Primary colors

[EdSawyer]

I have to put in a plug for my favorite color head, which is additive... The minolta Beseler 45A, which uses additive light from pulsed xenon flash tubes as a light source. Has a great analyzer built in, too. Hard to beat for $100-200 on Ebay, when it sold new for $2500-3000 back im the 90's. i use one and love it.

 

[lxdude]

Red, yellow, and blue have NEVER been classified as primaries, and never will be.

Yellow has been classed as a primary.

http://www.colormatters.com/color-and-design/basic-color-theory

http://en.wikipedia.org/wiki/RYB_color_model

 

[Bill Burk]

The "detectors" in the human eye are sensitive to blue, green and red. Depending on the proportions of blue, green and red stimulating these receptors, the brain interprets these mixes as all the colours we see.

You might find it interesting to find that the detectors in the human eye are NOT well-tuned to the primaries Red and Green as pure colors. The difference in Red/Green Cone wavelength sensitivity is very small. But the brain interprets the lousy Red/Green signals it gets as if they were pure colors. Of course in some people there's that defect in these Cones responsible for color blindness.

 

[bernard_L]

David, the additive colors for printing are cyan, magenta, yellow (and black)

These are subtractive primaries, not addditive.

 

[Steve Smith]

If both are right then can someone explain in straightforward, relatively simple terms how these can be reconciled and if the paint colour primaries are wrong then what is the flaw and why is it that red, yellow and blue are wrong?r

It's explained in the link I posted earlier. Red cannot be a primary as it can be made by mixing yellow and magenta paint or ink. Blue can be made by mixing cyan and magenta. Yellow is a primary.


Steve.

 

[analoguey]

There must be a Calvin and Hobbes cartoon about this somewhere? Or a xkcd one?

Sent from Tap-a-talk

 

[Bill Burk]

Red, Yellow and Blue are not necessarily "wrong" as a set of paints to play with.

But they don't cleanly mix as wide a gamut of colors as the subtractive primaries used in graphic arts... Yellow, Magenta and Cyan.

 

[Mr Bill]

If both are right then can someone explain in straightforward, relatively simple terms how these can be reconciled and if the paint colour primaries are wrong then what is the flaw and why is it that red, yellow and blue are wrong?

I've been trying to think of a relatively simple way of explaining this. Perhaps one way is to say there is really one set of primary colours as far as our eyes are concerned, and that the second set emerges as a way of "filtering" them.

I don't think that there IS a simple, straightforward way of explaining this. Well, at least in not less than an hour or so, with open-minded people. You have to be able to temporarily set aside what you "know" about this from color photography.

I see two ways of "knowing" about color (beyond what we directly see). The first uses what is known as "colorimetery," and one of the best "simple" explanations, to my way of thinking, was in the 1963 Feynman Lectures on Physics. Specifically chapter 35, "Color Vision." A brief excerpt,

Now a question is, what are the correct primary colors to use? There is no such thing as “the” correct primary colors for the mixing of lights. There may be, for practical purposes, three paints that are more useful than others for getting a greater variety of mixed pigments, but we are not discussing that matter now. Any three differently colored lights whatsoever can always be mixed in the correct proportion to produce any color whatsoever.

Be aware that Feynman is not saying exactly what it seems like. With 3 physical lights, there are always some colors that cannot actually be matched; it would be necessary to use negative values of some light(s). However, shining some of this light on the reference sample is mathematically equivalent to a "negative light" on the other side, so this demonstrates the mathematical "match."

He expounds on this...

We may ask whether there are three colors that come only with positive amounts for all mixings. The answer is no. Every set of three primaries requires negative amounts for some colors, and therefore there is no unique way to define a primary. In elementary books they are said to be red, green, and blue, but that is merely because with these a wider range of colors is available without minus signs for some of the combinations.

My second way of "knowing" is based on the spectral nature of light reaching the eye. A group called the CIE has a thing called the "Standard Observer," where something called "color-matching functions" have been experimentally determined. If one knows the spectral makeup of the light reaching the eye (multiply the spectral makeup of the light source times the spectral transmission of the filter), then it is possible to specify where this light would fall, in some color space, with respect to an adapted "white point."

The best "simple" explanation (to me) that I know of is in a booklet (1998? about 100 pages) called the "Color_Primer" by Fred Bunting, that I believe can be found on Xrite's website. I think most photographer's will find their eyes glazing over by about page 10, but if you can make it through, you will have a much deeper understanding of things.

I personally grew up with the system of color photography, with red, green, and blue (additive) primaries, and cyan, magenta, and yellow (subtractive) secondaries, and I had a very hard time getting over that. My breakthrough was when I first realized that there is no rigid physical definition of those colors, so they can sort of be whatever one wants them to be. The more important thing is that one designs the spectral sensitivities of the color paper, for example, to be matched with the dyes in the film, which are ideally the same as the printing filters.

ps: I don't know if I have the patience to argue these issues, which is why I give the references. But I'll hang in for a while if anyone cares to discuss. Thanks for reading, assuming you got to here!

