Talk:Color

Please note that the question of whether to spell the subject of this article as color or colour is covered by the Manual of Style and American and British English spelling differences. Proposals about spelling should be raised at Wikipedia:Village pump (policy). This talk page is for discussing the subject matter of its article. Comments about the spelling of color or colour may be regarded as off-topic and removed. Specifically: Wikipedia:Manual of Style#National varieties of English: “ Stay with established spelling -- If an article has been in a given dialect for a long time, and there is no clear reason to change it, leave it alone. Editors should not change the spelling used in an article wholesale from one variant to another, unless there is a compelling reason to do so (which will rarely be the case). Other editors are justified in reverting such changes. Fixing inconsistencies in the spelling is always appreciated. ” Also: Wikipedia:Perennial proposals#Enforce American or British spelling: “ Proposal: For consistency's sake, we should pick one style of spelling (British or American, generally) and stick with it. Reason for previous rejection: It is wildly impractical and there is no agreement on which style should be chosen, which has in the past resulted in repeated, needless edit warring. ”

Wavelengths of color.

The ranges given for the visible spectrum in this article are given by a reference to an atmospheric radiation textbook, not an optics textbook. I would like to propose a change of the values to fit optical or psychological scientific definitions. I have two science textbook references to a broader range for the visible spectrum, but I still haven't found digital examples of these values to make reference to. As an additional complication, one reference uses indigo and the other does not. The definition of these colors is, in many cases, chosen arbitrarily as explicitly expressed in the reference given for the current values listed on this page, which, sadly do not actually match those stated in the reference article. Also, neither of these references, nor this article, give any presence to Cyan, which may be the light-blue and blue-green described in the Russian and Chinese perspectives, respectively... but as they have no reference it is difficult to confirm.

The values I've found (in nanometers):

380-436 = Violet, 436-495 = Blue, 495-566 = Green, 566-589 = Yellow, 589-627 = Orange, 627-780 = Red (from Physical Aspects of Colour, P.J. Bouma, 1949)

390-446 = Violet, 446-464 = Indigo, 464-500 = Blue, 500-578 = Green, 578-592 = Yellow, 592-620 = Orange, 620-770 = Red (from The Principles of Optics, Hardy and Perrin, 1932)

I feel the need to point out that discrepancies such as these exist within the scientific community today largely due to a lack of consensus... an absence of communication... This is my attempt to reach out for communication. Perception of color is a biological reaction to a specific range of vibrations in the electromagnetic spectrum. If we fail to accurately understand how we perceive that range, how shall we ever recognize the underlying patterns? Given the ranges I listed above, if one translates them into frequencies through the use of the speed of light as their wave propagation velocity, there is a relationship that becomes apparent: the perceived range of color by humans almost perfectly fits into a binary exponential "octave".

I believe it is important to recognize the sacredness of this gift. Color brings emotion, definition, and, in most cases (WHO estimates 2.6% of the human population as visually impaired), to a large extent, defines our reality. In these days of endless color stimuli, to reach out for renewed recognition of the importance color has played in our evolution as a species, as a society, as a global consciousness seems paramount. The interrelationship between duality and color... between duality and existence itself... is ever present, and for the human race to go on ignoring the underlying patterns of our existence by allowing discrepancies such as these to propagate in our collective consciousness seems a terrible crime.

Luminaux (talk) 21:51, 14 December 2010 (UTC)

The assignment of color names to wavelength ranges is essentially arbitrary. We just need to say which source we're following, and follow it. Most don't include indigo, so I'd leave that out. None include cyan, because that's usually thought of as non-spectral, the color of a wide range of blue and green wavelengths combined. Dicklyon (talk) 00:20, 15 December 2010 (UTC)

Somewhere I saw a table of the association of color names to wavelength ranges, by several different authors. Such a comparison might be useful somewhere, but it would take making a slightly clever graphic to properly show it, I think. –jacobolus (t) 05:22, 15 December 2010 (UTC)

