PDA

View Full Version : 'Multiplicative' mixing and the 4 primary colors


Twinbee
02-17-2004, 02:27 PM
Hi all,

I'm more of a computer graphic artist than a traditional paint/ink artist but I love learning about color and color mixing. This is my first post to the forum, and I have a few questions and insights that some of you might find interesting.

I've looked through many of the forum posts, and there has been a lot of intelligent discussion on the subject. For example, WFMartin's posts at the following thread contain so many truths about color mixing, and is well worth a read if you haven't seen it already:
http://www.wetcanvas.com/forums/showthread.php?t=106936

Anyway, to the meat of my post:
As many of you probably know, mixing inks and paints is based around 'subtractive' mixing. This is certainly true for many inks (such as those used in professional printing etc.). But for more viscous inks and paints, I'm more inclined to think it's a cross between additive, /and/ subtractive mixing. This is one of the reasons why it's often tricky to predict the outcome of mixing two paints together.

To prove my point, if it was truly subtractive mixing, then mixing together red and yellow paint would result in red paint. It doesn't of course, you get orange. Therefore, there has to be a degree of additive (or rather, 'average') mixing in the mathematical process. Another example is mixing blue and yellow paint. Theoretically, this should be black, but you get grey instead. Yes, the blue and yellow probably aren't perfectly pure, but even if they were, the 'additive' mechanism would make a grey shade anyway.

It would also explain why you can't mix together ink primaries such as magenta and yellow, and get a wonderfully pure saturated red (or mix cyan and magenta and get a super-rich blue). Yes, the magenta and yellow might not be perfectly pure to start with, but the main reason that you don't get an incredibly rich red or blue is because mixing paints is an additive process as well as a subtractive one. Actually, there are probably paints that come very close to true subtractive mixing if you mix them carefully, but most of the time, they won't quite match the chroma of the deepest red, or blue straight 'out of the tube'.

Now my thoughts on the terminology 'subtractive' mixing. This is a misnomer really, as it should be called 'multiplicative' mixing, not subtractive. If the mechanism was really 'subtractive', you would often end up with negative numbers!

Here's an example with yellow and cyan (1 is full brightness, 0 is none):
Yellow = Red:1 Green:1 Blue:0
Cyan = Red:0 Green:1 Blue:1

Therefore: 1*0=0 (red), 1*1=1 (green), 0*1=0 (blue)
So: 0, 1, 0 (RGB) = Green

Plus it's easier to work out the answer too, helping people to understand what's really going on. A neat way of looking at it is in terms of percentage. Example:

r:100% g:50% b:0% (orange) multiplied by:
r:50% g:50% b:50% (grey) equals:
r:50% g:25% b:0% (dark brown).


Another thing I discovered is that there are really 4 ink primaries not 3. What's the other you may ask? Well it's white! White is needed as an 'ink' to obtain the others. The white of course comes in the form of the paper itself! :) Another way of looking at it is; pick any two of the three light primaries (red, green and blue), and you can obtain three of the four light secondaries (magenta, yellow and cyan). but if you mix all three primaries together, you get the fourth secondary - white.


One point from an old thread I'd like to comment on:

Of course when you mix a neutral (white) with any color it neutralizes that original color, because it is reflecting the red, green, AND blue thirds of the spectrum. However, when I began mixing Flake White with Cadmium Orange, an even more interesting phenomenon occurred. As I added progressively more white to the orange, it began plotting further and further toward red!


Ah, now this sounds more like a complex and unpredicatable chemical reaction in the inks. Either that, or the 'Flake White' that you used is veering towards orange or yellow, and there's a great deal of 'multiplicative' mixing going on. Do you know which is more likely to be the case?

Finally, I think discussion on the theory of colour would most benefit if everyone agreed on the colour of cyan (and stopped calling it blue), same goes for blue (stopped calling it violet), or calling magenta..... red. :)

JamieWG
02-17-2004, 03:04 PM
Welcome to Wetcanvas, Twinbee. It certainly sounds like you've made it to the right forum! I'm glad you've been checking out some of Bill Martin's posts. He's surely one of our golden assets here in the Color Theory forum.

Jamie

Patrick1
02-18-2004, 08:12 AM
Twinbee, that's an interesting post. But a caveat: I don't know if one can simply multiply the reflectances of the starting paints (at each wavelength) to determine the resultant reflectance of the mix. The Handprint site used to say that you do it that way, but a few months ago it changed to saying you use the square root of the product of the two (geometric mean). And remember...this will just give an estimate...it doesn't take real-world factors into consideration like paint scattering, surface glossiness, properties of the binder etc.

Micheal Skalka, or anyone else...I'd be interested if you multiply the two reflectances, or use the geometric mean.

Michael24
02-18-2004, 09:36 AM
Hi all,

Now my thoughts on the terminology 'subtractive' mixing. This is a misnomer really, as it should be called 'multiplicative' mixing, not subtractive. If the mechanism was really 'subtractive', you would often end up with negative numbers!

Here's an example with yellow and cyan (1 is full brightness, 0 is none):
Yellow = Red:1 Green:1 Blue:0
Cyan = Red:0 Green:1 Blue:1


Ah, now this sounds more like a complex and unpredicatable chemical reaction in the inks. Either that, or the 'Flake White' that you used is veering towards orange or yellow, and there's a great deal of 'multiplicative' mixing going on. Do you know which is more likely to be the case?

Finally, I think discussion on the theory of colour would most benefit if everyone agreed on the colour of cyan (and stopped calling it blue), same goes for blue (stopped calling it violet), or calling magenta..... red. :)

Combining additive and subtractive color theories results in confusion and misinformation. Blue and Yellow = Black has been discussed ad nausea. However as Mark Sabatella has pointed out so masterfully, this is a theoretical construct. No MONOSPECTRAL pigments exist in nature, so like it or not, in the subtractive world, Blue and Yellow will continue to hum along and make Green.

I agree that color printing using CMYK is a combined additive and subtractive process. The small size of the dots, when placed side by side, create an additive optical mixing effect while when overprinted create a subtractive effect.

Before you wish to do away with the term subtractive, take a look at a spectral response curve. All colorants reflect some part of the spectrum between violet (380 nanometers) and red. (740 nanometers) The combinations do not result in zero or negative values. The low values are just subdued. You can never have situations of *full brightness* or *zero*. Not even white has 100 percent reflectance, nor black displays 0 percent reflectance. That situation just does not exist in the real world. Some parts of the spectrum overpower or dominate others thus subtracting a part of the spectrum allowing other parts to reflect to a greater degree. Multiplicative mixing is an acceptable term when describing mixed colorants, but the physics demonstrates that all the colors are present but are subtracted out by the dominant ones. Therefore, the term subtractive in the color science term is more accurate and used by the color science community.

If Bill Martin was using inks instead of cadmium sulphoselenide then the company that sold him the tube of paint was cheating him.

Bill – if you have a chance to read this, I never responded to your observation on the red shift of cad orange plus white. White pigments have an absorption in the violet portion of the spectra. This may cause the cad orange to loose some of its green reflectance and shift more toward red. I will have to replicate this and measure the spectral curve.

A really good text on color theory is Billmeyer and Saltzman's Principles of Color Technology, 3rd Edition, edited by Roy Berns. Lots of math and graphs, but the science is first rate.

Hope this advances you understanding of the subject.

Michael Skalka
Conservation, National Gallery of Art, Wash. DC

Michael24
02-18-2004, 10:06 AM
Twinbee, that's an interesting post. But a caveat: I don't know if one can simply multiply the reflectances of the starting paints (at each wavelength) to determine the resultant reflectance of the mix. The Handprint site used to say that you do it that way, but a few months ago it changed to saying you use the square root of the product of the two (geometric mean). And remember...this will just give an estimate...it doesn't take real-world factors into consideration like paint scattering, surface glossiness, properties of the binder etc.

Micheal Skalka, or anyone else...I'd be interested if you multiply the two reflectances, or use the geometric mean.

Patrick:

Complex color mixing, accounting for absorption and scatter is a complicated mathematical formula. It is all based on Kubelka-Munk. Sorry, the geometric mean is not a quick and dirty little short cut to lead one to the answer to complex color mixing. Its a big math nightmare.

Michael Skalka
Conservation, National Gallery of Art. Wash. DC

Marc Sabatella
02-18-2004, 01:21 PM
To prove my point, if it was truly subtractive mixing, then mixing together red and yellow paint would result in red paint. It doesn't of course, you get orange. Therefore, there has to be a degree of additive (or rather, 'average') mixing in the mathematical process.


I don't understand why you'd say that. Assuming one understands how substractice mixing works, and know the actual reflectance curves of the red and yellow pigments you are mixing, there is every reason to expect an orange to result.

To whit: the red pigment is going to have high reflectance in red, low elsewhere. The yellow will have high in red, orange, and yellow, maybe some green, low elsewhere. If we use geometric mean as our model for how substractive mixing works (although Michael says, and I believe him, that it isn't that simple), a substractive mix should yield high reflectance in red, "medium" reflectance in orange and yellow, perhaps a bit in green, and low elsewhere. I don't see any reason to assume this wouldn't look exactly like it does - an orange that is not quite as intense as, say, cadmium orange.


Another example is mixing blue and yellow paint. Theoretically, this should be black, but you get grey instead.


Again, substractive mixing predicts what happens quite well. Blue pigment will tend to have high reflectance in blue, some in green, low elsewhere. Yellow will have high in red, orange, and yellow, maybe some in green, as described above. The geometric mean would predict some reflectance accross the board, but strongest in blue, green, and yellow - and there is no reason to think we wouldn't see this as green.


Now my thoughts on the terminology 'subtractive' mixing. This is a misnomer really, as it should be called 'multiplicative' mixing, not subtractive. If the mechanism was really 'subtractive', you would often end up with negative numbers!


Only if you imagin that "subtracting" in this ocntext is intended to always mean one reflectance minus the other at each wavelength. It doesn't - it just means that the reflectance of a mixture is less than that of the higher component - something has been "subtracted" from it. But it is true that as an approximation, since multiplication correlates with geometric mean, that it too could be an occassionally useful predictor.

Richard Saylor
02-18-2004, 04:39 PM
Twinbee, that's an interesting post. But a caveat: I don't know if one can simply multiply the reflectances of the starting paints (at each wavelength) to determine the resultant reflectance of the mix. The Handprint site used to say that you do it that way, but a few months ago it changed to saying you use the square root of the product of the two (geometric mean).
Suppose I have a pigment which reflects 50% green. I take two globs of this pigment and mix them together. Obviously I still have 50% green reflectance. However, if the reflectances multiplied, this 'mixture' would have a green reflectance of 25%, which is absurd.

Multiplication of reflectances for subtractive mixing is clearly incorrect (even absurd). The geometric mean at least makes sense.

WFMartin
02-19-2004, 02:17 AM
Twinbee,

Gosh, thank you for the nice plug! Welcome to wet canvas. We do have our fun discussing color theory in this forum, don't we? :) Three or four of us really get going sometimes in this understanding of color.

Let's address a couple of points, if I may be so bold. One reason that cadmium orange (out of a tube) becomes redder when mixed with white is that it is a simple characteristic of that particular pigment, and I have yet to understand the science behind it. It a characteristic of a pigmented paint, called an "overtone". On the other hand, I understand if you were to do a drawdown of a thin coat of the same paint, thus allowing the white of the canvas to show through, you would have the characteristic called an "undertone". I have yet to do a study of this effect, though; however, I have been led to believe that perhaps it might result in a different color bias.

Now, if you were to mix some cad orange with white, you'll see this phenomenon in that the tint takes on a reddish cast. I proved this to myself, by analysing the effect with a color densitometer, but it is so pronounced an effect that it is also quite visible to your eye.

Then, as an experiment (I did it just the other day), mix the same cad orange mass color, using W/N Permanent Rose 502 with W/N Transparent Yellow 653.
It'll take some white in the mix to accomplish this, but I guarantee, on the palette, you won't see the difference. However, take this mixed "cad orange" that you made, and add additional white to it, just as you did to the tube cad orange and, whammo!, you've got a TOTALLY different tint color than the tube cad orange produces. The overtone of this mix is decidedly YELLOW. And I mean a bunch!

Well, looks as though this might prove to be an interesting thread.

Bill :D

Michael24
02-19-2004, 08:32 AM
Multiplication of reflectances for subtractive mixing is clearly incorrect (even absurd). The geometric mean at least makes sense.


As stated in a posting above, the outcome of a mix of colors uses the Kubelka-Monk mathemetical model to extract the outcome. Its not a simple, quick, down and dirty formula. (See the Principles of Color - Billmeyer and Saltzman, 3rd Edition, ed. Roy Berns)

Michael Skalka
National Gallery of Art, Wash. DC

Michael24
02-19-2004, 08:58 AM
Dear Color Theorists:

Color Modeling is a multi-million dollar science that is in practical use every day. Do you think that any manufacturer would take thousands of pounds of cadmium yellow pigment at $50.00 per pound and mix it with another pigment at nearly the same price to make a plastic cup and not know without a doubt the color what the final product's color will be?

The complex math models for ploting subtractive color have been invented and are in daily use by industry. Subtractive color modeling is not a theory with undiscovered facets. It is a living, breathing working model that is the mainstay of color science engineers.

For all who wish to venture into redefining color theory:

Stop looking foolish and beating your head against the wall. Go out and read some books. Principles of Color by Billmeyer and Saltzman. Color Science in the Examination of Museum Objects, Tools for Conservators by Ruth Johnston-Feller. These books contain hundreds of references to other publications. It lays all this stuff out. It does not have to be reinvented in this forum. North American has already been discovered. You don't have to do it again! Yes, the math is daunting. Its all very technical. However, if you can get through some of it you will find that most of what is discussed as homegrown color theory nonsense in this forum has been written about scientifically and accurately in one of many text books.

For all who wish to discuss the application of color theory:

This is a wonderful place to marry scientific knowledge with practical artists knowledge. The interaction of colors, color bias, the strengths and shortcomings of pigments are wonderful points for discussion. They provide us with knowledge and a common ground for sharing viewpoints. Throw in some well researched theory and apply it to paints and that becomes the makings of fruitful discussion.

For all who are not concerned about the state of pigments and colors:

The world of pigments is changing before our eyes. New organic colors are coming into the marketplace. Our stable of old inorganic pigments will most likey become extinct within the next generation. (No you don't have to start hording tubes of cadmium paints quite yet!) Legislation like Prop. 65 in California has had a tremendous impact on the art materials business. Natural sources of earth pigments are drying up. Many are gone. Even some of the organic pigments out of fashion have been discontinued. Remember that artists paint companies are bottom feeders. They do not control market production of companies like BASF, Ciba, or DuPont. Paint makers buy small lots of pigment by comparison to General Motors. We as artists do not control the industry.

Within a short period of time we have lost lots of great colors: Manganese Blue, Chrome Yellow, Naples Yellow Genuine, are just a few. Some are gone for good reason, others because they are too costly to produce.

Thanks for taking the time to read and share. I have really enjoyed learning and discussing issues with most of you.

Michael Skalka, National Gallery of Art, Washington, DC

Richard Saylor
02-19-2004, 10:59 AM
As stated in a posting above, the outcome of a mix of colors uses the Kubelka-Monk mathemetical model to extract the outcome. Its not a simple, quick, down and dirty formula. (See the Principles of Color - Billmeyer and Saltzman, 3rd Edition, ed. Roy Berns)

Michael Skalka
National Gallery of Art, Wash. DC
Dear Michael,

I have the mathematical background (Ph.D., Rice Univ.) to understand the formulas in the Kubelka-Munk model, but I have no way to determine the values of the parameters used in the formulas. I guess this means I should shut up.

Michael24
02-19-2004, 01:30 PM
Dear Michael,

I have the mathematical background (Ph.D., Rice Univ.) to understand the formulas in the Kubelka-Munk model, but I have no way to determine the values of the parameters used in the formulas. I guess this means I should shut up.

NO! Not at all!! I am impressed. I do not claim to understand how to use the model. I am a humanities person not a mathemetician. Kubelka-Monk looks at each wavelength and takes into account scatter and specular reflection to derive the addition/subtraction of the two wavelengths. When I get a chance I will ask Roy Berns at RIT to explain it in simple terms to me. I have the formula within a spreadsheet model that does a fair job of color mixing. Roy created it for a project he was working on here and gave a copy to me. It works fairly well but does not account for luminance well. The math is one big hairy thing with cell references to fixed and variable numbers.

The most I do is take spectraphotometer readings, several times for each sample, average them, plot them in Excel, run them through another spreadsheet that determines their K/S ratios so that spectra may be compared to each other accounting for their luminance. I am still a low level student of the math involved. I will try to look for a decent explanation of how and where to plug in the numbers.

Later,

Michael Skalka, National Gallery of Art

Marc Sabatella
02-19-2004, 03:05 PM
As it happens, my degree is mathematics too, although only a BS (my Masters is in Computer Science, if that counts for anything :-). But I am not necessarily interested enough to dig that far into my past - I haven't used any higher math in 15 years - to really plow through a complex set of equations, even though I might be capable of it. Obviously, just performing the mixtures with actual pigment and seeing for myself what works is what it really comes down to. The delving into theory is more to give an idea of what things might be worth trying - to help form a hypothesis I can then test empirically, and if the hypothesis pans out, then it saves me the whole trial and error process of randomly picking pigments hoping for the results I want.