 

[David Allen]

As far as I know painting class instruction hasn't changed and painting teachers still persist with their "theory".

Red, yellow, and blue have NEVER been classified as primaries, and never will be.

It is really very simple. In terms of light, red, green, and blue are the primary colours and cyan, magenta, and yellow are the (subtractive) secondaries. As we require light to see something, the true primary colours are red, green and blue.

Where it gets a little more complex is in understanding how we perceive colours.The surface of a red apple is reflecting various wavelengths contained in light that we see as red and are absorbing the rest of the spectrum contained within light. Similarly, an object appears white when it reflects all the wavelengths contained in light and black when it absorbs them all.

In the history of painting, red, yellow, and blue have always been the primary pigments (and depending upon media have additionally had white or black pigments added). Note the word pigments here. Pigments work like any other object that we perceive as having a colour in that they absorb certain wavelengths of light and other wavelengths are reflected or scattered, which cause you to see the colours that you perceive. In short, a pigment appears to demonstrate a certain colour because of which parts of the light spectrum it reflects and/or absorbs.

So to repeat, light determines the primary colours by which we see our world. How that light is reflected by an object determines how we perceive the colour of an object. The reason painters believe that the primary 'colours' are red, yellow, and blue is because these pigmens, dyes or inks can be combined to create a wide spectrum of 'colours'. These are not, however, colours in the true sense but rather reflective surfaces that, when viewed in white light, will appear to us as a certain colour - i.e these surfaces will reflect certain parts of the light spectrum and absorb other parts of the light spectrum and we will interpret that as a light blue, or an orange or pale green, etc.

Bests,

David.
www.dsallen.de

 

[Mr Bill]

Correct. It's not like the eye responds to three specific wavelengths or anything that "tuned". But it is blue, green and red.

No, if you trust this graph found here, you'll find that Bill Burk is more correct. The "red" and "green" sensing functions overlap much more than they are separate.

http://en.wikipedia.org/wiki/Spectral_sensitivity

In that sort of business, it's traditional to designate with S(hort), M(edium), or L(ong) wavelength, or the Greek letters in place of r, g, and b.

 

[David Lyga]

OK, I started controversy (as I had presumed), but is there an objective, mathematical solution here? One that looks objectively at nanometers (billionths of a meter) of wavelength? This might be a way to start.

The difference in nanometers (nm) between blue (475nm) and green (546nm) is 71nm. The difference between blue (475nm) and yellow (580nm) is 105nm. Now, the difference between red (700 nm) and green (546nm) is 154nm. However, the difference between red (700nm) and yellow (580nm) is only 120nm (a lot closer to 105nm than 71nm is to 105nm).

What we have here is a closer alliance (from these particular 'poles') with yellow than with green, as the differences between those poles (the near-ends for human visual perception) blue and red, to the center (yellow) is only either 105nm or 120nm. Thus, it seems (to an 'ignorant' observer, such as myself) that green (546nm) is 'too close' to blue (475nm) to hold a discrete place in the category of 'primary'. However, is blue a true polar limit? NO!

On Wikipedia, for primary colors, the following quote is revealing: "No real display device uses such primaries, as the extreme wavelengths used for violet and red result in a very low luminous efficiency". So, (David Lyga, here) let's seek a new 'pole': violet's wavelength is 436nm, shorter than that for blue (475nm). Using these criteria for extending the 'poles', violet (436nm) is 264nm away from red (700nm). The mathematical 'midpoint' here would be 568nm, 132nm away from either 'pole'. This 568nm midpoint is conveniently located between the yellow (580nm) and green (546nm) wavelengths. So, with mathemantics being the only determinant for mitigating such hue-ignorance, maybe I have forced the issue to be answered by using mere numbers and not visual acuity (which can be notoriously indeterminant).

It looks as if there is a case for both yellow and green to join the category of primary. Should the primaries be deemed corrected as violet, yellow-green, and red? Indeed, the best answer would probably be a conflation of yellow and green (what would that color be called!!!???) In summation, probably color is more subjective that we want it to be, and different occupations call for different 'conveniences'. The high school art teacher was making $80K per annum (you who own homes are paying for this in taxes!) in a rather wealthy school district. One would HOPE that he was proficient in his knowledge! (But, I'll bet he could not tear this apart with numbers.) - David Lyga

 

[Photo Engineer]

Here is one simple graphical representation of color http://en.wikipedia.org/wiki/Munsell_color_system ; and here is some of the math behind it all http://en.wikipedia.org/wiki/CIE_1964_color_space.