Assigning names to the palette of colours is subjective, and varies in different languages. The article is primarily about the physics of colour, and it is misleading to imply that there are generally agreed boundaries for defining wavelengths associated with the words. The spectrum table in the article gives approximations, other sources may vary.--♦IanMacM♦ ^{(talk to me)} 07:58, 15 December 2010 (UTC)

Hmm, I’m not sure that’s quite fair. There are some relatively authoritative assignments of wavelengths to color names. For example, (a) there have been psychophysical experiments to find at which point some large number of observers most commonly put category boundaries when presented with a continuous spectrum, and (b) this 1943 JOSA paper by Kelly originated a very commonly reproduced diagram (for example, I’ve seen variants in documents put out by Adobe and other major companies) assigning color names similar to those used in the ISCC–NBS system, which was a cooperative project sponsored by all the major interested parties in government and industry (in printing, textiles, manufacturing, photography, agriculture, &c. &c.) in the US at the time. –jacobolus (t) 03:12, 20 December 2010 (UTC)

Here is the diagram from the Kelly paper, and here is a table from that paper which compares 24 different assignments of color name to wavelength made between 1866 and 1939. –jacobolus (t) 03:18, 20 December 2010 (UTC)

(In case it’s unclear, you’re right of course that it’s subjective and in some ways arbitrary, and we don’t want to imply that there’s some scientifically precise way to draw exact boundaries between color names, but I don’t think it’s fair to say that there’s no agreement, or that some sources aren’t better for Wikipedia’s purposes than others.) –jacobolus (t) 03:22, 20 December 2010 (UTC)

Thanks for all the quick responses, I'm happy to see that there are active editors on this page. I realize that the names and their corresponding "regions" are subjective and/or arbitrarily chosen, but the overall range of color wavelengths perceived by the average human eye should be much more accurately defined. The values I listed above define the red region ending at 770 or 780nm... this is very different from the generalized 700nm currently listed on this page. ~ Luminaux (talk) 06:18, 18 December 2010 (UTC)

The sensitivity of the L cones in nearly zero by 700 nm, but still not quite zero by 770 nm. Anything longer than 660 nm that gets detected is red, but not much gets detected beyond 700 nm. This much is easy to quantify, but there's no discontinuity, so different people specify nominal boundaries different ways. Dicklyon (talk) 17:00, 18 December 2010 (UTC)

This just further illustrates my original point... if two fields of optical science can't agree on the outer boundaries of our color perception, how is the general public ever supposed to understand it? I've seen the graph of the cone response, but the discontinuity is still there, from my perspective. The field of optical crystallography requires active use of the human eye and I find it extremely difficult to believe that they would note a larger range if they couldn't reliably measure and discuss it. And as far as the "nominal boundaries" that lay within the overall range... they could be defined by mass-consciousness-testing. The general public could be enabled to respond to a test on what they perceive a series of specifically defined wavelengths to be, essentially casting their vote on what they feel towards different wavelengths. The only problem is having a solid testing platform spread out across the entirety of human consciousness... but we're getting close already... most of the interface screens we use to access the internet embody the RGB color system as defined by the CIE 1931 color space, don't they? How accurate are LEDs and LCDs at producing these three singular wavelengths of Red, Green, and Blue (and Yellow apparently now too)? Why is there only a small reference to the CIE in this article? Wasn't that a public forum to define spectral lines? ~ Luminaux (talk) 18:23, 19 December 2010 (UTC)

This article needs massive expansion and much better explanation, in my opinion, but getting it to a state where I’d be happy with it would take a tremendous amount of work, and there is plenty else to work on, both in Wikipedia and elsewhere, so it’s not high on my list of priorities. To answer your question, no, computer displays/televisions/etc. don’t really try to “embody the RGB system as defined by the CIEXYZ color space”. Explaining why that’s a bad description is beyond the scope of this short comment though. –jacobolus (t) 03:36, 20 December 2010 (UTC)