To that end, I don't need to know the exact reflectance curves that the formulas would predict for a given mixture. What *woud* be interesting to know is how well it actually correlates with geometric mean, or any other simpler way of estimating reflectance. And by "correlation" here, I don't actually mean in a statistical sense, but more in the sense of whether it has some of the same basic mathematical properties.

For instance, I could characterize geometric mean in the following way: given the reflectance percentages for two different pigments at a given wavelength, the geometric mean will be somewhere between them, and will in fact always be closer to smaller of the two. That is, low reflectance in one of the pigments is hard to overcome with high reflectance in another. If the Kubelka-Monk model produces results that fit this same basic description, then I am satisfied that thinking in those basic terms - not actually computing geomtric mean, but just saying "somewhere between the two but closer to the smaller" - will at enable to me to come up with hypotheses about pigment mixtures that I can then test, with a better success rate than merely guessing. Meaning, ultimately, fewer experiments necessary.

BTW, for anyone shaking their heads in wonder at this point, I don't really mean I'd be going through this process for every color I am trying to mix in a given painting. I mean more in the sense of, if/when I start thinking about wanting changes to my palette, thinking perhaps I need to add a pigment to help me more easily mix a given type of color consistently, I would like to have a really good idea before actually going out and buying a tube of paint that the pigment I am buying will lead to the results I want.

For instance, when I was using a palette consisting of ultramarine, phthalo green, quinacridone rose, and azo yellow, I decided I needed help making a range of oranges more easily. I figured a pigment somewhere between the azo and quinacridone would be the key, but as far as choosing between cadmium yellow deep, cadmium orange, cadmium red light, naphthol, or pyrrol, or perhaps other choices, it was pretty much a crapshoot. Being able to make some hypotheses based on predictions extrapolated from reflectance diagrams might have helped in the decision making process. As it happens, I'm happy enough with the chocie I made - naphthol - not to be sweating it right now, but who knows how much mappier I'd have been with another choice? Sure, I could pick up a tube of each and find out for myself, but again, I'm looking to the theory to help me guess right the first time and save myself some time and money.

LarrySeiler
02-20-2004, 10:01 AM
Stop looking foolish and beating your head against the wall. Go out and read some books.

Your end signature credentials Michael certainly are impressive, and in responding potentially intimidating (so I'll ask for you to humor me here). It is of course a most welcome and enrichening thing for our community to find professionals in the business and field of art here. It adds to our credibility of a site...and I'll let you know up front..I am using your post to make a point. A springboard, if-you-will....

Let me now state a thing that will disgrace my credibility among this fine group here by no doubt "looking foolish"....(doing so for the sake of lurkers)

I think there seems to be some gap, some great chasm or gulf between those as color theorists...and those of perhaps lesser intellectual capacity wiring that prefer to paint minus the affections for convergent analysis.

For those that teach art or aware of learning models, and might be aware of left brain right brain thinking, we know the categorizing or stereotyping of the typical creative person as existing in the right brain world.

The feeling I get, and I don't intend to demean...but the "talk" of color as a science, seems IMHO to be more a left brain activity or concern. Really, there is room for such talk for there are all kinds of people that learn in all different manners. In a discussion that tends to grow intellectual, often egos get involved that soon attempt to "best" one another. It easy then to take on an appearance (even if that is not the intent) to be elitist as those more informed, and somehow superior as individuals from those of others less informed.

From simply an artist's perspective, the cost of cadmium as to how it concerns a business reasonably would concern left brained folks sitting at their calculators running this business and well it should, and thankfully so. Typically...(I'm speaking of your average artist), it is good we have a spouse that is more practically natured, who keeps lead in the soles of our shoes because our propensity for chasing after whim and dream is great. Those living in the right brained world do well to have a spouse that balances them (as well as the check book!)

That being said...how foolish I suppose (let's digress) for me all these years to have worked from a simple premise of the RYB colorwheel. How foolish still that I teach K-12 students this basic errant model...but with a twist of warm and cool variations of the primaries. I teach such for what results come of the art work...and not for its scientific potential to expand understanding. After all in my foolish less educated mind, I see the model as only that...a model to be toyed with, experimented until a suitable working order is come upon.

Goofier still is that my near 30 years of paintings have fooled so many to be good, no....even excellent! Foolish yet even more I suppose is how I have allowed the accolade that has followed my career to reinforce the sense of an imagined safety net to go on ignoring whatever benefits scientific thinking of color might bring me.

Oh....I'm not saying others do not have this right, nor that no benefit comes of it. I'm foolishly admitting that the emphasis of its importance is missing to me. I'm not indicting anyone here....I'm putting myself up on a cross to be nailed to it if you are so inclined.

I have worked from a premise of a thing looking right as a thing from an artist's aesthetic sensibility might think a right thing to be. One red looks this way...another looks that way. I have entrusted appearance to be reinforcement enough.

This red appears to be cool....and I anticipate it to contribute in one way. Another appears to be warm, and I anticipate it to do another thing from the cool.

I get out there in the field with an elusive sun fully intent to focus on my subject, fully expecting to get into a "zone" where all interference of thought other than my subject and drama at hand is quieted. The threat of the sun's drama at any moment vacating leaves little time for intellectual digression.

I fall into that color theory I suppose of "optical color"...where I see it, and I paint it. It looks this way...I paint it this way. That looks that way, I paint it that way. This leads me as an artist, not as a thinker to believe, and I suppose foolishly so...that I can trust my eyes to see a thing done.

I developed a sensitivity thanks to simple warm and cool thinking...to see cools where I anticipate them....warms where I expect them to be as well.

Why bother to write then to make comment? Well, it might well be only to represent other artists such as myself that have allowed time and experience to work a thing out...but did so, I suppose not having read all these fine scientific treatises on the subject.

For those lurking...that might not have read these books as well and suddenly feel that the indictment of "foolishness" might defer upon them I want to offer them hope. (Michael, it is not having used the word "foolishness" on your part that brings the sense of this to heart for many but the way the sense of emphatic insistence comes in these discussions that ends up infering such).

Speaking now to you unfortunates like myself, one might accidentally fall upon a working system, having contented yourself to learn thru the creative process. Do not be downcast in spirit. Let not your hearts be troubled. Foolish we perhaps might be, yet continue on in your energy. Learning comes thru thinking that finds credence in the doing...but learning also comes from the doing no doubt of things badly, that in doing still more slowly weeds the bad out. Intuition...some not quite so eloquent understanding comes bit by bit. Our limited thinking perhaps comes more in smaller bytes that reflects on one work finished comparitively with maybe a few other works.

I am by no way against the learning. I frown upon the assessment, the preclusions, the judgments that infer foolishness defers to those that do not engage understanding thru convergent ordered thinking. For you...I simply say, learn as you do....from painting to painting, work to work. From critique to the next. Some of us are wired to take in much information convergently, others divergently or by trippin' up or happy accident. Paint on....paint on... and like myself, foolish we shall be.


btw Michael...my comment is directed toward the use of the word "foolish" and in light of how technical discussions in color theory can get...I used your post to share thoughts intended for all. I do not for a moment think you actually feel folks such as myself foolish. I just pro-offer thoughts to remind all to leave room for all. Nothing wrong with enjoying indepth discussion...but in so doing we might forget the beauty of that which is yet simple.

As a side note to all....Its possible is it not...that someone might paint beautifully with great use of color and have no where near the knowledge of color theorists. Its also possible such an artist might paint better in his ignorance than those that are astute in color theory. For that possibility, we need obligate ourselves some room for others that don't quite get what the fuss is all about.

btw....I have edited this post to more adequately represent my intent, and apologize for initial knee jerk reaction. I come into this conversation later...

peace

Larry

Richard Saylor
02-20-2004, 11:52 AM
The feeling I get, and I don't intend to demean...but this "talk" of color as a science, seems IMHO to be more a left brain activity or concern. Really, there is room for such talk for there are all kinds of people that learn in all different manners. Its when a suggestion of "foolishness" arises though that pits one group as elitist, more informed, and somehow superior individuals from those of others that some humanitarian defense should come to aide.
Michael is just chiding some of us who appear to be trying to reinvent the wheel. Don't worry about it, Larry.

LarrySeiler
02-20-2004, 01:51 PM
Michael is just chiding some of us who appear to be trying to reinvent the wheel. Don't worry about it, Larry.

thanks....hoping Michael realizes I only springboarded off his post. I know Michael is a great guy! Appreciate his input and time here....
And..seriously our community benefits from folks with his proven background, Bill's printing industry background and so forth. WE have so many fine people here.

Still...a question though to ponder. What about artists that paint well without earning their dues in the color theory department, have they just been lucky and likely at any moment to start putting out mediocre laments not knowing why?

I and another member were discussing this very same thing but in the context of music. For example...I play blues and have had accomplished guitarists I've played with over the years with backgrounds in music degrees, music theories and accolade want me to spend time helping them understand how to play blues. Me? I can only read music if I recite acronyms "Every Good Boy Does Fine" or so forth....

Of course those that spend a lifetime dedicating themselves to the knowledge of music are going to defend its necessity or benefit to read music, and those that have played their lifetime not developing the ability to read music will resent the implication they are less capable because they lack the knowledge. Still...what do we do about those dadblasted musicians and artists that do not have the level of intimate knowledge we'd like them to have?

I remember when I taught tennis lessons both private and for the city of Green Bay...I played varsity high school and college level. "Get your racquet back early" "anticipate" "turn your body and step toward the net, hit the ball out front and follow thru" and so forth. Then along came that darned Bjorn "I don't turn my body" Borg. The strong western grip heavy topspin hitter that helped revolutionize grips and game play style. He was an instructor's nightmare.

Tell a student they need to get their racquet back early, not make their backswing part of their foreswing, turn their body and step into the ball, and they'd say, "well, what about Bjorn Borg?"

The size of Bjorn's thighs were massive, and his upper body strength incredible. Bjorn was able to break the rules/tradition and do what to that time ought not to be done.

take care

Larry

LarrySeiler
02-20-2004, 02:33 PM
FWIW...I also know more about color theory than my posts are letting on. Bill well knows that I'm sure.

I have chosen my warm & cool palettes for what I do, and I don't think out of ignorance; yet I will grant that I recognize any number might believe me to be ignorant for so choosing such a palette...hee heee...

I'm just trying to stimulate some thought, as well as some regard for folks that seem to produce a body of work that is exemplary inspite of lacking indepth color knowledge.

For everyone reading along...that we have a color theory forum is evidence enough for its importance...but humor me for a moment. I enjoy the rants and the read here in this forum, and often I lurk myself without participating. The task of trying to be understood is not always an easy one...and in the frustration of trying to make one's meaning clearer there are moments where things said, the bravado and so forth come across I'm sure with intent that is not meant.

In arguing for better understanding sometimes the benefit to be gained might project or infers one would become thereby a better artist due to having such knowledge. I lament that I wish I understood intimately the reading of music. Yet recently I participated in a pit band for a live play. I couldn't believe how inflexible people were when the director wanted minor changes because something wasn't written in the music. At the same time came surprises among the band that I did not need music at all to play along (nor would it have helped me), and I felt some resentments from a few members for that reason. During the four nights of the presentation, I simply enjoyed playing and taking in the moment, the play and so forth. Seemed like everyone else was sweating jot and tittle that everything would come together.

Now....from my position, an insistence to learn to read music is like suggesting I ought to be like these people, yet I somewhat recoil at the prospects of being creatively inflexible...a slave to the page at hand.

Thus...by sharing this music metaphor, and by having a life long body of work that has been a career for me in state and regional competition and galleries I hope you can appreciate that embracing the task of what seems greater understanding of color than what has been to me a working knowledge comes as a frightening prospect. You know that ole saying, "if its not broken don't fix it!"

Can some appreciate that...and like the music instructor, how would you go about attempting to convince one of the benefits? I ask this, because I always anticipate the lurkers and nonparticipants that are out there reading along. We have over 36,000 members and always there is only about 400 members participating at any one time it seems, so indeed...there must at any one time be a large contigency of folks reading along.

Thanks for your patience.....!!!

btw...I am impressed with the level of knowledge of this forum's participants, good people all....
I feel like the guitarist throwing this question out on a forum where music directors and masters are talking...
...but you know, us guitarist types are out there and some even think we're pretty good.

peace,

Larry

Richard Saylor
02-20-2004, 04:48 PM
Larry,

Nobody has to read music in order to play music or know color theory in order to use color. One would have to be terribly ignorant (or nuts) to dispute that. Personally I just think it's fun to play with color theory. (It's probably the mathematician in me.) In my studio (i.e., my bedroom :) ), I have a blast messing with cyan, magenta, and yellow. (I'll bet I can pretty closely match any hue you can mix with your six colors, but of course I will not necessarily be able to match the intensity or chroma which you can get.) The only concrete advantages of three colors for me are the ease of obtaining color harmony and saving a little money. Of course if I'm deeply immersed in a painting, I'm liable to reach for any color which does the job. Also, if I'm doing plein aire, I don't have the speed or the patience to restrict myself to three colors. I really admire Jaimie for having the fortitude to experiment with that.

LarrySeiler
02-20-2004, 05:59 PM
thanks for your response and that it is simply fun for you to discuss.

I should apologize to all I suppose for changing the tone of the conversation and what I will do is bow out and start a thread of my own on it.

To be honest...I find sometimes such conversations about color theory a bit intimidating. Not sure why because I fail to sense error or poor direction in my own work, and have very strong feelings/convictions about how and why I paint. Yet...I sense in the conversations that someone having such acquired higher knowledge about color should translate to better work. I really need someone to explain how my work might improve I suppose for me to see what the tuss is all about. So, in that sense...I guess I admire all of you quite a bit in this forum!

I am probably missing something...and am the poorer painter because of it.

peace

Larry

Marc Sabatella
02-21-2004, 12:55 PM
The only concrete advantages of three colors for me are the ease of obtaining color harmony and saving a little money.


In my opinion, though, those are both significant advantages of limited palettes, whether CMY or a dual RYB primary system or any other. You could also add convenience for plein aire painting, however, as has been observed, the CMY palette in particular ends up being a bit counterproductive in that department, as any hassle you save in carrying extra tubes around is probably lost in struggling to get the right dull greens & oranges for landscape work.

But this all gets to the heart of why this kind of discussion *does* have practical value. Assuming you buy into the notion of a limited palette - to improve color harmony, save money, or make plein aire painting more convenient - the question becomes, *which* colors are going to make sense. And one of the important end results of these discussions for me is the realization that, in practice, there are no definitively right answers. There are many sets of palette colors that can produce a wide range of colors. If you want to understand which set of palette colors is right *for you*, given your own personal style and preferences, understanding color theory can help you arrive at a good palette perhaps more quickly or with a bit more certainty than the trial and error process. But of course, either way works, and once you've settled on a palette that works, there probably isn't much if any added value in getting any deeper into this world unless you enjoy it for its own sake.

Which, I think, address Larry's complaint - no, there is no reason to infer that a better understanding of color theory would improve Larry's paintings, either by getting him to alter his palette, or make better use of the one he has. On the contrary, his understanding of color theory (whether the scientific stuff he admits to knowing or just the practical aspects of it) is what led to the palette that works for him, and there is no reason to suppose that any change would be for the better. But for someone who *hasn't* settled on a working palette, thinking about color theory issues now wouldn't be a bad idea, as it could save a lot of failed experiments down the road.

For example, one *potential* issue with a dual RYB primary palette with regards to color harmony is the simple fact that, assuming you actually used all of those colors in a given painting, there would be two entirely different reds, two entirely different blues, and two entirely different yellows making up your mixtures. When I've used this type of palette, I've been happier with the results I obtained if I laid all six colors out but chose just one of each primary for a given painting to the extent I could get away with it. For instance, I might want both a greener blue and a more violet blue in a painting, but if I can create both from, say, ultramarine (mixed with a dash of yellow to give me the greener blue), there is a harmony that isn't automatically present if I use both ultramarine and phthalo. Of course, the converse problem with an even more limited palette can be a tendency toward too much uniformity from painting to painting, as well as a potential lack of the excitement that comes from an occasional discordant color note in a painting. I think both problems can indeed be overcome, but one of the choices we make in settling on a palette is which problem we think we can deal with better (or, which would bother us more if we failed to deal with it).

DuhVinci
02-21-2004, 06:40 PM
As one of the lurkers, I must say that I enjoy eavesdropping on the discussion even though I understand less than half of it. I'd like to make two points:

1. The greatest athletes probably know nothing of the physics or mathematics involved in their acomplishments. In fact its well known that if an athlete thinks too much about how he or she accomplishes something it can hurt. Even Tiger Woods, when he works on his golf swing, is working on his body mechanics, not on theories of mass, force, airodynamics, etc. So in trying a new swing, he's like an artist trying a new pallet, not one studying the physics of color.

2. Maybe the physics behind color theory can be used to explain why an artist's paintings work so well but it should not be asked how an artist could paint so well without an understanding of the physics behind color theory.
Of course the artist should have a practical model of color theory which will differ from the models used by other, equally effective, artists. Just as Tiger's golf swing differs from other professional golfers. The physics remains the same but the practical models will, justifiably, differ from user to user.