Now, a gross simplification. This is wrong but can work for us here. Additive systems are used with devices that emit light such as a TV set. Subtractive systems are used for reflection prints and newspapers. You can see holes in this already, but it is a fair starting point.

PE

 

[DREW WILEY]

Steve - magenta contains blue. It's half blue. Therefore you CANNOT mix it with ANYTHING to obtain pure red. Cyan is only half blue, the other
half is green, across the color wheel, and therefore it CANNOT ever produce true blue with it, no matter what you mix it with. My gosh. First you folks set up units of measurement based on how far some drunken Druid priest could throw the head of an ox, and now you want to rewrite the laws of physics. Or maybe you just enjoy some form of recreation which entails walking into art stores and changing labels on the tubes of pigment.

 

[DREW WILEY]

There is no controversy, David. It's simple. If you want to print additive color, you need as pure a possible RGB. If you want to print subtractive, the purer the CYM, the better. With subtractive there tends to always be a little white light contamination spilling over into all
three, so it cannot be engineered quite as pure as additive. And there are different ways of making filters. I happen to use the "sandwich"
method of pairing a set of dichroic filters, meaning three sets, which each allow just a narrow band of R,G,or B to pass through. You can't
make the bandwidth too narrow, however, because the peak sensitivity of dichroic filters changes with temperature. In an additive system,
if you use very dense narrow-band single filters, that means a lot of heat, and pretty soon you start chasing your tail. I know someone who
made that mistake. Not that I expect you'll make an additive colorhead yourself. Durst came damn close to it just before they ceased mfg
of all their true industrial enlargers. The leftovers filter and electronic components seem to have gone to the NSA, who can obviously afford
things like big custom enlargers. It's tricky. But if I were to substitute a yellow filter for the deep green one, it would be impossible to correctly balance color film. Not hard. Impossible. Light and color just don't work that way.

 

[Mr Bill]

On Wikipedia, for primary colors, the following quote is revealing: "No real display device uses such primaries, as the extreme wavelengths used for violet and red result in a very low luminous efficiency". So, (David Lyga, here) let's seek a new 'pole': violet's wavelength is 436nm, shorter than that for blue (475nm). Using these criteria for extending the 'poles', violet (436nm) is 264nm away from red (700nm). The mathematical 'midpoint' here would be 568nm, 132nm away from either 'pole'. This 568nm midpoint is conveniently located between the yellow (580nm) and green (546nm) wavelengths. So, with mathemantics being the only determinant for mitigating such hue-ignorance, maybe I have forced the issue to be answered by using mere numbers and not visual acuity (which can be notoriously indeterminant).

David, I think you're just blowing a lot of smoke, and not realizing it. You can't just make up wavelengths to use and ignore what the human eye sees. Let me invent an example. You have suggested using "mere numbers" to come up with a "568 nm midpoint," and I presume you are suggesting to use that wavelength as one of the primary colors. (Am I correct in this presumption?) Ok, now let's say that someone has studied human vision, and discovered that humans could not see this wavelength. Of course this is not true (see the link in my prior post), BUT WHAT IF IT WERE TRUE? If a human could not see this, it clearly could not be used as a primary color! This should be convincing evidence that one cannot ignore human vision and simply calculate wavelengths.

By the way, Caltech has made the Feynman lectures available online for free. People learn differently, but if your brain works in the right way to follow along, I think it's a terrific explanation. If it's too tedious, section 35-3 is the part I previously quoted from, and (I think) especially worth reading. I doubt that the high school art teacher can explain as well as the physicist did 50 years ago. What do you think?

http://feynmanlectures.caltech.edu/I_35.html

 

[Photo Engineer]

The chromaticity diagram in that Feynman Lecture is a good example of the problems with defining color. There is a discontinuity across the bottom from the ~400 to ~700 nm region in which the curves do not join. I have used other sources to define this position which is the combination of blue + red (magenta).

PE

 

[BrianShaw]

Just get a basic primer on color theory. ...

And/or get a PhD in this topic if you want to know how really complex a topic it is!

 

[David Lyga]

David, I think you're just blowing a lot of smoke, and not realizing it. You can't just make up wavelengths to use and ignore what the human eye sees. Let me invent an example. You have suggested using "mere numbers" to come up with a "568 nm midpoint," and I presume you are suggesting to use that wavelength as one of the primary colors. (Am I correct in this presumption?) Ok, now let's say that someone has studied human vision, and discovered that humans could not see this wavelength. Of course this is not true (see the link in my prior post), BUT WHAT IF IT WERE TRUE? If a human could not see this, it clearly could not be used as a primary color! This should be convincing evidence that one cannot ignore human vision and simply calculate wavelengths.