The boundary of "red" in the Kelly paper (and this is fairly typical) is defined as something like the complementary dominant hue to 494 nm. This is typically written something like 494c. Indeed, a “unique red” color, which appears to have no part of yellow or blue in it, is extraspectral, something like 493c. –jacobolus (t) 03:32, 20 December 2010 (UTC)

A recent documentary about the latest research on colour perception is quite relevant to this. It basically found that colour is perceived by the brain based on experience and the usefullness of knowing of the existance of the colours. For example the Himba of Namibia have only five words to describe different colours and they can not see blue very easily, variously describing it as black (the sky) or white (the blue of water). When asked to pick out the odd coloured card when presented with 11   coloured cards and one   coloured card they could not easily pick the odd card and sometimes said they couldn't see a differently coloured card. Other experiments they conducted indicated that the colour an individual sees for a particular wavelength is not exactly the same colour as that seen by another individual for the same wavelength. I cant remember the scientists names but the paper is written and an editor may be able to find it. Wayne (talk) 09:31, 14 December 2011 (UTC)

S receptor peak?

In various places I've seen the peak of the S receptor range given as 420nm (this article ) and as 420-440nm (in cone cell). The response graph seems to show the response peaking at 450nm, which is considerably different. What's up with that?

I see estimates from 419 to 450 nm here. Dicklyon (talk) 06:37, 26 February 2011 (UTC)

The discrepancy, as far as I can tell from reading Stockman & Sharpe (2000), which seems to be the best source to date, depends on what is being graphed/measured. The spectral sensitivity of just the S-cones peaks at about 420 nm, but the lens absorbs some light before it ever reaches the cones, especially at short wavelengths, and so the “cone fundamental” for S-cones (a model for the sensitivity to light coming through the eye) has its peak at something like 440 nm. There’s also inter-observer variation, changes to lens/macular color as people age, variation between the precise center of the retina and areas further out, some genetic variations in the cones, differences between “typical” observers and the color blind, and possible effects based on experiment design, equipment used, random noise, etc. –jacobolus (t) 08:58, 26 February 2011 (UTC)

Indigo/Violet

I propose that instead of eliminating indigo while keeping violet, which literally leaves out the section of the spectrum named indigo, we use the term that describes both colors, purple. -Calibas (talk) 05:59, 26 February 2011 (UTC)

Better read purple; it's non-spectral. Dicklyon (talk) 06:33, 26 February 2011 (UTC)

I figured something like that was the case, though in common usage it says I'm correct. I still think it's a good idea to change it to something less confusing for 90% of people who read the article. We can even rename it to purple, and then add a note explaining why for the small fraction of people who don't consider violet to be purple. --Calibas (talk) 17:02, 26 February 2011 (UTC)

The word “purple” is used as a catch-all for colors between red and blue  . In describing a hue circle, it's an appropriate label. The word “violet” is used both to describe colors nearer to blue than red  , and also to describe very short wavelengths. This article is certainly misleading in putting such heavy emphasis on relation of colors to wavelengths up front, since that isn't really the way color is experienced, but in the context of naming wavelengths with color terms, “violet” is appropriate but “purple” would not be. The real solution is to add more material about the organization and experience of color to the front of this article, before the “physics” section. –jacobolus (t) 17:28, 26 February 2011 (UTC)

black and white

I cruised on in here to resolve a workplace argument about the status of black and white as colours. Sadly the article didn't resolve the dispute (seriously guys - what else is Wikipedia for? :)

If someone knowledgable enough can address this question within the article in an encyclopediac fashion, it would be a welcome addition. Manning (talk) 00:36, 27 January 2012 (UTC)

That black and white are clearly “colors” is implied by every reasonable definition of “color”. Cheers. –jacobolus (t) 08:07, 27 January 2012 (UTC)

Can you come to my office and tell that to the guy who sits next to me? :) Manning (talk) 09:20, 27 January 2012 (UTC)