Twinbee
02-21-2004, 11:58 PM
Hi all,

Thanks all for the interesting responses!...



Combining additive and subtractive color theories results in confusion and misinformation. Blue and Yellow = Black has been discussed ad nausea. However as Mark Sabatella has pointed out so masterfully, this is a theoretical construct. No MONOSPECTRAL pigments exist in nature, so like it or not, in the subtractive world, Blue and Yellow will continue to hum along and make Green.

Don't you mean grey? It would only make a (muddy) green if the blue was not real blue, but had a tinge of green/cyan in. Either that, or there's some unpredictable chemical reaction in the inks that causes the slight green hue.

Because blue is at the end of the visible spectrum, I would imagine that it's fairly easy for paint manufacturers to avoid green pollution in blue paint, but I may be wrong.



I agree that color printing using CMYK is a combined additive and subtractive process. The small size of the dots, when placed side by side, create an additive optical mixing effect while when overprinted create a subtractive effect.

Right, yep that makes sense. For those who want to experiment with this type of dot mixing, I found a good site at:
http://www.learner.org/teacherslab/science/light/color/dots/index.html



Before you wish to do away with the term subtractive, take a look at a spectral response curve. All colorants reflect some part of the spectrum between violet (380 nanometers) and red. (740 nanometers) The combinations do not result in zero or negative values. The low values are just subdued. You can never have situations of *full brightness* or *zero*. Not even white has 100 percent reflectance, nor black displays 0 percent reflectance. That situation just does not exist in the real world. Some parts of the spectrum overpower or dominate others thus subtracting a part of the spectrum allowing other parts to reflect to a greater degree. Multiplicative mixing is an acceptable term when describing mixed colorants, but the physics demonstrates that all the colors are present but are subtracted out by the dominant ones. Therefore, the term subtractive in the color science term is more accurate and used by the color science community.
Could you give me an example with the orange and grey I posted? For me, 'subtractive' makes no sense because the resulting color changes according to which order you do the calculation. For example, red minus yellow (black if you count negative numbers as zero), is not the same as yellow minus red (green). Surely 'multiplicative' is a more appropriate term instead?



A really good text on color theory is Billmeyer and Saltzman's Principles of Color Technology, 3rd Edition, edited by Roy Berns. Lots of math and graphs, but the science is first rate.

Sounds good. Hmmm... what I would like is a diagram showing the response for each eye cone at a given wavelength. I've found six so far on the net, and they all contradict each other slightly :)
http://henrysturman.com/images/spectra.jpg
http://www.siggraph.org/education/materials/HyperGraph/color/images/colorspc.gif
http://www.marine.maine.edu/~eboss/classes/SMS_491_2003/sound/em-spectrum_human-eye_asu_380x300.gif
http://webvision.med.utah.edu/imageswv/spectra.jpeg
http://davis.wpi.edu/~matt/courses/color/p2i7.jpg
http://teachpsych.lemoyne.edu/teachpsych/faces/script/Ch09_HTM/Trad_curves.gif





To prove my point, if it was truly subtractive mixing, then mixing together red and yellow paint would result in red paint. It doesn't of course, you get orange. Therefore, there has to be a degree of additive (or rather, 'average') mixing in the mathematical process.

I don't understand why you'd say that. Assuming one understands how substractice mixing works, and know the actual reflectance curves of the red and yellow pigments you are mixing, there is every reason to expect an orange to result.

Well, in terms of light filters and certain inks (say, felt-tip pens and the professional ink printing process), red and yellow will make red, not orange. Though it was a mistake of mine to say that mixing more opaque paints also relies on this subtractive mechanism. I think it's because most schools and websites mistakingly equate paint mixing with the term 'subtractive', and then proceed to give the classic light filter diagram (http://www.skytopia.com/stuff/cmy.png) as a demonstration of this.

Like you, and others in this thread have said, it would seem mixing paint is a lot more complex than simple addition, 'subtraction' or even finding the geometric mean.



Again, substractive mixing predicts what happens quite well. Blue pigment will tend to have high reflectance in blue, some in green, low elsewhere. Yellow will have high in red, orange, and yellow, maybe some in green, as described above. The geometric mean would predict some reflectance accross the board, but strongest in blue, green, and yellow - and there is no reason to think we wouldn't see this as green.

Yes, I was referring to mixing with a pure blue, containing no green 'pollution'. I'm not sure how close ink manufacturers have got to this ideal, but would be interested to know...



Only if you imagin that "subtracting" in this ocntext is intended to always mean one reflectance minus the other at each wavelength. It doesn't - it just means that the reflectance of a mixture is less than that of the higher component - something has been "subtracted" from it.
I'm realising that the subtractive term is more of a 'general' term, rather than an exact mathematical process. But it would be better if the nomenclature was consistent with the term 'additive' mixing, where the wavelengths are properly added together in a mathematical way. Hehe, I guess it's more or less pointless for me to redefine the term though, now that the definition of subtractive is so widespread :)



Suppose I have a pigment which reflects 50% green. I take two globs of this pigment and mix them together. Obviously I still have 50% green reflectance. However, if the reflectances multiplied, this 'mixture' would have a green reflectance of 25%, which is absurd.

Multiplication of reflectances for subtractive mixing is clearly incorrect (even absurd). The geometric mean at least makes sense.

Fair enough. I admit I didn't think of using the same colour paint to mix with itself. One should only use the 'multiplicative' term for light filters etc.

I'm going to print a list of different types of mixing. Please let me know if I'm mistaken in any of them:

Dyes and inks (especially felt-tips): multiplicative mixing
Light filters: multiplicative mixing
Monitor displays (LCD/CRT): additive mixing
Colour splitting (e.g. chess board style): 'average' mixing
Professional ink printing: multiplicative cross with average mixing
Paints: Kubelka-Monk model, (approximates down to..?)



Let's address a couple of points, if I may be so bold. One reason that cadmium orange (out of a tube) becomes redder when mixed with white is that it is a simple characteristic of that particular pigment, and I have yet to understand the science behind it. It a characteristic of a pigmented paint, called an "overtone".
I wouldn't mind seeing a picture of a graduation from cadmium orange to this pink tone.



As stated in a posting above, the outcome of a mix of colors uses the Kubelka-Monk mathemetical model to extract the outcome. Its not a simple, quick, down and dirty formula. (See the Principles of Color - Billmeyer and Saltzman, 3rd Edition, ed. Roy Berns)

I wonder if this super-complex formula can be /roughly approximated/ down to a relatively simple one. I realise that some paints have massively different attributes, but assume a generic selection with 'idealized' properties.
There's something in maths called the 'power' or 'generalized' mean. It subsumes all other types (harmonic, geometric, arithmetic, root-mean-square etc., and everything in between) - with just a simple change of the exponent value. Perhaps the right value gets fairly close to the number that the Kubelka-Monk model churns out.



To that end, I don't need to know the exact reflectance curves that the formulas would predict for a given mixture. What *woud* be interesting to know is how well it actually correlates with geometric mean, or any other simpler way of estimating reflectance. And by "correlation" here, I don't actually mean in a statistical sense, but more in the sense of whether it has some of the same basic mathematical properties.

Yep, I would be very interested to know too. Is any relatively simple combination of subtraction, multiplication, addition, division (!), and the various types of averages enough to roughly approximate the Kubelka-Monk model...?

Richard Saylor
02-22-2004, 12:38 PM
Fair enough. I admit I didn't think of using the same colour paint to mix with itself. One should only use the 'multiplicative' term for light filters etc.

This is getting interesting. You are making me think, Twinbee. :)

With stacked filters, you are combining spectral transmissions, which indeed combine multiplicatively. With mixed paints, you are combining spectral reflectances. Transmission and reflectance are like apples and oranges.

In your examples, you should specify whether it is transmission or reflectance that is being combined. With inks/dyes, it would seem that you are combining transmissions. The same is true when glazing one transparent color over another, which is a common procedure with watercolors. This probably explains why I can get a higher chroma green by glazing pthalo blue over lemon yellow than I can by mixing the paints. With glazing, I am combining transmissions multiplicatively; with mixing I am combining reflectances (for which the geometric mean is probably a reasonable rough approximation). If it were reflectances which were being combined in both instances, I would expect the opposite: higher chroma with mixing than with (multiplicative) glazing.

Before Michael or somebody else gets on my case, let me add that the above remarks involve some oversimplication. In real life the situation is vastly more complicated. It's like in elementary physics. If you want to show that a projectile describes a parabolic arc, you ignore air drag because this simplifies the equations of motion enough that they can be easily solved by hand.

Patrick1
02-22-2004, 01:32 PM
Because blue is at the end of the visible spectrum, I would imagine that it's fairly easy for paint manufacturers to avoid green pollution in blue paint, but I may be wrong...
Heh heh...you be the judge:

Here's Phthalo Blue GS:
http://www.handprint.com/HP/WCL/IMG/RC/rcPB15.jpg

Prussian Blue:
http://www.handprint.com/HP/WCL/IMG/RC/rcPB27.jpg

Cobalt Blue:
http://www.handprint.com/HP/WCL/IMG/RC/rcPB28.jpg

Ultramarine Blue:
http://www.handprint.com/HP/WCL/IMG/RC/rcPB29.jpg

Cerulean Blue:
http://www.handprint.com/HP/WCL/IMG/RC/rcPB35.jpg

Indanthrone Blue:
http://www.handprint.com/HP/WCL/IMG/RC/rcPB60.jpg

Marc Sabatella
02-22-2004, 05:42 PM
Don't you mean grey? It would only make a (muddy) green if the blue was not real blue, but had a tinge of green/cyan in.


This is true of every blue pigment in the world. In fact, pretty much all pigments reflect all wavelengths to some degree. They generally reflect best at some particularly wavelength, also reflect all the surrounding wavelengths pretty well too. Anything in the real world that reflects enough blue to be seen as blue is going to reflect a pretty significant amount of green as well. Which is why, in fact, pretty much any blue paint, mixed with yellow, *will* produce some sort of green. Not as brilliant as, say, phthalo green, but very clearly and obviously green.

For me, 'subtractive' makes no sense because the resulting color changes according to which order you do the calculation. For example, red minus yellow (black if you count negative numbers as zero), is not the same as yellow minus red (green). Surely 'multiplicative' is a more appropriate term instead?


Not if you take "subtractive" in the sense I described - when two colors are mixed, the amount of light reflected is *less* than the amount of light reflected by the components. Less = subtraction, in colloquial usage if not in any exact sense.


what I would like is a diagram showing the response for each eye cone at a given wavelength. I've found six so far on the net, and they all contradict each other slightly :)


BTW, I don't see the one from the Handprint site on your list - that's the site that gets used as reference most around here, for better or for worse.


Well, in terms of light filters and certain inks (say, felt-tip pens and the professional ink printing process), red and yellow will make red, not orange.


This isn't exactly true. First, the usual printing process doesn't use red - it uses magenta. And orange *is* created by mixing magenta and yellow - how else would it be created?

As for felt tip pens - an essentially transparent medium - I suspect that what you are seeing in your red and yellow mixtures isn't so much a factor of subtraction or multiplication as the fact that darker colors always win. But I confess I haven't given this a lot of thought.


I'm realising that the subtractive term is more of a 'general' term, rather than an exact mathematical process. But it would be better if the nomenclature was consistent with the term 'additive' mixing, where the wavelengths are properly added together in a mathematical way. Hehe, I guess it's more or less pointless for me to redefine the term though, now that the definition of subtractive is so widespread :)


Well, if you must rename it, then, call it Kubelka-Monk mixing :-) The appeal of the term "subtractive" in this context is that it seems an obvious converse of "additive", but of course neither "subtractive" nor "multiplicative" actually describe what happens mathematically particularly well.


Dyes and inks (especially felt-tips): multiplicative mixing
Light filters: multiplicative mixing


I still don't see what you think "multiplicative" mixing is - what you've described in the literal sense *never* happens. It might be interesting to discuss what actually happens with transparent layers of color as in felt tips and light filters, but I don't see any reason to assume anything is being multiplied...

Richard Saylor
02-22-2004, 08:36 PM
For me, 'subtractive' makes no sense because the resulting color changes according to which order you do the calculation. For example, red minus yellow (black if you count negative numbers as zero), is not the same as yellow minus red (green).
It makes perfect sense. Subtraction is not commutative. 5-3=2, but 3-5=-2. Reversing the order of subtraction changes the result into its opposite.

Actually, you made a miscalculation with your ideal colors. It is true that Y-R=G. However, R-Y=M (magenta, not black). Note that magenta and green are complements: M=-G. So, analogous to what happens with numbers, Y-R=G, but R-Y=-G.

Please, please don't anyone tell me that this is nonsense because ideal colors don't exist. It's also true that projectiles are usually subject to air drag, which is ignored in a first approximation when computing their trajectories. (Sometimes I feel like I'm walking on eggshells.)

WFMartin
02-22-2004, 10:44 PM
Hey CMY Guy,

When you're right, you're right, and.......you're right. ;)

By the way, for whomever mentioned it, the laying down of transparent filter combinations on a white substrate represents one of the best examples of subtractive use of color that I can think of. The white substrate (paper, canvas,etc.) becomes the full spectrum, in this case, and each successive filter blocks out various parts (thirds, if the filters represent cmy)as it is added, creating new colors. Where there is to be a lighter color, the halftone dots are simply created smaller, allowing more of the white substrate to show through. It works much better than the mixing of white paint with a pigment, because then the pigment begins to exhibit the effect of overtones, which I mentioned earlier. The lithographers' cmy inks as well as the filter example perform the same function, because they are both transparent, and work by each subtracting one third of red, green, or blue from the spectrum.

The yellow printing inks of the old days used to actually be quite opaque, but today, most litho companies actually print with a Black, Cyan, Magenta, Yellow sequence of lay down. Yellow goes ON TOP. Yellow printers' inks are currently very transparent, and do, in fact act as subtractive filters in conjunction with the white of the paper stock, thereby creating the thousands of colors we see on a printed page. Where inks values are printed as near solids the resulting colors are nearly totally created by the subtractive process. In very light tints, the dots are very small, and print side by side more than overlapping. So, for example, a tint of green gets produced by printing small dots of yellow and cyan side by side in a light tint. The results is always grayer than a solid made with the same ink in a solid area. There are many reasons for this "graying" effect, and it is known as "proportionality failure".

Bill :D

Richard Saylor
02-22-2004, 11:33 PM
Hey CMY Guy,

When you're right, you're right, and.......you're right. ;)


Thanks, Bill. I needed that. :)

thorkil
02-23-2004, 02:51 PM
A note on Bill's much earlier post regarding cadmium orange: I had to see for myself if Cad Orange was the only cadmium color that falls off to pink as it is tinted out. I did 12 strips, from pure color to mostly white (to show the 'pink' effect if present.) I used OH and some vintage Grumbacher Finest and Pre-Tested, plus one Windsor/Newton color, all mixed with Grumbacher Superba White (Titanium) :

Transparent Yellow (Aureolin) (Grum Finest) -- tints to Yellow
Cad Yellow Light (Grum Finest) --tints to Yellow
Cad Yellow Medium (Grum Finest) --tints to Yellow
Cad Yellow Deep (Old Holland) --tints to PINK
Cad Yellow Orange (Old Holland) -- tints to PINK
Cad Yellow Extra Deep (Old Holland --tints to PINK
Cadmium Orange (Old Holland) -- still PINK
Cadmium Orange (Grum. Finest) PINK
Cadmium Orange (Grum Pre-Test, 1960's tube) PINK
Cadmium Orange (Grum. Pre-Test, 1970's tube) PINK
Cadmium Orange (Grum Pre-Test, current production) PINK
Cadmium Orange (Grum Gainsborough- current production) PINK
Cadmium Scarlet (Winsor Newton) PINK

The first three stay a nice yellow all the way. All three are cadmium sulfide pigments. All of the rest are cadmium seleno-sulfide, and are decidedly pink when you get them mixed with lots of white. I'm not surprised by the last one (scarlet) as it should make a nice pink. But the cadmium yellow deep and extra deep were a surprise.

I will have to get tubes of Old Holland Cad Yellow Light and Medium to verify it, but I expect they will stay yellow.

Anyway, I learned something, so thanks, Bill, for your having mentioned this.

Michael24
02-23-2004, 02:52 PM
btw Michael...my comment is directed toward the use of the word "foolish" and in light of how technical discussions in color theory can get...I used your post to share thoughts intended for all. I do not for a moment think you actually feel folks such as myself foolish. I just pro-offer thoughts to remind all to leave room for all. Nothing wrong with enjoying indepth discussion...but in so doing we might forget the beauty of that which is yet simple.

btw....I have edited this post to more adequately represent my intent, and apologize for initial knee jerk reaction. I come into this conversation later...

peace

Larry

Larry and Others:

Larry:
Cmykguy put it succinctly. This is a response to someone who wants to reinvent the already discovered.

I mean no disrespect to any who wish to discuss color. However, the artists industry has been done a great disservice over the years by a number of experts who use poor science and faulty reasoning to influence a number of artists to make major changes in their selection of materials and techniques.