By the way, Caltech has made the Feynman lectures available online for free. People learn differently, but if your brain works in the right way to follow along, I think it's a terrific explanation. If it's too tedious, section 35-3 is the part I previously quoted from, and (I think) especially worth reading. I doubt that the high school art teacher can explain as well as the physicist did 50 years ago. What do you think?

http://feynmanlectures.caltech.edu/I_35.html

Mr Bill, of course I was speaking with convenience, using only numbers to 'make my case'. Yes, what the human eye actually perceives is what really counts. I stand (potentially) corrected thus. But this is an interesting topic, as it conflicts with pure logic. Our intuitive perception of color is really all that matters here and my 'argument' did not include this criterion. Thanks. - David Lyga

 

[DREW WILEY]

Understanding the color wheel is pretty simple. Buy a book, any number of em. What that color wheel implies in the real world, well, that's where things get fun. I remember attending a trade show back when spectrophotometers were just beginning to move from research machines into practical applications. And they were huge expensive things back then. But because I knew how they worked, I took along a bunch of color samples that I knew would drive the machine batty, along with the salesmen and engineers who were trying to sell its features. Things with fluorescence or iridescense, or that would otherwise handle color differently from different angles of incidence. Things highly prone to metamerism. It was fun, at least for me. Not so fun where we actually bought a machine. By that point they operated on intense krypton flash tubes rather than pre-tuned halogens, that required quite a bit of power build-up first. When one of the capacitors
blew, it knocked out not only every surge control device all along the circuit backwards, but half the buildings worth of office computers.
Them was the days.

 

[Steve Smith]

But because I knew how they worked, I took along a bunch of color samples that I knew would drive the machine batty, along with the salesmen and engineers who were trying to sell its features

I like the way you think!


Steve.

 

[DREW WILEY]

David - a good color theory book will also briefly describe some of the basics of human color vision, including phenomena like successive contrast, simultaneous contrast, the way the cones fatigue (just like taste buds) - all of this, and a whole lot more, factors into the physiology and well as the psychology of color vision, and is valuable to know even when planning how to print a subject or choose a film and paper. The genius of someone like Van Gogh with color was really just a simple handful of tricks that any good housepainter would know. That fact won't turn a house painter into a Van Gogh. But the color mixing theory itself isn't all that difficult. The psychological and physiological effects of doing it in a particular way are, by contrast, as complicated as you want to make them, or as interesting as you wish. I personally find it all quite fascinating. And with this kind of background, it amazes me how, in this day of instant honey and jam atop sugar
cubes saturated courtesy of Fauxtoshop, just how few photographers understand how, the more colorful they try to make things (in a naive
manner), the less effective the color becomes, largely due to the way the eye fatigues, just like taste buds reject too much sweetness all
at once.

 

[snapguy]

rock it

High school teachers don't teach rocket science. I'm sure they say "red, green and blue" because they know this kids more or less know what those are. Cyan? Forget it. And I don't think high school kids spend much time making exacting art works to be printed on million dollar presses.
I bet you think Polar Bears are white. They are not. They are black. The fur has no pigment and many times, depending on the light falling on it, it looks white. But since there is no pigment, the beasts are not white. Their skin is black. Polar Bears are black. And Lou Gehrig did not have Lou Gehrig's Disease.

 

[David Lyga]

Thank you, Drew. This is a subject that is not so prone towards objectivity. And that makes it difficult to pin down with finality.

All who posted here contributed more than I had a right to expect and I do hope all this helps others as well. Oftentimes, when printing, if I see that the photo needs more 'red' or more 'green' or any other color I will be a bit disappointed with adding what I feel to be the appropriate filtration. That is because my perception of what the color really is is not necessarily just what the paper might perceive also. Color is a mighty challenge. I remember back in the seventies Fuji film had an advertisement that stressed that their film was better because Japanese eyes could perceive correct color better that others' eyes. This might well have been true and those engineers just might have had special training in order to accomplish this feat (perhaps analogous to 'perfect pitch'). I do not know for sure but that specialty might have involved unique training; training that Western theory (with its 'objectivity always') does not think to be of sufficient importance. - David Lyga

 

[DREW WILEY]

Color printing carries its own set of variables, in addition to all the above. For example, fine tuning color with my subtractive colorhead is a very
different ballgame than doing it with the additive heads, and certain hues simply will not respond to subtractive. But what looks good at the end of the process is very subjective. A better method or hypothetically more accurate film might not produce the "best" results, because so much of this is personal and subject. Most of the time I am aiming for very clean hues that simply weren't possible with color neg films and papers in the past, but now seem to be if enough hoops are jumped thru. But there are times I absolutely the look of faded old negs and chromes of yesteryear, replete with their flaws. There was some "flavor" there that just can't be duly replicated with some "skanky old film app" in PS. Either you hate these kinds of darkroom challenges or love them. I love em.

 

[Photo Engineer]

Gee, is everyone ignoring the Bartleson Breneman effect?

Size does matter!