Colors Not Found In Nature

[RSPIII]

...we all understand what he means to say even if inaccurately stated.

I don't. Can you interpret it for me?

[Polar01]

RSPIII as to not take over Andrews threat, I would suggest if you have not done so already that you look at the Photography Tutorials say the one titled Colour Management and Printing.

Cheers: Allan

[davidedric]

It's an interesting question, but I think actually many questions, a lot of which have been picked up. I guess it will come down to "it all depends". If we limit ourselves to colours produced directly by plants and animals through chromophores, then I'd be surprised if there were not a limited number, at least in theory. Once you include refractive colours such as feathers and butterfly wings the number must shoot up enormously, and I am still talking visible light. I very much doubt if there are any odd significant "gaps".

Someone mentioned the "granuality" of colour, which I think is a quantum mechanical question relating to energy level differences in electrons around atoms. Here by definition there are a limited, though very high, number of theoretical possibilities. Limited because that is the essence of quantum theory.

Then there is the question of colour perception, as referred to by Andrew, which as he suggests is a huge subject.

All that is without considering the limits of our current technology in reproducing colour.

I think you could set this as a University exam question in a range of subjects from Philosophy to Physics, and get some very different answers.

Dave

[chauncey]

I don't. Can you interpret it for me?

It simply means that he's asking us to walk in his shoes, to interpret colors as he does...no big deal.

[pnodrog]

I think Andrew made it perfectly clear what he wishes to do. However we are unable to provide a precise answer. His own response to colours and what seems natural or not will probably be his best guide in achieving his objective. A bit like setting a white balance that in spite of all the theory is ultimately the photographers choice.

If photography is an art then aesthetics not theory should be the most important factor.....

[FootLoose]

Andrew, Margulis is mainly concerned with PP in the Lab colour space but in one of the videos (not sure which one) he does talk a bit about Lab's ability to describe colours not found in nature. If I recall correctly (and I probably don't) he refers to a colour that is simultaneously pure black and pure red, and is not a mixture of any colours in between.

I gathered from his remarks that he was doing more work in this area because he is interested in it.

[krisadamstv]

That is, software may produce certain colors that won't print, certain colors that won't transmit faithfully via the internet, and certain colors that are not found in nature.

you say, "... certain colors that won't transmit faithfully via the internet." Do you mean to say that certain colors cannot be digitized? Or that our current digital expressions of color (such as r=x, g=y, b=z) are not sufficient for some colors?

C'mon guys, let's cut Andrew some slack...we all understand what he means to say even if inaccurately stated.

I don't. Can you interpret it for me?

I know I am 3 years late to the party, but I can't leave this question unanswered.

My interpretation of what Andrew was saying with this comment has to do with "Web safe colours" (if a link to wikipedia is allowed: https://en.wikipedia.org/wiki/Web_co...eb-safe_colors)

Basically, the internet has a set amount of colours that are "safe" to use in the HTML code / CSS of a website. This traditionally doesn't include the full list of colours a monitor is capable of displaying.

[Manfred M]

I know I am 3 years late to the party, but I can't leave this question unanswered.

My interpretation of what Andrew was saying with this comment has to do with "Web safe colours" (if a link to wikipedia is allowed: https://en.wikipedia.org/wiki/Web_co...eb-safe_colors)

Basically, the internet has a set amount of colours that are "safe" to use in the HTML code / CSS of a website. This traditionally doesn't include the full list of colours a monitor is capable of displaying.

Welcome to CiC, Kris(?). Would you mind clicking the the "My Profile" tab at the top of this page and at least filling out your first name and where you are from on your profile?. We tend to go on a first name basis here at CiC; we are a pretty informal lot.

Andrew hasn't been back to see what was written after posting his question last year. I suspect that he was not referring to to web safe colours. Take a look around and tell me how often you see magentas and cyans, especially deep tones in those colours in nature.

[Steaphany]

Has anyone yet considered the distinction between Spectral Colors, what constitutes the visible spectrum, versus non-spectral colors or colors that we can see and differentiate but which can not be created from spreading white light into it's components ?

If you designate Natural colors as those from elementary spectral line emissions, colors created through the effects of refraction or diffraction then one possible definition for colors not found in nature could be the non-spectral colors. The problem is Purples, Magentas, and Pinks are found and see being the result of pigments.

[xpatUSA]

I'm investigating the phrase "colors not found in nature."

I'm not certain that your actual questions have been addressed so far, so I'll give them a try one-by-one:

How can I determine which colors or RGB settings would be considered colors not found in nature?