Example: In the 1920s Max Doerner used faulty sicientific technique and misled art historians and conservators as to the use of the pigment naples yellow.

Further, Doerner and De Mayerne and lots of other were proponents of secret mediums that they claimed the Old Masters used to make their paintings look the way they do.

Extensive work over many years by the International Rembrandt Committee, scientists at the National Gallery London, the Getty and the National Gallery, Washington, have published reviewed papers that refute these claims of a secret medium.

I can visit literally thousands of entries in wet canvas that use Doerner and De Mayerne faulty science to support their argument of the existence of leaded oil, resin mediums were the secret of the masters. No matter what volume of ignorance is shared back and forth in these web messages, it does not change the truth. Credible, well executed scientific study of hundred of paintings have not come up with the secret mediums so freely discussed today as gospel truth. The conservation community does not enjoy a wide distribution of its published works. Further, even when presented with the scientific analysis, lots of painters still just state that the material presented to refute modified oils is wrong, no reason, just plain wrong.

With regards to color theory:

When opinion and theory made from thin air is touted as research it raises the specter of foolish thinking and misplaced logic. This seems to come from one pattern of thought.

For the overly zealous, there seems to be some glee at toppling established knowledge and replacing it with their own version of the truth. They seem to proclaim that established color science somewhere got it all wrong, that the theorems and mathematical models that are in use by companies today are nothing but smoke and mirrors. It is no wonder that you cannot get expert color scientists to participate in forums like these. So many of the comments are so off the wall that member of this community do not wish to bother with forums like this one. I do not state this to mock anyone for ignorance. However, totally home-grown color theory or painting theory is unsettling to those with even the most rudimentary knowledge of color theory and practical mixing of color.

Bottom Line: One needs to do a bit of homework before coming to class. Ignorance is not a sin, but sticking to a faulty notion when a huge body of literature exists to refute it does smack of the foolishness that I spoke of in a previous post.

However, if my comments, while not directed at any individual have personally offended anyone. I apologize. I do not intend to be personally insulting. My comments are directed at the general notion by forum postings that reinvent already established color theory and try to replace it with material that does not have any foundation.

Trust me, I make foolish comments and postings all the time and expect my colleagues to challenge my statements. Many in this forum have called me to task on my comments and I believe we have discussed issues without hostility or attempting to one-up each other. From those who have something to offer I have learned much.

Posting do not have the advantage of quick sharing or voice tone, despite the stupid smiley faces we can paste on the end of text. So the shorthand writing and piece-meal way of writing back and forth does not allow for subtle debate.

A lot of the forum debates have good practical vs color theory components in them. In hindsight it would probably be better just to ignore off the wall color theory comments and respond only when a posting has some substance.

My greatest fear is that many visitors will pop in on 1 or 2 of those kind of post, get the notion that blue and yellow equals grey and spin off yet another branch of scientifically unsupported color theory to their fellow painters and others.

It amazes my colleagues and me as to how much misinformation flows about artists’ circles. I know that it is one main reason why conservators and conservation scientists will not participate in these types of forums where so much could be learned.

Larry: I have no problem with the comments you made and the means by which you teach color theory. I believe much the same as you do with regards to using practical means of mixing color and experimenting with color to achieve desired results. I agree with what you say and do not find cause to argue with your facts or methods.

However if you think it best not to call totally off the wall color theory generally foolish then I will abide by that and decide to limit my comments to the posts that I choose; ignore comments that are poorly researched or get out of Wetcanvas all together and spend my time elsewhere.

Michael Skalka, National Gallery of Art, Wash. DC

JamieWG
02-23-2004, 04:27 PM
Michael, you have made a tremendous difference here on Wetcanvas. It's hard to calculate, with 36,000 artists here, how many you have touched in a meaningful way. I had so many questions when you first poked your head into the Colorful Conversations thread here in the CT forum. You were so patient in explaining changes in materials and methods, and how to achieve archival soundness. I was unable to find real answers to very legitimate questions in any of the standard books and materials. In the short time you've been here, my own working methods have changed in significant ways. I know other artists have made changes as a result too. Yes, in addition to reinventing color theory, there will always be those who will continue to reinvent Maroger medium and discuss it in the forums. (You already know where you can find them, right?) You can't save 'em all! But there are far more who are listening quietly in the background, taking notes, and communicating about sound practices.

I know to seal masonite now and why.
I banished Damar from my studio.
I switched to alkyd mediums.
I use synthetic varnish.
That's not all, but it's a sampling of results brought on by information you have given to us here, not to mention all the worthwhile and interesting color discussions we've had.

You know better than anybody why we need you here. Most of us are here with a true desire to learn, and to support one another through the process.
I've so enjoyed our discussions about color, surfaces and mediums. I especially appreciate how everything you say is always in the context of actual painting, and not abstract theory without application potential. I'm sure you've been around the block with various artist's forums, and I think you can agree that Wetcanvas is the most congenial place to be found. There's a reason why so many more artists are on Wetcanvas than on the other online sites.....and we do hear you, believe me. Your presence has had a tremendous influence on many artists here. Moderate your posts as you need to for your own sanity, but please stay!
Jamie

LarrySeiler
02-23-2004, 08:36 PM
However if you think it best not to call totally off the wall color theory generally foolish then I will abide by that and decide to limit my comments to the posts that I choose; ignore comments that are poorly researched or get out of Wetcanvas all together and spend my time elsewhere.

Michael Skalka, National Gallery of Art, Wash. DC

Its like this Michael...(and I appreciate your thoughts, really), folks can talk and talk...but truth is best promoted by proof of the pudding, you know?

I look at the insistences of some for what they say is THEE only way color may rightly be thoughut of...and then go to look at their work. My reaction? Well...often it is that if their work is any indication of the true way to think about color as they suggest being IMHO as poor as perhaps their work is, then I'll take the wrong ways to think about color they are ranting against...thank you very much! hahaha.... you know?

In short...truth becomes somewhat self-evident by the testimony of that which is accomplished thru it.

By the same token...you need not use the word "foolish" to describe ideas. Your authority and professionalism will speak loud and true enough that the very "idea" of foolishness earned by someone is one you can rest assured we as members might conclude on our own without your having to use that descriptive adjective.

My position as a moderator requires by duty of that office to be sensitive and take offense when community rules appear to be ignored and one member attacks another.

I was the debate forum moderator for geez...let's see, maybe 3-4 years until recently. Over and over again I found myself repeating to participants in a reminder that they were free to attack the opinions of others, but not the opinionators.

It would be certainly well within reason to consider ideas foolish...however, decorum and protocol beg members to carefully word their thoughts before hitting send...so that it is clear that disagreement on an idea is in no way attaching itself to the sentiment that the bearer of such disagreement is not in character being attacked.

Have enough confidence in your own knowledge and experience to simply disagree....and let us as readers determine if that person opposed has ideas that are foolish or not. See what I'm saying?

I know that it is a difficult adjustment sometimes....and I suppose some professionals require a particular preferred treatment to cater to their presence. I am not in the least suggesting Michael that describes you...but hear me out if you will. I know professionals who just don't have the temperment to be patient, kind. They are often simply direct and tell it like it is and if everyone else can't handle it, too bad. But see...

....this is first and foremost a community, and a vision of its founder Scott Burkett. Rules are designed to create a level playing field. We benefit from your presence here...truly we do. I am sure the truths of expertise you share will yet be received and recognized for the gems they are without the need to appear judgmental of other member's character. That "appearance" of such I speak of....is when descriptive pronouns and adjectives take that slippery slope to describe people rather than their views.

To have "foolish" ideas says something negative about a person. Ssheesh, it requires a degree of foolishness to embrace foolish ideas. To say why you disagree with an idea and present your reasons why....isn't that productive in and of itself? I for one...hope you'll opt to stick around. If you choose to go though...then may you have good health and blessings that abound for you.

peace

Larry

Richard Saylor
02-24-2004, 12:40 AM
Larry, I certainly think that Michael should conform to the exact same standards that are expected of everybody else, but not stricter standards.

Here is what I mean. If I, cmyguy, were to describe someone's ideas as foolish, it would probably just be regarded as the opinion of some stupid southern illiterate and therefore not worth being concerned about. :) However, if Michael uses the word 'foolish,' then it has more intensity because of his position in the art world and is therefore more likely to generate ill-feelings.

I just have the feeling that perhaps Michael is being subjected to a higher standard of verbal expression than the rest of us. If that is not the case, then I will be more than glad to be corrected.

WFMartin
02-24-2004, 02:09 AM
A note on Bill's much earlier post regarding cadmium orange: I had to see for myself if Cad Orange was the only cadmium color that falls off to pink as it is tinted out. I did 12 strips, from pure color to mostly white (to show the 'pink' effect if present.) I used OH and some vintage Grumbacher Finest and Pre-Tested, plus one Windsor/Newton color, all mixed with Grumbacher Superba White (Titanium) :

Transparent Yellow (Aureolin) (Grum Finest) -- tints to Yellow
Cad Yellow Light (Grum Finest) --tints to Yellow
Cad Yellow Medium (Grum Finest) --tints to Yellow
Cad Yellow Deep (Old Holland) --tints to PINK
Cad Yellow Orange (Old Holland) -- tints to PINK
Cad Yellow Extra Deep (Old Holland --tints to PINK
Cadmium Orange (Old Holland) -- still PINK
Cadmium Orange (Grum. Finest) PINK
Cadmium Orange (Grum Pre-Test, 1960's tube) PINK
Cadmium Orange (Grum. Pre-Test, 1970's tube) PINK
Cadmium Orange (Grum Pre-Test, current production) PINK
Cadmium Orange (Grum Gainsborough- current production) PINK
Cadmium Scarlet (Winsor Newton) PINK

The first three stay a nice yellow all the way. All three are cadmium sulfide pigments. All of the rest are cadmium seleno-sulfide, and are decidedly pink when you get them mixed with lots of white. I'm not surprised by the last one (scarlet) as it should make a nice pink. But the cadmium yellow deep and extra deep were a surprise.

I will have to get tubes of Old Holland Cad Yellow Light and Medium to verify it, but I expect they will stay yellow.

Anyway, I learned something, so thanks, Bill, for your having mentioned this.

Hey, Thorkil,

Thanks for your support. I actually stumbled upon that phenomenon known as "overtone" while trying to prove (or disprove) the idea that white is a "cool" color. My contention was of course, it's "cooler" than orange, and should skew it's plot on a color wheel closer to the center the more white that got added to it, or perhaps remain at an "orange" plot on a color wheel, while simply becoming lighter.

However, to the contrary, I found the opposite to be true. It, instead stayed near the outer edge of the color wheel, and plotted progressively toward red. I learned from someone on another forum that this effect in a paint is know as an "overtone", and now I find myself using that effect extensively. On this forum someone awhile ago was asking just why they were having a difficult time lightening burnt sienna with white. "It keeps making PINK", They said. My advice was that it probably represented an overtone, and they'd simply need to add some yellow to keep it toward the burnt sienna with which they started.

Of course, this is not to suggest that every color to which you add white will take on a pink-er cast. I haven't experimented with other colors, but I would guess that the chances are that any color might actually move in ANY direction, toward ANY hue, and possibly even toward neutrality.

However, I find your chart absolutely fascinating, and I'd like to thank you for posting that. I use the knowledge of "overtones" quite often, now that I've experienced the effect, and find it quite time-saving and useful, to say the least.

Bill :)

thorkil
02-24-2004, 01:56 PM
However, I find your chart absolutely fascinating, and I'd like to thank you for posting that. I use the knowledge of "overtones" quite often, now that I've experienced the effect, and find it quite time-saving and useful, to say the least.

Bill :)

Thanks, Bill. I also think the real value of any of this type of discussion is it's application where the pigment hits the canvas. I will spend less time getting to the right color next time I do a sky or water reflection, just knowing about overtones.

regards,
Chuck

LarrySeiler
02-24-2004, 04:52 PM
I just have the feeling that perhaps Michael is being subjected to a higher standard of verbal expression than the rest of us. If that is not the case, then I will be more than glad to be corrected.

Certainly no higher standard than I expect from myself or is expected of other staff members I'm sure...but, I hear ya.

Its not easy sharing things we feel very empassioned about sometimes without offending. It is an effort. We have many folks here that think and create differently and not all at an equal level of understanding/growth. Doubly tough because we many wear our emotions on our shirt sleeves.

We all should be willing to take some useful criticism so that we are aware of how our intent might be taken wrong or offend others. I take your comments as just that very thing cmyguy...and thank you.

In the world as competitive as it is...we find ourselves in the position of having to best one another to simply develop a reputation. Certainly it is difficult to find things disagreeable an easy pill to swallow. I am no more comfortable offending Michael than I am a painter that expresses what I think is poor information or wrong ideas about color.

Certainly, Michael is deserving of grace and mercy to be spared from any potential embarrassment from me. I can't help but think we should likewise be concerned with the inferred high standard we set, a standard that suggests if one errors in sharing a wrong notion they too will suffer humiliation for it from us. So, I hear ya cmyguy..thanks. I gladly recant and relent perhaps over-reacting with Michael. Michael, please accept my apologies.

Speaking to all...and with reasonable standards to expect in mind, as a personal request, could we then agree to work also on not holding so high a standard as concerns our expectations of those perhaps not as astute or savvy to the knowledge of art that we might have...giving thus the same grace and mercy we're speaking here of?

I find myself constantly having to demonstrate patience among students teaching art as I do. Sometimes I cannot believe the level of ignorance or foolishness, but I'm paid to keep those thoughts to myself. Such silence does not intimate automatic agreement with such false notions. Instead it challenges me to find ways to bring about better changes. I don't always succeed....

peace

Larry

Twinbee
02-24-2004, 05:30 PM
Firstly, thanks everyone for your patience and the opportunity to learn in such a thoughtful and friendly forum!


In your examples, you should specify whether it is transmission or reflectance that is being combined. With inks/dyes, it would seem that you are combining transmissions.
So using light filters or inks, if cyan and yellow make green through transmission, what do those colours make through reflectance? If you could give a few colour mixing examples, this would be much appreciated.


Heh heh...you be the judge:
Lol, well that just about sums it up then :) Perhaps technology will come up with a purer blue paint in the future. What would make those graphs more useful is if it gave 3 bars - indicating the response for each eye cone.
Hmmm... how about "Cobalt Blue Deep"? (see http://www.handprint.com/HP/WCL/wheel.html).



BTW, I don't see the one from the Handprint site on your list - that's the site that gets used as reference most around here, for better or for worse.
Do you have the URL at hand?




Well, in terms of light filters and certain inks (say, felt-tip pens and the professional ink printing process), red and yellow will make red, not orange.
This isn't exactly true. First, the usual printing process doesn't use red - it uses magenta. And orange *is* created by mixing magenta and yellow - how else would it be created?

Sorry, yes I was referring to a theoretical red ink, not the 'red' that's produced by the mix of magenta and yellow.




Dyes and inks (especially felt-tips): multiplicative mixing
Light filters: multiplicative mixing

I still don't see what you think "multiplicative" mixing is - what you've described in the literal sense *never* happens. It might be interesting to discuss what actually happens with transparent layers of color as in felt tips and light filters, but I don't see any reason to assume anything is being multiplied...

It's true - the colour transmission of light filters is effectively getting multiplied - if only in a strict mathematical sense. See my first post on how orange and grey make dark brown through multiplicative mixing.




For me, 'subtractive' makes no sense because the resulting color changes according to which order you do the calculation. For example, red minus yellow (black if you count negative numbers as zero), is not the same as yellow minus red (green).

It makes perfect sense. Subtraction is not commutative. 5-3=2, but 3-5=-2. Reversing the order of subtraction changes the result into its opposite.

Yes, that's what I implied. The point is, the type of subtraction we've just discussed is never what is usually meant as 'subtractive mixing'. Like others have said, subtractive mixing generally refers to the mixed brightness being lower than either of the two paints to start with. I actually prefer the definition we've discussed, but I doubt the world will ever adopt it :)



Actually, you made a miscalculation with your ideal colors. It is true that Y-R=G. However, R-Y=M (magenta, not black). Note that magenta and green are complements: M=-G. So, analogous to what happens with numbers, Y-R=G, but R-Y=-G.

It is magenta if you 'shift' the numbers so that 0 is the lowest value. I got black because red (1,0,0) minus yellow (1,1,0) equals 0,-1,0. I took the minus one as zero to make 0,0,0. But of course, as you found out, if you add one to 0,-1,0, you get 1,0,1 which is magenta.



Please, please don't anyone tell me that this is nonsense because ideal colors don't exist. It's also true that projectiles are usually subject to air drag, which is ignored in a first approximation when computing their trajectories. (Sometimes I feel like I'm walking on eggshells.)
I'm the same =) I love talking about subjects in a theoretical way first, and then add any real world complexities afterwards, once the basics have been clarified.



