms. Was there a discussion on this terminology choice in some Talk page? an RfC perhaps? --Noleander (talk) 23:57, 25 September 2011 (UTC)

Yes, neither name is particularly great. Better would be to call such articles “blue colors” or “red colors” etc., which would have the added advantage of emphasizing that color categories represent a whole range of colors, though this might also be confusing. There was some discussion someplace with several alternatives suggested. I don’t remember where. Maybe at this page? Maybe at talk:Green? –jacobolus (t) 00:14, 26 September 2011 (UTC)

"Red colors" etc is certainly a good contender: comes in first place in just about every WP:NAMINGCRITERIA criteria: conciseness, precision, commonness. Regarding past discussions of the name "Variations ..": I cannot find any discussion of the choice. I'm toying with the idea of a WP:Request for move on the five or six "Variations of ..." articles, but I don't want to be hasty. Does anyone object to a RM on this set of articles? If we go the RM route, I would post a notice in all the Talk pages, as well as on this project page. --Noleander (talk) 01:43, 26 September 2011 (UTC)

Summary of article name possibilities:

..anyone should feel free to add to this list of tentative candidates. If there is any issue over which candidate is best, we could do an RfC before the RM. --Noleander (talk) 01:47, 26 September 2011 (UTC)

You’re welcome to collect comments, but I think mainly the discussion is a bikeshedding distraction; I’m not going to participate much further on it. Cleaning up the content of color articles is much more worthwhile. –jacobolus (t) 10:46, 26 September 2011 (UTC)

Like I said, the word "Variations" just strikes me as wrong, so improving the titles is improving the articles. But it sounds like you have no objections to finding better titles, and you even would endorse the "Red colors" option, so I'll leave the proposal out here for awhile and see if there are any more thoughts. --Noleander (talk) 13:52, 26 September 2011 (UTC)

One of the arguments in favor of the "Shades of ..." nomenclature is that it is consistent with both (1) the numerous color categories found in Category:Shades of color; and (2) the numerous footer navbox templates found in Category:Shades of color templates. They all use "shades of ...". About the only downside to "Shades of ..." is that some color experts who know that the term "shade" also has a specific technical meaning (see Tints and shades) may get confused. On the other hand, any expert who knows that specific technical meaning of "shade" is probably savvy enough to not get thrown off by an article name "Shades of red". --Noleander (talk) 21:04, 26 September 2011 (UTC)

If there is no further input, I propose to do the move to "Shades of ..", based on the several factors: (1) most meaningful to readers; (2) consistency with Category:Shades of color templates; (3) consistency with Category:Shades of color; and (4) the other meaning of "shade" found in Tints and shades will be known only to color experts, who will have no problem understanding the alternative usage in "Shades of red.." --Noleander (talk) 17:23, 28 September 2011 (UTC)

Unless you get some further consensus, I’m opposed, and would suggest changing the names of the templates if the conflict seems like a problem. Using the word “shades” is inherently confusing here (since its widely used specific meaning w/r/t color is contradictory). In my opinion “variations” is an unambiguously better term than “shades”, but I concede that both are pretty bad, and a better term might be found. Cheers. –jacobolus (t) 21:39, 28 September 2011 (UTC)

Okay, it sounds like we are in agreement that the articles & templates & categories should be uniform. It sounds like the choice is "Shades of .." or "Variations of ..." (I think the "Red colors" choice is not so great, but maybe others would endorse it?). Maybe I'll start an RfC to get some other editors to share their thoughts. --Noleander (talk) 00:57, 29 September 2011 (UTC)

Circular sourcing

I noticed that one of perbang.dk's sources is en.wikipedia. It has even inherited some of en.wikipedia's dubious lists of Crayola colours. Should perbang.dk be trusted as a source for RGB values? bobrayner (talk) 19:32, 25 September 2011 (UTC)

Let's consider "Amaranth deep purple" as an example.

It was originally "sourced" to a photo of a flower; [2]
At some point, perbang.dk scraped RGB values and names from en.wiki; [3]
One editor (quite rightly) removed the swatches and infoboxes based on photos of flowers; [4]
Another editor restored the "Amaranth deep purple" infoboxes using values from perbang.dk. [5]

It's a disgrace. bobrayner (talk) 19:50, 25 September 2011 (UTC)

Well, if there are no other sources for "Amarath deep purple", you are certainly free to remove it from the Aramanth article. I'm not sure what you can do to prevent it from returning someday, other than putting Aramanth on your watch list. Now that the perbang.dk web site defines "aramanth deep purple" it is a bit beyond our control, unless you can contact the admin of that website and ask for it to be removed, but that looks unlikely since the web site appears to have the goal of capturing every color ever mentioned anywhere. :-) --Noleander (talk) 20:11, 25 September 2011 (UTC)

No, it should not be considered a reliable source. –jacobolus (t) 23:12, 25 September 2011 (UTC)

Agree: No, it should not be considered reliable. (Lost track of the original question in my prior comment). --Noleander (talk) 23:14, 25 September 2011 (UTC)