Pointer's Gamut is supposed to include all common natural colors:

http://www.tftcentral.co.uk/articles...ters_gamut.htm

Therefore, anything outside that gamut is a "color not found in nature."

Can I create such colors by using RGB settings?

The short answer is "YES"

The longer answer is "probably" with many caveats as stated by others in this thread. As I hope you know, an RGB color space has three primaries - quess what - red, green and blue. If any of those primaries is "inside" Pointer's Gamut, you can not create all such colors by using RGB settings.

How does anyone know if a particular color is in fact not found in nature?

Anyone can find out, as follows. Take a shot of the suspect color - with as neutral color settings as possible - especially saturation. Open it up in your editor and note the color-picker values.

Let's say they are RGB = 231, 140, 59 and let's say you're in the sRGB working space.

Now you go to Bruce's famous color calculator:

http://www.brucelindbloom.com/index....alculator.html

Set the Ref. White to D65, check the "Scale RGB" box and leave everything else at default. Enter the color picker values and click on the RGB button. By magic all the fields get filled with numbers. Look at the two fields next to the xyY button. Make a note, because these numbers represent a point on the CIE chart (the one with Pointer's Gamut on it that you've already carefully downloaded and saved somewhere). My example gives you x,y = 0.49, 0.41. If those co-ordinates are outside the said gamut. they are not natural.

Not everyone agrees that Pointer got it right but easily useable alternatives are few and far between.

[Manfred M]

Ted - I thought Pointer's gamut was supposed cover the whole human visual range, not the subset of "colours found in nature".

[revi]

Examples of "colours not found in nature" could be two of the primaries (green and blue) in the Prophoto colour space, which are outside the "horseshoe" of human visible colours...

Pointer's gamut reproduces the gamut of visible colours obtainable by diffuse reflection off a surface. That wouldn't take into account the colours found on e.g. some beetles, which are obtained by specular reflections on a grating. And that implies that Pointer's gamut covers only a subset of "colours found in nature".
The CIE "horseshoe" is supposed to cover the whole human visual range.

But that's only talking about colours we might see in nature. Which of these are actually found in nature? No idea. And then there's the whole problem of how e.g. insects see colours (UV gives a whole extra range of colours found in nature, e.g. on flowers...)

[xpatUSA]

Ted - I thought Pointer's gamut was supposed cover the whole human visual range, not the subset of "colours found in nature".

The whole human visual range is all those chromaticities (x,y values) found inside the spectral locus of the 1931 CIE xyY diagram. Pointer's Gamut is found well inside of that spectral locus and is used for comparing with the various color space gamuts (monitors, printers, etc) to determine their efficacy in presenting a natural scene.

Reading the whole of the linked article will amplify the above statement. There, you will find lots of diagrams of all kinds - all enclosing Pointer's by a good margin.

http://www.tftcentral.co.uk/articles/content/pointers_gamut.htm

Revi has provided a clarification re: the coverage by the gamut of "colors found in nature" and of course his mention of flower spectral reflectance is quite relevant. In fact, in "nature" - according to Revi, it is quite likely all possible natural colors, and type thereof, will fill the entire CIE diagram. Indeed in "nature," there is much radiation that vastly exceeds the human visual gamut and is not even "color", by definition.

None of which helps the OP much, I would have thought

[AlanP]

The CIE diagram contains all colors visible to the human eye. The color space is xyy where x and the first y are imaginary colors used for mathematical purposes. The second y is luminance.

If you adjust colors in sRGB or ARGB you cannot produce imaginary colors. If you adjust colors in ProPhoto or LAB you can produce imaginary colors.

If you use Photoshop in these color spaces Photoshop will render the colors to the monitor color space. You will not see anything strange because of the rendering.

Since the colors being rendered are imaginary they are not defined and the resulting colors are arbitrarily chosen (by Photoshop). When the same imaginary colors are rendered by the printer the colors a different arbitrary rendering may take place and the printed colors may be quite different than the monitor colors. In both cases you will see something that contains only real colors.

Alan

[xpatUSA]

Has anyone yet considered the distinction between Spectral Colors, what constitutes the visible spectrum, versus non-spectral colors or colors that we can see and differentiate but which can not be created from spreading white light into it's components ?

If you designate Natural colors as those from elementary spectral line emissions, colors created through the effects of refraction or diffraction then one possible definition for colors not found in nature could be the non-spectral colors. The problem is Purples, Magentas, and Pinks are found and see being the result of pigments.

Good points, Steaphanie.