My greatest fear is that many visitors will pop in on 1 or 2 of those kind of post, get the notion that blue and yellow equals grey and spin off yet another branch of scientifically unsupported color theory to their fellow painters and others.
Thanks to you and the others here, I've realised that the paints reflect a relatively large amount of wavelengths around a specific hue.
But rather than say that blue and yellow make green, wouldn't you say it's better to teach that they make grey. /And/ that the reason they make green in the world is because true blue paints don't exist as yet? (so one is actually really mixing cyan/blue and yellow)

Or perhaps I'm wrong. Assuming absolute pure blue paint could be produced. If this was mixed with yellow, I presume this would make grey?

Also, can you confirm that 'subtractive' is more of a colloquial term, rather than an exact mathematical process. I'm tempted to think that no inks, paints or filters use subtraction in any strict mathematical sense, but maybe you can clarify this.

Marc Sabatella
02-24-2004, 06:01 PM
[ re: handprint site graph of cone response for different wavelengths ]
Do you have the URL at hand?


No; you'll have to browse around handprint.com. But I doubt their chart is significantly different than the others you found.


It's true - the colour transmission of light filters is effectively getting multiplied - if only in a strict mathematical sense.


How so? It seems it has already been demonstrated quite convincingly that this isn't true regarding mixing of opaque media, although the assumption is still that geometric mean makes a reasonable approximation if you don't care too much about specifics. But I have yet to see any sort of argument that light filers operate by multiplying at all. I might imagine that mathetmically, the "minimum" function might actually be the closest simple approximation. That is, the transmission at a given wavelength will be the minimum of the transmission of the source filters at that wavelength. This at least would explain why red + yellow appears to remain red - although I'm just taking your word for it there.


Thanks to you and the others here, I've realised that the paints reflect a relatively large amount of wavelengths around a specific hue.
But rather than say that blue and yellow make green, wouldn't you say it's better to teach that they make grey. /And/ that the reason they make green in the world is because true blue paints don't exist as yet?


To me, this would be like teaching people "If you throw a baseball, it will stay in the air forever. The only reason it falls in the real world is that anti-gravity devices don't exist as yet".

The real world has gravity, and while it can indeed be useful to discuss what would happen in the absence of gravity, you would normally treat this as the exception. For most purposes, it makes more sense to discuss how balls behave here on earth.

Similarly, while there might be some theoretical world in which a blue paint exists that reflects only blue light (and same for other colors), and it can indeed be useful to discuss how paints would behave in such a world, that isn't our world, so I don't see any reason to go around pretending it is. It isn't just some quirk of history that we haven't happened to discover a pure blue pigment. Pretty much *nothing* on this earth is pure color in the sense of reflecting only one color of light. Might as well just accept this reality.


Or perhaps I'm wrong. Assuming absolute pure blue paint could be produced. If this was mixed with yellow, I presume this would make grey?


If this paint truly reflected only blue wavelengths, and the yellow paint mixed only yellow wavelengths, the resulting mix would be essentially black.

Richard Saylor
02-24-2004, 07:11 PM
But I have yet to see any sort of argument that light filers operate by multiplying at all.

Let's start with a very simple example. Neutral density filters transmit the full spectrum without favoring any single color. They just reduce the amount of light. Suppose I have two neutral density filters, each of which transmits 50% of the light. If I stack them and shine a light through them, what comes out of the first filter has 1/2 the intensity of the original light. Now this already filtered light passes through the second filter, and again the intensity is cut in half. Obviously the combined effect of the stacked filters is to pass 1/2 of 1/2 of the original source, or 1/4 (25%).

Obviously the same computation works if, say the first filter passes, say, 75% and the second filter passes, say 25%. The source hits the first filter, which passes 3/4 of the light, which hits the second filter, which passes 1/4 of what hits it. The combined effect of the two stacked filters is to pass only 1/4 of 3/4 of the source, which would be 3/16.

I think it's clear that the same argument works for filters which pass different percentages of different wavelengths. E.g., if one filter transmits 50% blue and another transmits 25% blue, then if they are stacked, together they transmit 1/4 x 1/2 = 1/8 or 12.5% blue.

Richard Saylor
02-24-2004, 07:38 PM
Twinbee, I'll try to answer some of your questions.

So using light filters or inks, if cyan and yellow make green through transmission, what do those colours make through reflectance? If you could give a few colour mixing examples, this would be much appreciated.
It's still 'subtractive' color mixing (i.e., absorbances get combined, not reflectances), so you would get green. Cyan pigment absorbs red, yellow pigment absorbs blue. Mix them, and red and blue get absorbed, so the mixture will reflect green. By the way, subtractive mixing really just refers to adding negatives (absorbances).

It is magenta if you 'shift' the numbers so that 0 is the lowest value. I got black because red (1,0,0) minus yellow (1,1,0) equals 0,-1,0. I took the minus one as zero to make 0,0,0. But of course, as you found out, if you add one to 0,-1,0, you get 1,0,1 which is magenta.

Don't blame R-Y=M on Marc. :) I don't know what sort of subtraction you're doing. If you are doing binary subtraction, 100-110=-010, the complement of green, which is magenta. If you are subtracting number triples (like rectangular coordinates), then (1,0,0)-(1,1,0)=(0,-1,0)=-(0,1,0), which again is the complement of green.

Patrick1
02-24-2004, 10:17 PM
What would make those graphs more useful is if it gave 3 bars - indicating the response for each eye cone.
I was thinking the same thing. There isn't a cone reponse chart for every pigment on that site, but there is for a few selected ones:

http://www.handprint.com/HP/WCL/color2.html#hue
http://www.handprint.com/HP/WCL/color3.html#opponent

Here's a neat site to play with...you can shape the reflectance curve to any shape you want, and you'll get the resultant cone response (the 'Metamers' demo is probably the best for this):

http://www.cs.brown.edu/exploratories/freeSoftware/catalogs/color_theory.html

Michael24
02-25-2004, 08:38 AM
Twinbee:


Sorry, I can’t quite get a hang of the quote boxes like you have so bear with me as I comment on your inquiries.

Twinbee said:
Don't you mean grey? It would only make a (muddy) green if the blue was not real blue, but had a tinge of green/cyan in. Either that, or there's some unpredictable chemical reaction in the inks that causes the slight green hue.

Because blue is at the end of the visible spectrum, I would imagine that it's fairly easy for paint manufacturers to avoid green pollution in blue paint, but I may be wrong.

My response:
I am not sure what you mean by a real blue? Blues like ultramarine come with a spectral response that is high in the blue as well as red, the same for cobalt blue. Indanthrene has a small red response but a high blue green reflectance pattern. I can’t comment on inks because I have not tested them. However, let me focus on where I am coming from with pigments. No pigment has a single narrow wavelength of spectral reflectance. All pigments reflect or absorb from 370 nm to about 730 nm. Each pigment reflects in its own way so that mixtures of different combinations of yellow and blue pigments will form a range of green hues, some low chroma, some high chroma, some yellow-green and blue-green hues, etc., because each pigment combination will react differently when mixed. I do not believe that the changes are chemical in nature.
Regarding blues position in the spectrum: Search the net and find a visible light spectrum for reference. It will help visualize this. I paste one on many of my spectral curves just to keep my interpretation straight. The visible light spectrum starts at 370- violet, 450 - blue violet, 480 – blue, 590 – cyan, 540 – green, 580 – yellow, 600- orange, 650 – red. (I go from short waves to longer. You may see a lot of web sites with long to short wave order.) So blue does not appear at the end in terms of wavelength. Green pollution is unavoidable since spectra as I mentioned up top is not narrow banded so blue being next to green in wavelength is present to some degree in blue colorants.


Twinbee said:
Could you give me an example with the orange and grey I posted? For me, 'subtractive' makes no sense because the resulting color changes according to which order you do the calculation. For example, red minus yellow (black if you count negative numbers as zero), is not the same as yellow minus red (green). Surely 'multiplicative' is a more appropriate term instead?

Response:
I will need to dig up your posting and find something on this. Meanwhile, I think I know where you are coming from on the calculation dilemma. A lot of books show RGB response curves and cutoffs for how the eye sees and this may be what has entered into your argument. It is called subtractive because the dominant portions of the spectra win when two colors mix. Put otherwise, the dominant SUBTRACTS out the subordinate parts of the spectrum. Its not an ON / OFF, positive and negative relationship. No negative numbers are involved when measuring spectra with a spectrophotometer. It would be messy, but I could post the response curve in nanometers for a selected color. Better yet, I could send you an Excel spreadsheet with a colors spectra and how it graphs. Let me know.

Twinbee said:
Yep, I would be very interested to know too. Is any relatively simple combination of subtraction, multiplication, addition, division (!), and the various types of averages enough to roughly approximate the Kubelka-Monk model.

Response:
I need to find this out for myself and others who have asked. It isn’t as simple as taking the mean of the two spectral readings. I have another spreadsheet on that and the formula for making the calculation would choke a horse. I will see if I can find a simpler explanation that is understandable. So far no luck with this.

Generally, don’t get caught up with eye cone response and subtractive color. The two are not interchangeable. A really well written book on this, AND it addresses artists works is Margaret Livingstone’s *Vision and Art: The Biology of Seeing* Its got spectra and the biology of how the eye responds and everything else.

Happy Color Hunting

Michael Skalka, National Gallery of Art, Washington, DC

Richard Saylor
02-25-2004, 11:13 AM
If you are doing binary subtraction, 100-110=-010, the complement of green, which is magenta. If you are subtracting number triples (like rectangular coordinates), then (1,0,0)-(1,1,0)=(0,-1,0)=-(0,1,0), which again is the complement of green.
I'm wrong, neither ordinary binary subtraction nor bitwise binary subtraction works. The above computation is a fluke. However, under no circumstances will you get anything like black. In the real world, most red pigments reflect red and (to a lesser extent) yellow. If you remove the yellow reflectance, you just get a redder red. If the original red has a significant blue component (like alizarin crimson), removing yellow will make it lean more toward magenta.

Marc Sabatella
02-25-2004, 11:48 AM
say the first filter passes, say, 75% and the second filter passes, say 25%. The source hits the first filter, which passes 3/4 of the light, which hits the second filter, which passes 1/4 of what hits it. The combined effect of the two stacked filters is to pass only 1/4 of 3/4 of the source, which would be 3/16.

I think it's clear that the same argument works for filters which pass different percentages of different wavelengths. E.g., if one filter transmits 50% blue and another transmits 25% blue, then if they are stacked, together they transmit 1/4 x 1/2 = 1/8 or 12.5% blue.


Makes perfect sense; thanks for explaining what, I suppose, should have been obvious.

Michael24
02-25-2004, 01:09 PM
Thanks to you and the others here, I've realised that the paints reflect a relatively large amount of wavelengths around a specific hue.
But rather than say that blue and yellow make green, wouldn't you say it's better to teach that they make grey. /And/ that the reason they make green in the world is because true blue paints don't exist as yet? (so one is actually really mixing cyan/blue and yellow)

Or perhaps I'm wrong. Assuming absolute pure blue paint could be produced. If this was mixed with yellow, I presume this would make grey?



Twinbee:

Sorry I am either so bad a speed reader or really stupid. I get what you mean about grey and in reality the pure blue and pure yellow (again if one could obtain pure colors) would form blocks, thus a void and reflect nothing - therefore black. But since that can't exist right now in the real word we are stuck with green as a reflectance.

In your first paragraph above, I am not sure what you mean by *And/ that the reason they make green in the world is because true blue paints don't exist as yet?* The graphs of the spectra should indicate why any blue and yellow will make a green. Mind you, they may not be a high chroma green or a warm or cool green you desire, but it will make a green. Interesting to note - the blue pigment is cobalt blue which as you can see has a high reflectance in the red region. This mutes any greens made with in and yellow since the high degree of red is subtracting out the chroma of the green that is produced. - thus a muted green, but green none the less.

Questions on the spectra. Let me know.

Michael Skalka
National Gallery of Art, Wash. DC

thorkil
02-25-2004, 01:44 PM
If this paint truly reflected only blue wavelengths, and the yellow paint mixed only yellow wavelengths, the resulting mix would be essentially black.


Not only that, but if the theoretical 'pure' blue & yellow pigments were reflecting only a tiny, tiny slice of the spectrum (what are "pure" colors anyhow?) they would appear very dark, perhaps nearly black, in any normal light. (There is only so much light of a particular wavelength available in a given illumination to hit the pigment, right?) So, the more wavelength-specific a theoretical pigment is, the less reflectance. I think. :D

In other words, I think our human eyes would see a 'pure' yellow pigment as black. Kinda weird, and not too useful for artists.

(Exception to the above would be a theoretical situation where the pigment is isolated in a non-reflective environment, and the illumination is brought up to a sufficiently high level that the one wavelength reflected to our eyes becomes visible. Again, not too applicable to our everyday reality, right?)

Chuck

Richard Saylor
02-27-2004, 01:16 PM
The graphs of the spectra should indicate why any blue and yellow will make a green. Mind you, they may not be a high chroma green or a warm or cool green you desire, but it will make a green. Interesting to note - the blue pigment is cobalt blue which as you can see has a high reflectance in the red region. This mutes any greens made with in and yellow since the high degree of red is subtracting out the chroma of the green that is produced. - thus a muted green, but green none the less.
I took your spectral graphs and superimposed them. (The magenta line is the curve for green.) This shows clearly that the reflectance of the mixture is not the geometric mean. In fact, it follows very closely the minimum of the yellow and blue curves. I wonder whether this is generally true of mixtures.

Patrick1
02-28-2004, 07:19 AM
Thanks Richard for superimposing the graphs...much easier to compare.

This shows clearly that the reflectance of the mixture is not the geometric mean. In fact, it follows very closely the minimum of the yellow and blue curves. I wonder whether this is generally true of mixtures.

I wonder too...at least to get a rough estimate. It's interesting that around the middle, the reflectance of the mix is less than either the blue or yellow.

It looks like Handprint greatly overstates the value of geometric mean or ''the visual average shifted somewhat toward the darker reflectance profile at each wavelength'' to get an estimate.

Richard Saylor
02-28-2004, 11:14 AM
Hi, Domer. I'm just trying to get a handle on some sort of rough approximation. I have the feeling that the reflectance of mixtures may generally lie somewhere between the geometric mean and the minimum curve. The Kubelka-Munk formulas are complicated but don't involve anything more mathematically sophisticated than the exponential function. Anyone experienced with a scientific calculator could work them. However, the formulas are practically useless because they use information other than simply the reflectance spectra, such as the absorption spectrum, scattering spectrum, sample thickness, and the reflectance spectrum of the substrate.

Richard Saylor
02-28-2004, 12:29 PM
It's interesting that around the middle, the reflectance of the mix is less than either the blue or yellow.
Yes, at around 550nm, blue and yellow are at 50% and green is below that. That could be accounted for if the spectrophotometer just takes a finite number of readings at selected wavelengths and a continuous reflectance curve is constructed by interpolation.

Michael24
03-01-2004, 02:32 PM
Yes, at around 550nm, blue and yellow are at 50% and green is below that. That could be accounted for if the spectrophotometer just takes a finite number of readings at selected wavelengths and a continuous reflectance curve is constructed by interpolation.

cmyguy: You are just too clever!!! I never thought about superimposing all three curves. Interesting to see the result. Yes, the Hunter spectrophotometer takes a reading at every 5 nm and interpolation makes the continuous curve via Microsoft Excel.

As you can see its not just the geometric mean. Every single and mixed color has it own spectral response that has a bit of predictability and some unpredictability. I don't know if scatter, surface distortion or luminosity play a role in the spectral, but they do have some surprises.

We note the *red tail* (an abrupt increase in reflectance on the right end of the spectral curve) on many pigments does play an important role in how a color will appear when mixed. It have nearly everything to do with the failure of inorganic pigments (cad yel. and ultramarine or cobalt blue) to make clean high chroma greens. The do grey out but not because of chemical or theoretical issues. They grey because the red reflectance in the cobalt blue neutralizes the green of the cad yel and cobalt mixture. It is as if you are adding a bit of red to the cad yellow cobalt blue mix. That's how much red is being reflected in cobalt blue.

Again, love what you did with the graphs.

Michael Skalka
Conservation, National Gallery of Art, Wash. DC

Einion
03-01-2004, 04:28 PM
The Kubelka-Munk formulas are... practically useless because they use information other than simply the reflectance spectra, such as the absorption spectrum, scattering spectrum, sample thickness, and the reflectance spectrum of the substrate.
I'd go further, they're of no use to artists whatsoever (regrettably!)

I'm just trying to get a handle on some sort of rough approximation. I have the feeling that the reflectance of mixtures may generally lie somewhere between the geometric mean and the minimum curve.
I can certainly appreciate why you'd want to do this but I'm afraid anything one comes up with can only approximate certain results and if I'm thinking about this correctly won't account for specific 'oddball' mixtures and other effects.

I took your spectral graphs and superimposed them. (The magenta line is the curve for green.) This shows clearly that the reflectance of the mixture is not the geometric mean. In fact, it follows very closely the minimum of the yellow and blue curves. I wonder whether this is generally true of mixtures
Nope, it's certainly about right in specific instances but it's not widely true: the relationship of subtractive mixing complements is very irregular, hence no simple idea can account for all mixes; off the top of my head there are bowed mixing lines, plus the unexpected hues in some near-complement mixes aren't accounted for. I think, again, this is accounted for by the undercolour, the information for which isn't inherent to reflectance profiles.

I think I've posited this previously in a discussion with Bill, but even if we had measurements of more of the parameters of pigments than just the masstone reflectance, the results still couldn't be collated and point to anything more than broad relationships (much less be a useful predictor) because of the unusual effects one can get at different concentrations of two or more colours, the influence of the binder's colour and refractive index, substrate/underlying colour, film thickness, surface gloss etc. etc. There are just too many variables to take into account in paints to do anything other than, maybe, predict a mix at a given proportion applied at complete opacity with a specific surface finish.


Michael, great to see you here and have your input.

Einion

Richard Saylor
03-01-2004, 10:33 PM
Michael and Einion, thanks for the clarification.

Twinbee
03-08-2004, 03:32 AM
Hi all,



Thanks to you and the others here, I've realised that the paints reflect a relatively large amount of wavelengths around a specific hue.
But rather than say that blue and yellow make green, wouldn't you say it's better to teach that they make grey. /And/ that the reason they make green in the world is because true blue paints don't exist as yet?

To me, this would be like teaching people "If you throw a baseball, it will stay in the air forever. The only reason it falls in the real world is that anti-gravity devices don't exist as yet".

Hehe.. that's not a bad analogy.
But if someone wants to learn the basics of color theory, then I can imagine it might be somewhat confusing to have two different results for the same mix. Especially if they try to equate what's happening with what a colour wheel would imply. The main problem we have is that there are two definitions of blue; one that's true blue, and the other that's: "let's call it blue, because it's close enough blue" ;)


The real world has gravity, and while it can indeed be useful to discuss what would happen in the absence of gravity, you would normally treat this as the exception. For most purposes, it makes more sense to discuss how balls behave here on earth.
Actually, it's simpler to discuss the effects without gravity, because the mathematics involved would then be easier to learn. How gravity effects things should be taken into account afterwards. The same goes for the 'true blue' argument.


If this paint truly reflected only blue wavelengths, and the yellow paint mixed only yellow wavelengths, the resulting mix would be essentially black.
True. But judging by the complexity of the Kubelka-Monk model, might it be a dark shade of grey?



It's still 'subtractive' color mixing (i.e., absorbances get combined, not reflectances), so you would get green.
Okay. This is what I imagined, so I was hoping you'd say that :)



By the way, subtractive mixing really just refers to adding negatives (absorbances).

I found this site which might explain how the term 'subtractive' first came into use. For many decades, photographers have always measured logarithmically. And obviously, subtracting (or adding) logarithmic values is equivalent to normally multiplying. For more info, see this site (http://www.gimlay.org/~andoh/cg/faq/ColorFAQ.html#RTFToC22)



Don't blame R-Y=M on Marc. :-) I don't know what sort of subtraction you're doing. If you are doing binary subtraction, 100-110=-010, the complement of green, which is magenta. If you are subtracting number triples (like rectangular coordinates), then (1,0,0)-(1,1,0)=(0,-1,0)=-(0,1,0), which again is the complement of green.
Rectangular coordinates it is...



The visible light spectrum starts at 370- violet, 450 - blue violet, 480 - blue, 590 - cyan, 540 - green, 580 - yellow, 600 - orange, 650 - red. (I go from short waves to longer. You may see a lot of web sites with long to short wave order.) So blue does not appear at the end in terms of wavelength.
Am I right in saying that a mono-spectral wavelength at the very end of the visible spectrum (deep violet - around 370nm) activates the red cone slightly as well as the blue cone?



If you are doing binary subtraction, 100-110=-010, the complement of green, which is magenta. If you are subtracting number triples (like rectangular coordinates), then (1,0,0)-(1,1,0)=(0,-1,0)=-(0,1,0), which again is the complement of green.

I'm wrong, neither ordinary binary subtraction nor bitwise binary subtraction works. The above computation is a fluke. However, under no circumstances will you get anything like black. In the real world, most red pigments reflect red and (to a lesser extent) yellow. If you remove the yellow reflectance, you just get a redder red.

My reasoning for saying red minus yellow equals black was based on a purely abstract level since yellow itself contains red (in the RGB model anyway). Think of R minus RG, and then count zero as a cutoff point.




Thanks to you and the others here, I've realised that the paints reflect a relatively large amount of wavelengths around a specific hue.
But rather than say that blue and yellow make green, wouldn't you say it's better to teach that they make grey. /And/ that the reason they make green in the world is because true blue paints don't exist as yet? (so one is actually really mixing cyan/blue and yellow)

Or perhaps I'm wrong. Assuming absolute pure blue paint could be produced. If this was mixed with yellow, I presume this would make grey?

Sorry I am either so bad a speed reader or really stupid. I get what you mean about grey and in reality the pure blue and pure yellow (again if one could obtain pure colors) would form blocks, thus a void and reflect nothing - therefore black. But since that can't exist right now in the real word we are stuck with green as a reflectance.

Yep, that's right. Incidentally, I think that a near fully saturated true blue (that is; a blue with equal green and red pollution, say... 100% blue, 20% green, 20% red), is possible with real paints. This would make a dark grey if mixed with yellow. So in other words, we can get a real blue hue with paint, but at the cost of the saturation level :)



In your first paragraph above, I am not sure what you mean by *And/ that the reason they make green in the world is because true blue paints don't exist as yet?* The graphs of the spectra should indicate why any blue and yellow will make a green.

Yes, they only make green because the 'blue' isn't true blue. It contains a green bias in.

I'll attempt to rephrase the paragraph:
"True blue and yellow will always mix to black/grey, but the closest blue paints have a green bias in them, which is why you get a greenish grey instead."

Just for the record, I realise 'true blue' (100% response of the blue cone, and 0% red and green cone response) isn't possible even with a mono-spectral laser. The red and green cones would need to be disabled completely for anyone to see the truest possible blue.

Marc Sabatella
03-08-2004, 01:12 PM
The main problem we have is that there are two definitions of blue; one that's true blue, and the other that's: "let's call it blue, because it's close enough blue" ;)


True, but given that most people will never see the former in their entire lives, I'm not sure it makes sense to give it that much attention. See below.


Actually, it's simpler to discuss the effects without gravity, because the mathematics involved would then be easier to learn. How gravity effects things should be taken into account afterwards. The same goes for the 'true blue' argument.


Depends on what you're learning. If we're talking about just getting the basics across to a group of schoolchildren, I don't see the point of confusing them with things that don't actually exist in reality. It's not going to helpthem understand the real world one bit. And even to the extent it makes sense to start getting into the math involved with, say, a high school student, I still don't see this as coming down to any more than spending 30 seconds telling them that without gravity, you could throw a ball to the moon, but then spend the next week working out the equations for how gravity factors in. Of course, we're still ignoring the drag of air resistance; this strikes me as perhaps similar to using geometric mean or min or some such approximation to the full calculation of reflectances.

But the main point here is, I don't see any value in giving more than a passing thought on the things that *don't* happen in the real world except if you're making a lifelong study of this. Spending more than 30 seconds talking about situations in blue and yellow don't make green has no value whatsoever to 99.99999% of people on earth, and even 99.9% of artists.


Incidentally, I think that a near fully saturated true blue (that is; a blue with equal green and red pollution, say... 100% blue, 20% green, 20% red), is possible with real paints. This would make a dark grey if mixed with yellow.


I have yet to see any sort of even remotely blue pigment that doesn't create something we'd all recognize as green when mixed with anything we'd recognize as yellow. You'll have to name some actual pigments to convince me otherwise. Even ultramarine mixed with cad yellow deep looks green. Not even a slightly green grey, but a color that a child would look at and instantly call green.

Richard Saylor
03-08-2004, 02:19 PM
My reasoning for saying red minus yellow equals black was based on a purely abstract level since yellow itself contains red (in the RGB model anyway). Think of R minus RG, and then count zero as a cutoff point.

I see exactly what you mean. Rather than subtraction (the inverse of addition), that's what I would think of as the Boolean product (multiplication/set-intersection) of red and not-yellow. I.e., all reflected colors common to red and blue, which would indeed be nuttin' (black).

Richard Saylor
03-08-2004, 03:15 PM
But the main point here is, I don't see any value in giving more than a passing thought on the things that *don't* happen in the real world except if you're making a lifelong study of this. Spending more than 30 seconds talking about situations in blue and yellow don't make green has no value whatsoever to 99.99999% of people on earth, and even 99.9% of artists.
The point is that colors like blue, yellow, and green are abstractions, like numbers. A pound of potatoes is never exactly one pound. Colors in the real world are bluish, yellowish, and greenish, not blue, yellow, and green.

Let's get away from yellow and blue and onto a very similar issue. Kids learn early on (at least I did) that red and blue make purple. Let's say you're teaching an oil painting class, and one of your students notices that one of the most popular reds (cadmium, which came in the oil paint set which s/he got for Christmas) does not make purple when mixed with blue. Are you going to tell her/him that it is just an anomaly, and that s/he is only 1 out of 1000 people who is weird enough to be concerned about this?

I may be strange, but it seems to me that exactly what we mean by 'red' is quite important, especially to artists. (Of course, I'm also strange enough to wonder about what is exactly meant by the number 'six,' so maybe my opinion is automatically disqualified.)

Marc Sabatella
03-08-2004, 06:29 PM
Let's get away from yellow and blue and onto a very similar issue. Kids learn early on (at least I did) that red and blue make purple. Let's say you're teaching an oil painting class, and one of your students notices that one of the most popular reds (cadmium, which came in the oil paint set which s/he got for Christmas) does not make purple when mixed with blue. Are you going to tell her/him that it is just an anomaly, and that s/he is only 1 out of 1000 people who is weird enough to be concerned about this?


Not at all - that's the level at which I *do* believe it makes to sense to think about color for most people. There are very real practical reasons cadmium red makes a violets, and that is precisely what I advocate concentrating on what happens in the real world. We can teach about this - using reflectance diagrams if necessary, or just the Michael Wilcox simplification using the color wheel - without ever resorting to talking how some mythical color that reflected only a single wavelength would behave.

Richard Saylor
03-09-2004, 09:14 AM
I found this site which might explain how the term 'subtractive' first came into use. For many decades, photographers have always measured logarithmically. And obviously, subtracting (or adding) logarithmic values is equivalent to normally multiplying. For more info, see this site (http://www.gimlay.org/~andoh/cg/faq/ColorFAQ.html#RTFToC22)
That's a good read. Thanks. What they are calling the density of a filter is called the absorbance in other contexts.

A = -log(T) = log(1/T)

where the transmittance T is the ratio of the output to the input of the filter. (The following is an explanation for the non-mathematically inclined.) When you are mostly interested in blocking certain wavelengths, A is really handy and intuitive. The basic unit is A = 1, the absorbance corresponding to a transmission of .1 (10%). If A = 5, the effect is the same as stacking 5 filters, each having an absorbance of 1. For perfect 100% transmission, A = 0. If the filter completely blocks the wavelength, A = infinity (like stacking an infinite number of A = 1 filters, ha-ha).

Yeah, photographers do use logarithms a lot, especially base 2 (which are proportional to base 10, so there's no difference in the computations) but less and less as photographic equipment becomes more and more automated (sigh). Being a mathematician and an amateur photographer, I like to use about the most non-automated (some would say antiquated) equipment available, my basic camera being a totally manual Leica rangefinder. :)

Einion
03-09-2004, 02:21 PM
Incidentally, I think that a near fully saturated true blue (that is; a blue with equal green and red pollution, say... 100% blue, 20% green, 20% red), is possible with real paints.
This isn't a particularly useful way of thinking about colour unless you're working in the digital realm, but if you want to examine this colour specificially why don't you plug in the numbers to a graphics program and have a look at it?

This would make a dark grey if mixed with yellow. So in other words, we can get a real blue hue with paint, but at the cost of the saturation level
If you check 51, 51, 255 in RGB values you'll see immediately that it isn't a saturated blue for a start... which should be reasonably easy to see because of the proportion that amounts to white light (20% in case it's not evident). And in paint it won't mix anything like grey with any yellow, even a yellow earth. If you continue your research you should easily be able to learn that you can mix a colour like this pretty easily and it will indeed mix a definite green just as Marc says.

While it's interesting to examine the underlying principles of colour relationships I think one should always work at it from the basis of looking for something with a practical application, so any imaginings of mono-spectral colours and so forth is only of academic interest (pun intended).

Einion

Einion
03-09-2004, 02:26 PM
A pound of potatoes is never exactly one pound. Colors in the real world are bluish, yellowish, and greenish, not blue, yellow, and green.
http://www.wetcanvas.com/Community/images/20-Aug-2003/3842-thumbsup.gif

Let's get away from yellow and blue and onto a very similar issue. Kids learn early on (at least I did) that red and blue make purple. Let's say you're teaching an oil painting class, and one of your students notices that one of the most popular reds (cadmium, which came in the oil paint set which s/he got for Christmas) does not make purple when mixed with blue. Are you going to tell her/him that it is just an anomaly, and that s/he is only 1 out of 1000 people who is weird enough to be concerned about this?
Examples like this are at the crux of my argument about the value of teaching better colour theory than the basic RGB model. Teaching the elementary principles of colour bias, just for a start, would easily inform why one mix is better than another.

But I do have to point out that with the exception of a mix of a few specific colours (Phthalo Blue GS and Cadmium Red Light for example) what you will get is in fact violet in hue, it just might be low in chroma. Once we get to a given level (I don't know what age this would be appropriate for, maybe 12 or so these days?) I see no reason not to teach the concepts of hue, saturation and value which would allow an accurate description of what one does get and hence gives the student the visual vocabulary to better understand what they see. But at the basic level they can at least be shown that mix A, Ultramarine and Quin Rose, gives a good violet, while B, Phthalo Blue GS and Cad Red, gives a much duller one and a simple explanation of why.

I may be strange, but it seems to me that exactly what we mean by 'red' is quite important, especially to artists.
Yes indeed it should be. That way we could stop thinking about scarlet and crimson as simply two members of one imaginary family and expecting similar behaviour from them.

Of course, I'm also strange enough to wonder about what is exactly meant by the number 'six,' so maybe my opinion is automatically disqualified.
Hehehe.

Einion

Twinbee
03-13-2004, 04:10 AM
But the main point here is, I don't see any value in giving more than a passing thought on the things that *don't* happen in the real world except if you're making a lifelong study of this. Spending more than 30 seconds talking about situations in blue and yellow don't make green has no value whatsoever to 99.99999% of people on earth, and even 99.9% of artists.
Perhaps it's just me then :) Somehow, I see some kind of universal simplicity and beauty in teaching the way that there are 3 basic pairs of complementary colours. And the way it all matches with our 3 RGB eye cones, and the way that light sources such as television display 3 of them, and that inks use the other 3 in a transmissive process.

The only disadvantage is that someone who's learning colour for the first time would mix yellow and blue paint and see that it doesn't make a hueless colour (and so it would need to be explained that the blue paint they use isn't 100% pure in hue). But I think that's a small 'price' to pay for teaching such a cool concept...



I have yet to see any sort of even remotely blue pigment that doesn't create something we'd all recognize as green when mixed with anything we'd recognize as yellow. You'll have to name some actual pigments to convince me otherwise. Even ultramarine mixed with cad yellow deep looks green. Not even a slightly green grey, but a color that a child would look at and instantly call green.
Yes, ultramarine blue activates the green cone quite a lot according to this diagram (http://www.handprint.com/HP/WCL/IMG/RC/rcPB29.jpg) - (not to mention red!). A 100% blue hue would look relatively violet compared to that hue.


Let's get away from yellow and blue and onto a very similar issue. Kids learn early on (at least I did) that red and blue make purple. Let's say you're teaching an oil painting class, and one of your students notices that one of the most popular reds (cadmium, which came in the oil paint set which s/he got for Christmas) does not make purple when mixed with blue. Are you going to tell her/him that it is just an anomaly, and that s/he is only 1 out of 1000 people who is weird enough to be concerned about this?

I may be strange, but it seems to me that exactly what we mean by 'red' is quite important, especially to artists. (Of course, I'm also strange enough to wonder about what is exactly meant by the number 'six,' so maybe my opinion is automatically disqualified.)

Hehe, that wouldn't happen to be related to whether one should start counting from zero or one would it? :)




Yep, that's right. Incidentally, I think that a near fully saturated true blue (that is; a blue with equal green and red pollution, say... 100% blue, 20% green, 20% red), is possible with real paints. This would make a dark grey if mixed with yellow. So in other words, we can get a real blue hue with paint, but at the cost of the saturation level
If you check 51, 51, 255 in RGB values you'll see immediately that it isn't a saturated blue for a start...
That's what I said, yes. The saturation isn't full, but the hue can be.



which should be reasonably easy to see because of the proportion that amounts to white light (20% in case it's not evident).

And in paint it won't mix anything like grey with any yellow, even a yellow earth.
I'm inclined to think that the reason that mix produces a green hue is because the blue was not really 100% blue in hue. Correct me if I'm wrong...



If you continue your research you should easily be able to learn that you can mix a colour like this pretty easily and it will indeed mix a definite green just as Marc says.
If what you say is true (that a not fully saturated, but a blue that's 100% blue in terms of hue) still makes green when mixed with yellow, then this leads me to believe that there is some complex chemical reaction going on which 'distorts' the outcome of the mixed hue.