Perhaps it would be more simple if firstly we say that any emittance outside of the spectral locus and the line of purples is not a 'color', by definition. Then secondly we can say that any emittance in nature could occur inside of the spectral locus and the line of purples.

If both statements are true then the questions in the OP have been fully answered by no less an august body than the CIE itself !! Therefore, the xyY or Yuv diagrams say it all with no more to be added.

Inside = a color that can be found in Nature. Outside = not a color. Easy, ennit?

[Simon Garrett]

My goodness, this is a can of worms! Trouble is, people mean different things by "colour".

There are many definitions of colour - some useful, some less so - but here's the one I find most useful.

Colour is what human beings perceive (in our brains) when some combination of wavelengths within the visible spectrum (roughly 400-700nm) impinges on our eyes. The resulting stimulation of the L, M and S cone cells is further processed before we perceive "colour". This perception varies from person to person (even those with "perfect" colour vision"), so we need to consider some average person. This idea of colour is meaningless except to human beings.

It's about perception, not wavelength, pigment, spectral distribution, RGB value or whatever.

By that definition, there can be no colour not found in nature. The CIE diagram shows all possible colours at one brightness level. The points around the curved edge correspond to colours that can be created by single wavelengths of light. These are the colours of the spectrum - or colours of the rainbow. All other points are non-spectral - they must be created by a combination of wavelengths. Usually there is no unique combination of wavelengths to make a colour (except those on the curved edge).

Points outside the horseshoe shape are not colours. They are not "colours not found in nature", they are simply not colours.

Hence if it's in the horseshoe shape it's a colour of nature and can be created. If it's not in the horseshoe it can't be created, but it's not a colour!

The Pointer's Gamut is not "all colours found in nature", but might be regarded as "all common colours". It's a subset of all the colours in the CIE diagram consisting of those that can normally created by reflected light. As virtually all the colour we see is reflected (except direct sunlight, direct starlight or direct artificial light), it's a pretty good approximation of "all common colours". In other words: colours outside the Pointer's Gamut generally can't be created by pigments - only by directly emitted light. Possible on a suitable monitor, not in a print.

However, as my definition of "colour" is not the only one, some might disagree.

[Simon Garrett]

Has anyone yet considered the distinction between Spectral Colors, what constitutes the visible spectrum, versus non-spectral colors or colors that we can see and differentiate but which can not be created from spreading white light into it's components ?

Good point. In my previous post, I described "non-spectral colours" as those that can be created only by more than one spectral wavelength.

If you designate Natural colors as those from elementary spectral line emissions, colors created through the effects of refraction or diffraction then one possible definition for colors not found in nature could be the non-spectral colors. The problem is Purples, Magentas, and Pinks are found and see being the result of pigments.

Depends what you mean by "found in nature", of course. Spectral colours - corresponding to single spectral wavelengths - are arguably pretty rare in nature. Direct sunlight is a broad band of wavelengths and most artifical light is a combination of wavelengths. Reflected or difracted light is likely to be a subset of the sunlight or artificial light impinging on the reflecting or refracting medium, so is also almost invariably a range of wavelengths.

People posting here have come up with a variety of ways regarding "colours of nature", all of which may be useful in the intended context. It's a bit of a minefield.

[xpatUSA]

Good points, Steaphanie.

Perhaps it would be more simple if firstly we say that any emittance outside of the spectral locus and the line of purples is not a 'color', by definition. Then secondly we can say that any emittance in nature could occur inside of the spectral locus and the line of purples.

If both statements are true then the questions in the OP have been fully answered by no less an august body than the CIE itself !! Therefore, the xyY or Yuv diagrams say it all with no more to be added.

Inside = a color that can be found in Nature. Outside = not a color. Easy, ennit?

As I read my above post again, it seems perhaps a bit too technical. Please read the one following it for a clarification in layman's terminology ;-)

[xpatUSA]

The CIE diagram shows all possible colours at one [zero] brightness level.

Here it is at all levels of lightness (hold onto your seat):



Do notice the change in horizontal cross-section area with increasing Y !!

Even at "half-brightness" (going upward) the number of available chromaticities (x,y values) is much less. I think the line up the middle is the equal energy white point.

This animation also shows how limited is the information offered by the standard diagram that everyone, including myself, has quoted here. That is to say, what may look like a 'color' in the 2D view can become a 'not-color' as Y is increased ! So, wot we want now is a 3D Pointer's Gamut

Well, here's sRGB just for info and Pointer's is a bit "bigger" than that . .

[Simon Garrett]

Thanks for that, Ted. Very revealing!