Marc Sabatella
03-14-2004, 05:07 PM
If what you say is true (that a not fully saturated, but a blue that's 100% blue in terms of hue) still makes green when mixed with yellow, then this leads me to believe that there is some complex chemical reaction going on which 'distorts' the outcome of the mixed hue.

Depends on what you mean by "100% blue in terms of hue". If you mean, one that simply *looks* as blue as can be, then no fancy chemical reaction is needed to explain why this makes green when mixed with yellow - it is simply yet another example of the basic subtractive mixing principles we all know and love. Because our eyes have never seen a single-wavelength blue paint, and interpret colors with quite a bit of green and vioelt in them as still being a blue as can be if some sort of average wavelength is squarely in the blue range. If you by "100% blue" you mean a pigment or mixture that reflects only a single wavelength of light, again, never in human history has any such thing been seen nor will it likely ever be, so it's not worth worrying about.

Einion
03-15-2004, 03:51 PM
Yes, ultramarine blue activates the green cone quite a lot according to this diagram - (not to mention red!). A 100% blue hue would look relatively violet compared to that hue.
Why do you think it would look more violet-biased?

Also, I think it would help to think of colour not in percentages for something like this, but rather in terms of specific hue descriptors, or better yet hue angles, because in paint the issue is irrelevant - as has been pointed out a number of times previously there are no actual reflective colours that are monospectral so in a very real way there is no such thing as a "100% blue" so as Marc says in his last post it's not worth worrying about :)

That's what I said, yes. The saturation isn't full, but the hue can be.
The implied "full hue" here is basically the percentage idea put in different terms and doesn't appear to be helping you understand things.

I'm inclined to think that the reason that mix produces a green hue is because the blue was not really 100% blue in hue. Correct me if I'm wrong...
A number of the posts above explain in various ways why a mix of blue and yellow paints make greens.

I presume you are trying to get towards an understanding of actual subtractive colour interactions (i.e. paint mixing)?

If what you say is true (that a not fully saturated, but a blue that's 100% blue in terms of hue) still makes green when mixed with yellow, then this leads me to believe that there is some complex chemical reaction going on which 'distorts' the outcome of the mixed hue.
Why does it require a complex chemical reaction? You've seen what the reflectance curves of blue paints look like and they all make greens when mixed with yellows right? So any blue colour mixed or created to fall in a specific hue/chroma/value position will also mix a green.

On the basic level it's quite easy to imagine what's happening inside paint films in simple two-colour mixes. The reflectance that gives us the perception of blue and yellow in each paint cancel each other out to a given degree, largely leaving green reflectance related to the amount that both paints had to begin with. So a green-biased blue and a green-biased yellow - PB15:3 and PY3 for example - mix good greens because they both have quite a bit of green reflectance and less 'pollution' from other parts of the spectrum (particularly little violet in this case).

In a mix of a violet-biased blue and an orange-biased yellow - PB29 and PY83 say - the green reflectance to begin with is lower and there is a lot more magenta reflected light (red light + violet light) so the mix is consequently duller. There are other ways of thinking about it but they must have a realistic starting point and follow certain principles related to paints as they exist in the real world if they're to be useful.

Hope it helps,
Einion

Patrick1
03-15-2004, 04:52 PM
Yes, ultramarine blue activates the green cone quite a lot according to this diagram - (not to mention red!). A 100% blue hue would look relatively violet compared to that hue.

Why do you think it would look more violet-biased?

Here's the cone response for ultra. blue:

http://www.handprint.com/HP/WCL/IMG/coneblu.gif

If the G cone were stimulated less (let's say to the same level of the R cone), I also think that it would look more violet-ish in hue...I can't see why it wouldn't.

I did a test by glazing different blues & violets over a stripe of orangish yellow (looks like PY 73). I turned out I'd get a dark grey using a blue-violet colour...but I wouldn't call it either blue or violet. Einion, I rememebr before you said something like that the printing industry is starting to call this colour blue-violet, rather than just 'blue'. If this is the case, I think it's a very good change.

Marc Sabatella
03-16-2004, 12:39 AM
Here's the cone response for ultra. blue:

http://www.handprint.com/HP/WCL/IMG/coneblu.gif

If the G cone were stimulated less (let's say to the same level of the R cone), I also think that it would look more violet-ish in hue...I can't see why it wouldn't.


For one thing, you're assuming the green cone response is due to the green wavelengths only. It probably isn't - the red wavelengths are also likely contributing to the green cone response.

Second, while it is almost certainly true that reducing the green reflectance would make the color look more violet, that would not be a "true blue" either - you'd have to also eliminate the red reflectance. And doing so would quickly eliminate any sense of the color being violet at all.

Richard Saylor
03-16-2004, 11:09 AM
Second, while it is almost certainly true that reducing the green reflectance would make the color look more violet, that would not be a "true blue" either - you'd have to also eliminate the red reflectance. And doing so would quickly eliminate any sense of the color being violet at all.
This is a little off-topic, but that last sentence has raised a question in my mind. We do perceive spectral violet as violet, don't we? Spectral violet is at the opposite end of the visible spectrum from red. I wonder how that works. Do extremely short wavelengths stimulate the red cones in a special way?

Marc Sabatella
03-16-2004, 06:01 PM
This is a little off-topic, but that last sentence has raised a question in my mind. We do perceive spectral violet as violet, don't we? Spectral violet is at the opposite end of the visible spectrum from red. I wonder how that works. Do extremely short wavelengths stimulate the red cones in a special way?

Not according to anything I've read, which does make you wonder how it is that we perceive it as similar to the color you get by mixing a little red into a less violet blue. However, spectral violet is still pretty blue - and also pretty dark. There is also, of course, a whole range of redder violets that simply don't exist in the spectrum.

Richard Saylor
03-16-2004, 06:44 PM
Not according to anything I've read, which does make you wonder how it is that we perceive it as similar to the color you get by mixing a little red into a less violet blue. However, spectral violet is still pretty blue - and also pretty dark. There is also, of course, a whole range of redder violets that simply don't exist in the spectrum.
From a physics standpoint, that seems to be a weird range of colors: reflectance from both ends of the visible spectrun. They are beautiful and useful, however. I mean, I'm a magenta guy.

Twinbee
03-17-2004, 04:32 AM
Depends on what you mean by "100% blue in terms of hue". If you mean, one that simply *looks* as blue as can be, then no fancy chemical reaction is needed to explain why this makes green when mixed with yellow - it is simply yet another example of the basic subtractive mixing principles we all know and love. Because our eyes have never seen a single-wavelength blue paint, and interpret colors with quite a bit of green and vioelt in them as still being a blue as can be if some sort of average wavelength is squarely in the blue range. If you by "100% blue" you mean a pigment or mixture that reflects only a single wavelength of light, again, never in human history has any such thing been seen nor will it likely ever be, so it's not worth worrying about.
When I say "100% blue in terms of hue", I mean that the hue is at 240 degrees. This could be anything from saturated blue to white-ish blue to white. We know that paints can't achieve 100% saturated blue whilst maintaining the exact 240 hue degree. However, one can achieve a pure blue hue in paint if the saturation is lower (white-ish).

So, imagine the blue eye cone response is at 100%, and the green and red eye cone response is anything from 0 to 100%, but where they must equal each other (so for example: red:20%, green:20%, blue:100% .....or..... red:50%, green:50%, blue:100%). Now if that 'blue' were to mix with a true yellow hue (60 degrees), in either additive or 'subtractive' mixing, then green would not result - you'd get a hueless colour.



Why does it require a complex chemical reaction? You've seen what the reflectance curves of blue paints look like and they all make greens when mixed with yellows right? So any blue colour mixed or created to fall in a specific hue/chroma/value position will also mix a green.

All those supposedly blue paints weren't at 240 degrees, they were more cyanish/greenish - something like 220 or 210 degrees.




Here's the cone response for ultra. blue:

http://www.handprint.com/HP/WCL/IMG/coneblu.gif

If the G cone were stimulated less (let's say to the same level of the R cone), I also think that it would look more violet-ish in hue...I can't see why it wouldn't.

For one thing, you're assuming the green cone response is due to the green wavelengths only. It probably isn't - the red wavelengths are also likely contributing to the green cone response.

I think those bars in the GIF picture represent the eye cone response levels directly (rather than specific wavelengths which have to go through further conversion).



This is a little off-topic, but that last sentence has raised a question in my mind. We do perceive spectral violet as violet, don't we? Spectral violet is at the opposite end of the visible spectrum from red. I wonder how that works. Do extremely short wavelengths stimulate the red cones in a special way?

It's a very interesting question, and I'd love to know the answer. In fact, I posted the same question on the 4th page in this very thread. I've searched Google before, but this time, I decided to give it a more thorough look.

Just one page had the info I was looking for...
http://howthingswork.virginia.edu/supplements/paint.pdf

The crucial quote:

The response of the red cone cells to short wavelength light allows you to perceive violet light.

Remember all those cone response diagrams I linked to earlier (on page 2) ? Well, it looks like this PDF document has a much more accurate one, this time where the red 'curve' gets two helpings! You were right :-) And it's something I've suspected for some time too - nice to see it confirmed from that PDF doc.

I've grabbed the picture from the PDF and uploaded it to here:
http://www.skytopia.com/stuff/rgbcone.jpg

Now I know that the other 6 pictures I posted were wrong ;-)

Marc Sabatella
03-17-2004, 12:23 PM
When I say "100% blue in terms of hue", I mean that the hue is at 240 degrees.


So, you are referring to the *apparent* hue, meaning the color may well still be reflecting some green adn some violet.


So, imagine the blue eye cone response is at 100%, and the green and red eye cone response is anything from 0 to 100%, but where they must equal each other (so for example: red:20%, green:20%, blue:100% .....or..... red:50%, green:50%, blue:100%).
[/QOUTE]

I wouldn't assume the above described color would produce this cone response at all. Even a pure spectral blue is going to trigger green more than red as far as I know. And if the blue in question is reflecting a little violet as well as a little green, so the perceived hue is the same as the spectral blue, that's definitely going to trigger green more than red.

[QUOTE]
Now if that 'blue' were to mix with a true yellow hue (60 degrees), in either additive or 'subtractive' mixing, then green would not result - you'd get a hueless colour.


Again, there is simply no reason to say this. The color you are describing *is* reflecting green if it is exists in the real world, and any similar yellow paint (that *appears* to be the same hue as what you are calling true yellow) is also reflecting green. So the mixture would look green. The only to mix a blue and yellow and have them not produce a green is if neither of the component colors reflects any green whasoever, and this is just plain impossible in the real world. No pigment or combination of pigments will produce anything remotely like that.


All those supposedly blue paints weren't at 240 degrees, they were more cyanish/greenish - something like 220 or 210 degrees.


Of course, there may well be pigments out there somewhere that really do appear to be the hue you are describing - but I can absolutely guarantee they will still reflect enough green to produce green when mixed any yellow out there you might find that appears to be exactly the hue you want. note that ultramarine is even further toward violet than "true" blue, but even it produces green.

Einion
03-18-2004, 04:33 PM
When I say "100% blue in terms of hue", I mean that the hue is at 240 degrees... a true yellow hue (60 degrees)...
Now if that 'blue' were to mix with a true yellow hue (60 degrees), in either additive or 'subtractive' mixing, then green would not result - you'd get a hueless colour.
Okay, this is a good starting point - blue at 240° and yellow at 60°. However, you can't draw parallels with many subtractive and additive mixing examples, plus remember that light of this hue is not going to be exactly the same as the light reflected from paint of that hue (or any coloured surface for that matter) because of metamerism. A paint's hue is effectively an average of its total reflectance, it is all too easy to forget that it's reflecting all this other light 'unrelated' to the apparent hue before beginning to imagine what happens when it's mixed with another colour where the same is equally true.

Yes a mix of blue and yellow light gives us an achromatic colour - because it amounts to B + (R + G) which we know makes an analogue of white light - but you cannot extrapolate from this the outcome of a subtractive mix with colours of the same hues. Additive and subtractive principles should not be mixed up and expected to predict results.

All those supposedly blue paints weren't at 240 degrees, they were more cyanish/greenish - something like 220 or 210 degrees.
First, I referred to blues as a whole so how can they all be closer to cyan?! Second, you said, "one can achieve a pure blue hue in paint if the saturation is lower (white-ish)" in your opening paragraph. Given that this is true then obviously what you think to be a logical follow-on is shown to be incorrect since we all know that mixtures of ANY blue and ANY yellow paint will make green.

We keep on going over the same territory time and again so just try it for yourself: mix a blue to the hue you want (as I said before, it's easy) and then mix it with any of the number of yellow paints at roughly 60°... you'll get a green I assure you. Simple, end of discussion. Any argument you come up with that tries to explain why one doesn't get the result you're saying we should, "oh but the hue wasn't exactly right" or, "it was a mix so it wasn't 'pure' enough" or, "the saturation was too low" are just sophistries and you need to recognise them as such :)


Again, there is simply no reason to say this. The color you are describing *is* reflecting green if it is exists in the real world, and any similar yellow paint (that *appears* to be the same hue as what you are calling true yellow) is also reflecting green. So the mixture would look green....
No pigment or combination of pigments will produce anything remotely like that.
Right on http://www.wetcanvas.com/Community/images/20-Aug-2003/3842-thumbsup.gif

As you said, the only way to mix a blue and yellow and not get green is for neither of the starting colors to reflect any green whatsoever, which is of course not possible in the real world. But even more importantly, the green light is necessary for the perception of something as yellow because of the sensitivity of the green cone in this area. Drop out the peak green response and the colour isn't seen as yellow any more so we're no longer talking about what we started talking about! :D

Of course, there may well be pigments out there somewhere that really do appear to be the hue you are describing - but I can absolutely guarantee they will still reflect enough green to produce green when mixed any yellow out there you might find that appears to be exactly the hue you want. note that ultramarine is even further toward violet than "true" blue, but even it produces green.
Again right on.


Twinbee, we've now had ample explanation of why you won't get a grey or black in terms of the underlying science, certainly more than enough to sway any reasonable person don't you think?

Einion

Twinbee
03-19-2004, 05:14 AM
Well, sorry to say, I'm still 90% sure that 240 degree blue and 60 degree yellow will make a hueless colour in additive or 'subtractive' mixing (at least ignoring any potential complex chemical reaction). I'm willing to be proved wrong, but so far I remain unconvinced. The fact that two of you seem so sure brought it down from 99% certainty ;-)




So, imagine the blue eye cone response is at 100%, and the green and red eye cone response is anything from 0 to 100%, but where they must equal each other (so for example: red:20%, green:20%, blue:100% .....or..... red:50%, green:50%, blue:100%).
I wouldn't assume the above described color would produce this cone response at all. Even a pure spectral blue is going to trigger green more than red as far as I know.
The percentages that I gave were intended to be the actual cone responses, rather than a point on the visible spectrum, or wavelengths needing to go through additional conversion.



Again, there is simply no reason to say this. The color you are describing *is* reflecting green if it is exists in the real world,
No more than it's reflecting red. A semi-saturated blue (white-ish blue) will send a response to the green eye cone - agreed. But if it's really at 240 degrees, it will also send an equal response to the red cone. So there's no more reason to believe blue and yellow will make green any more than it will make red. This is of course assuming that the blue hue really is at 240 degrees.


and any similar yellow paint (that *appears* to be the same hue as what you are calling true yellow) is also reflecting green.
Again, a 100% yellow hue will not reflect any more green than it does red.

Here's another way of looking at it. We know that blue can be created by using cyan and magenta in the 'subtractive' process. If we mix cyan, magenta and yellow together - guess what? We get a hueless colour.

To summarize, the blue (or yellow) hue would need to be biased towards green at the start to create a slightly greenish hue.



note that ultramarine is even further toward violet than "true" blue, but even it produces green.

Even ultramarine blue is only about 225 degrees in hue.



Yes a mix of blue and yellow light gives us an achromatic colour - because it amounts to B + (R + G) which we know makes an analogue of white light - but you cannot extrapolate from this the outcome of a subtractive mix with colours of the same hues.
With inks (such as the professional ink process), blue (which is cyan and magenta) and yellow will definitely make a hueless colour (near black in this case). I'm 90% convinced the same thing happens with paints too. Think of Y + (C + M).



We keep on going over the same territory time and again so just try it for yourself: mix a blue to the hue you want (as I said before, it's easy) and then mix it with any of the number of yellow paints at roughly 60°... you'll get a green I assure you. Simple, end of discussion.

Not quite :) Theoretically at least, it should make a hueless colour. I would like to try and test out that mix, but I can't trust my eyes (and haven't got the measuring equipment to confirm the blue hue would be at precisely 240 degrees, or that a yellow is at precisely 60 degrees).



As you said, the only way to mix a blue and yellow and not get green is for neither of the starting colors to reflect any green whatsoever,

I didn't quite say that. They can reflect green /and/ an equal amount of red too, and thus the resulting mix (of blue and yellow) will be a hue at a point from the continuum of blue to hueless to yellow.



Drop out the peak green response and the colour isn't seen as yellow any more so we're no longer talking about what we started talking about!
I recall reading somewhere that to make what we see as 'white', more green is needed than red, and more red than blue. I assume these are the required wavelengths in terms of power to stimulate the cones to equal levels. However, in all my posts, I continually refer to the cone response rather than the wavelength power needed to produce the cone response. Perhaps this is where some of the confusion may lie?

Marc Sabatella
03-19-2004, 01:22 PM
A possibly major key point buried in here that I thought worth repeating at top.


Even ultramarine blue is only about 225 degrees in hue.


Here, then, you're going to have to define what you mean by this. I've been taking these degree figures as a semi-useful abstraction to gauge where the colors appears to be on the color wheel, and so I haven't questioned the specific numbers. I simply assumed when you equated 240 degrees with "true blue", you were referring to a color similar to what the world knows as blue. But if you are claiming that ultramarine is *greener* than the color you are describing, you're obviously working from some sort of model that is completely unfamiliar to me. It seems that what you are calling "240 degrees" would actually be perceived as a violet, not a blue. I'm not even sure if it would be a spectral violet or not - might be too red for that. And indeed, if you've biased the color so far toward red that it does reflect a significant amount of red wavelengths, that would affect the results. No one is claiming that mixing a violet with yellow should produce a green. We were talking about blue, which is normally defined as a color somewhat greener than ultramarine.


The percentages that I gave were intended to be the actual cone responses, rather than a point on the visible spectrum, or wavelengths needing to go through additional conversion.


I realize that, and that's why I said you shouldn't assume the color in question would actually produce those cone responses. It almost certainly would not, as described below.


No more than it's reflecting red. A semi-saturated blue (white-ish blue) will send a response to the green eye cone - agreed. But if it's really at 240 degrees, it will also send an equal response to the red cone.


Anything that we perceive as blue is almost certainly going to reflect more green than it does red. Note, though, it isn't clear what you mean by "send a response". The paint sends *light*; it is the cone that determines the response. Obviously, the paint sends the same light to all three cones. But the different cones will respond to those wavelengths different. And in particular, because the G cone is more responsive than the R cone to the blue light light as well as the green light reflected by the paint in question, you'll get more of a response from the G cone than the R cone from those wavelengths. Depending on whose chart you believe, you might get more of a response from the R cone from the violet wavelengths reflected by this paint, but there is no reason to assume this would cancel out the G response from the green and blue wavelengths.

Now, if the paint actually does reflect a significant amount of red wavelengths - something we wouldn't assume a blue paint would do, since red is on the opposite end of the spectrum from blue - this would of course trigger the R cone. But it would trigger the G cone pretty well too. It's anyone's guess as to how all this would play out in terms of which cone would actually give the higher response. Which is why I said there is no reason to assume the R and G responses would be equal - there are far too many variables here. It certainly seems likely the G cone would be triggered more than the R cones if we are talking about an actual blue, not a color so violet that it reflects a lot of red.


So there's no more reason to believe blue and yellow will make green any more than it will make red. This is of course assuming that the blue hue really is at 240 degrees.


If this color is actually violet and not blue, I agree completely.


Again, a 100% yellow hue will not reflect any more green than it does red.


This much is true as well. However, since real blues blue don't reflect nearly as much red as they does green, that is why a mixture of the two will result in a color that is perceived as green. Whereas if you mix that yellow with a color that reflects more red than green - say, a red violet, you'll indeed get a red out of it.


Here's another way of looking at it. We know that blue can be created by using cyan and magenta in the 'subtractive' process. If we mix cyan, magenta and yellow together - guess what? We get a hueless colour.


Not in the real world you don't, at least not in with ratios of cyan to magenta that produce something most people would identify as blue. The only reason this works in the theoretical realm is that "magenta" is defined as a color with no green reflectance whatsover, so the blue it creates with cyan has in theory no green reflectance either. This won't happen in real life, ever - the magenta will reflect *some* green, as will the blue it creates when mixed with cyan. When mixed with yellow, it is the green reflectance that will dominate, and it will be more than enough to obviously create a green. You can make this hueless only by pumping up the magenta content enough that the reflectances start to even out.

However, this is definitely on the tright track - by mixing magenta and cyan, you are likely to get a blue that reflects much more red than a blue pigment would, although in order for it to still read as blue and not violet, there is going to have to be proportionately more green reflectance than normal blue paints would reflect.


With inks (such as the professional ink process), blue (which is cyan and magenta) and yellow will definitely make a hueless colour (near black in this case).


I would bet that would depend entirely on the actual ratios of these colors. As with paints, if you first mixed the blue so that it actually looked blue and not like the red violet color you seem to be describing as "240 degrees", and then added yellow, you'll get green.


I would like to try and test out that mix, but I can't trust my eyes (and haven't got the measuring equipment to confirm the blue hue would be at precisely 240 degrees, or that a yellow is at precisely 60 degrees).


Again, I've been assuming we were talking about blue, because that is the word you used to describe the color, but if "240 degrees" actually describes a color more violet than ultramarine, we are talking different languages. It is indeed possible the color you are describing wouldn't make much of a green when mixed with yellow. Even if you can't trust your eyes to mix these colors exactly, presumkably you have some diea of what these colors you look like. Mix something that appears to err on the non-green side of each color.

Richard Saylor
03-19-2004, 01:43 PM
Mark, a 240º blue is as blue as it gets, consisting of blue light, with equal parts of green and red light. For example, (25,25,210) would be at 240º. A hue angle < 240º would appear more greenish, and > 240º more violet-reddish. Yellow is a narrow sector centered around 60º, which is 180º from 240º, so, theoretically, mixtures of yellow and 240º blue lie on a straight line passing through the center of the circle, which is hueless.

Marc Sabatella
03-19-2004, 04:35 PM
Mark, a 240º blue is as blue as it gets, consisting of blue light, with equal parts of green and red light.


Is this perceived as the same color as a blue light with no green or red light? If so, which wavelength of blue light are we talking about? And is ultramarine really on the green side of this? Looking at the pure blue produced by my monitor, it doesn't *look* more violet than my tube of ultramarine, but it's definitely closer than I imagined it would be.

Had I thought about it, I would have noticed that 240 is diametrically opposite 60 on the color wheel. Of course these wouldn't produce green when mixed by default. But I don't normally consider blue and yellow to be complements, either, although I realize Munsell defines them this way. To what extent is the spacing of these colors on the color wheel subjective?

Anyhow, if this is all true, and ultramarine really is greener than true blue, it can't be by much. It wouldn't take much added red to make it appear a pure blue. And I'd still bet that there is more green reflectance in the resultant color than red, and that the result would look green. But maybe not. So perhaps it really does hinge on this definition of what exactly true blue is.

Richard Saylor
03-19-2004, 04:48 PM
Is this perceived as the same color as a blue light with no green or red light? If so, which wavelength of blue light are we talking about? And is ultramarine really on the green side of this? Looking at the pure blue produced by my monitor, it doesn't *look* more violet than my tube of ultramarine, but it's definitely closer than I imagined it would be.

If my source is correct, 240º is the pure blue produced by the monitor. Starting with red at 0º, the light primaries and secondaries are spaced at 60º intervals around the circle. http://www.zianet.com/jpierce/Panel.html
To my eyes, ultramarine does not look as if it lies on the green side of 240º.

Twinbee
03-19-2004, 05:31 PM
Just Googled Ultramarine blue to see if I could find any actual hue value for Ultramarine Blue.

The first site I found was here:
http://tx4.us/mh/mh227.htm
However, it's called "French Ultramarine Blue", so I'm not quite sure it's 100% the same. It gives a value of 227.7 degrees.

The second site is none other than the handprint.com site. For Ultramarine Blue, it gives a range of hues from 282 to 294 degrees. As the hue rotation starts at magenta (instead of what we were using - red), you minus 60 degrees, so this actually translates to 222 to 234 degrees - averaging at 228, so I wasn't far off when I said 225 :)
http://www.handprint.com/HP/WCL/waterb.html

And of course Domer gave that diagram showing that Ultramarine Blue was biased towards green:
http://www.handprint.com/HP/WCL/IMG/coneblu.gif

Marc Sabatella
03-19-2004, 09:12 PM
However, it's called "French Ultramarine Blue", so I'm not quite sure it's 100% the same.


It is. Virtually all ultramarines you can buy are the synthetic imitation of the real lapis lazuli, which is supposedly chemically identical, although one or two high end manufacturers still use the original.

OK, so all these sources seem in agreement in defining blue to be a color more violet than ultramarine. I was obviously talking about a different color also popularly known as blue, but the question of how this more-violet-than-ultramarine-blue would mix with yellow is still valid.

If you actually pushed it toward violet with as much red would be necessary to get it to 240, then, the red *might* indeed cancel out the green. But even so, I wouldn't take it as a given, because green really is right next to blue on the spectrum, but red isn't - we just pretend it they are closer than they really are by joining the opposite ends of the spectrum to form a circle. It still seems entirely likely to me that you could get this 240 degree color that reflected more green than red if it reflected enough of the violets just to the left of blue but little of the actual reds. The reflected violet wouldn't have nearly as much of an effect on how the mix looked as the green would, since the green would be reflected significantly by the yellow but the violet might not be. But I don't know just how far toward the end of the spectrum 240 really is. If it's basically at the end of the visible spectrum, then any green reflectance would indeed have to be balanced by actual red reflectance.


And of course Domer gave that diagram showing that Ultramarine Blue was biased towards green:
http://www.handprint.com/HP/WCL/IMG/coneblu.gif

Again, since these are cone responses, they don't really show any such bias. They may simply show what we already know to be true: that the G cones are more responsive to the blue wavelengths and the greenish wavelengths than R cones are, and neither cone responds particularly well to the more violet wavelengths. I suspect you'd get the more or less same results out of a monospectral blue, or a color that had a very narrow centered on this blue, simply because what's to the left of blue on the spectrum is *not* red - it's violet. Red is way over on the other side of spectrum, so it shouldn't be surprising we don't get get much of an R cone resposne from a blue.

Patrick1
03-19-2004, 11:06 PM
According to these diagrams from Handprint, the far (violet) end of the visible spectrum does stimulate the R cones more than G:

http://www.handprint.com/HP/WCL/IMG/conesens1.gif
http://www.handprint.com/HP/WCL/IMG/RC/rcPB29.jpg

I suspect that this is why violet wavelengths look more reddish than blue does, even though in a linear sense it's farther away from red. But this is violet-blue or violet, not blue.

And an important question: How does one define the hue of true blue?

-no apparent hue bias towards green or red?
-equal stimulation of R & G cones, and higher stimulation of B cones?
-0, 0, 255 or 51, 51, 255 (etc.) on a colour monitor?
-a paint which seems to mix greens and violets equally well?
-the average person's personal idea of true blue?

drollere
03-20-2004, 03:02 AM
i feel a little like the guy who tuned into the superbowl *after* the halftime show. even so, please indulge a rant.

the kubelka-munk "theory" is just a mathematical model for predicting paint mixture color if the reflectance profiles and light scattering behavior of pigment particles are known for dispersions in a given vehicle at a specific paint thickness. it is basically a mechanism to extrapolate from that baseline case to layers of different thicknesses or dispersions of different pigment concentrations, and little else. it rests on several important simplifications or restrictions of the problem that quite often make its predictions unrealistic in actual cases. for example, it assumes a paint "layer" exists (which does not hold in watercolors), and that the pigment particles are randomly and evenly dispersed in the layer (which doesn't hold for glazing or canvas rather than palette mixtures); and that the pigment particles have a constant size and homogenous refractive properties (which they never do, especially in crystals); and so on. in fact the kubelka-munk is inaccurate enough often enough that it is rarely used to address color appearance issues in practical color engineering: ask any automotive paint chemist.

on the other hand, the geometric mean of two pigments, though not a sufficient method for determining a mixture reflectance profile from individual pigment reflectance profiles, is a useful ("quick and dirty") method to understand how two specific pigments will combine to reflect or absorb different parts of the spectrum. (for example, it can help you understand why chromium oxide green mixes so well with warm paints, and so poorly with the cobalts.)

the fundamental contrast in its sharpest form: believing you can know *abstractly* what is going on with color in every possible situation, using mathematical or logical or geometrical models; and knowing how to address a color problem *perceptually* in a specific, practical situation. i'd guess 99% of what is written about color is not useable or accurate because that basic distinction isn't respected.

for example: skalka wrote, "Do you think that any manufacturer would take thousands of pounds of cadmium yellow pigment at $50.00 per pound and mix it with another pigment at nearly the same price to make a plastic cup and not know without a doubt the color what the final product's color will be?" no, i don't think that ... because what they do in practice is, they mix up a one ounce color sample in a single prototype cup, and see how it looks, and if they like it *then* they mix the large batches -- and they retest the large batches to make sure the color stays the way they want it.

i subscribe to the radical view that as of today we understand everything important there is to understand about color, and that nearly all of that understanding can be boiled down to a single rule: "color depends .. on the colormaking materials and the color viewing observer in the color viewing situation." no single theory, no mathematical model, can do it justice. but often *any* theory is better than sheer ignorance. you just have to learn when the cure, in large enough doses, becomes poisonous.

artists use color the way carpenters use lumber -- they cut to fit. to the limited extent that abstractions can guide experiential learning, "color theory" is like wittgenstein's ladder: you use it to climb up to a new level of awareness, and once you've attained that awareness you no longer need the ladder.

twinbee wrote me a few months ago with the "multiplicative" concept and his "binary" logic, and it strikes as a good example of how "color questions" can be projected the domain of pure logic, or pure geometry, or pure paint, or some other "pure" or "ideal" case, where abstractions can be talked about with purity and logic -- in hopes of knowing *abstractly* what is going on with color in every possible situation.

two problems: "purity" usually means there aren't any *practical* issues driving the discussion (like, "how do i paint better?"); and "purity" or "logic" or "geometry" loses sight of the fact that color is an *animal experience,* that is, it's the kind of thing we "do" as organisms to answer specific issues related to the world we live in. what has "purity" got to do with biology?

subtractive mixtures *nearly always* reduce the amount of light reaching the eye, while additive mixtures *nearly always* increase the light, and that contrast is all that is intended by the terms "subtractive" and "additive." yet twinbee insists on interpreting a broadly qualitative contrast as denoting narrowly quantitative arithmetical operations ("you'd end up with negative numbers"). what is the specific problem the nomenclatural innovation is trying to solve?

there is a world of difference between treating the "ideal" case in physics (an artillery shell with or without air drag) and the "ideal" case in color. the problem: there *is no* ideal realm of color. in most cases, an "ideal" case in physics really can apply in some situations (space travel, for example). but what is a "pure" blue or an "ideal" green supposed to refer to?

perhaps the "primary" colors in standard colorimetry give us a clue: the primaries are invisible! this is simply done as a mathematical convenience, so that the third "primary" expresses the luminance of the color, and any visible hue/chroma combination can be expressed as two terms. but this is the problem: you can't use some colors, including "primary" colors, to explain other colors. it's a circular argument, begging the question. ("what creates colors?" "why, primary colors." "what creates primary colors?" "why, they're colors that don't have a cause." "where do they come from if they don't have a cause?" "er, well ... ") again, the difference between the abstraction and the perception can be huge. talking about abstractions as if they were simply the most pure form of perception is, well ... platonism. and plato hated painters.

the concept of a blue or green paint "polluted by other colors" is one of those tiny bits of color logic that is packed with a long history of misinformation. the essential misconceptions can be found in michel-eugene chevreul's book on color contrast, but actually date from the 18th century, for example:
* the "color" is in the light (yellow color equals "yellow" wavelengths, blue color equals "blue" wavelengths, etc.)
* all colors arise from "primary" colors
* "primary" colors are the colors you can't mix from other colors
* "impure" "primary" colors are the reason for all the problems in subtractive color mixing
* "impure" colored pigments are the reason why we can't mix the bright colors found in nature (for example, the colors of gems)
* we can at least partially solve the problem of "impure" pigments by mixing two "primary" pigments that are "polluted" with each other (a yellowish red mixed with a reddish yellow makes the brightest orange)
* "primary" colors exist outside pigments or light
* "primary" colors outside pigments or lights are still *real* -- in the sense that laws of nature or principles of euclidean geometry are "real".

not only is every one of these "color theory" dogmas false as intended, they are in some cases false in every context in which they are used. but that's another discussion.

my guess is that cadmium orange tints to pink rather than yellow because the smaller pigment particles have a redder hue than the larger particles, and these smaller particles have a higher tinting strength than the large particles, so they dominate in a mixture with an opaque white paint. right or wrong, my explanation rests on a specific physical mechanism, in response to an artist who was *looking at how paint behaves*. it seems to me that is the only way artists make any progress: look for the effect, and find the mechanism that made it happen.

if theory can't lead that process along, or follow the facts very closely, it had better get out of the way.

Richard Saylor
03-20-2004, 11:49 AM
Maybe we shouldn't be too dogmatic about yellow and blue always making green. Maybe this is true for all the single pigment blues which we know about, but there most certainly can be blue hues which combine with yellow to make what would appear to be a neutral grey. I do it all the time. I take Pthalo Blue and mix in Quinacridone Violet until I get a hue which certainly appears to be blue. (It is no more violet-looking than Ultramarine Blue.) Then I can make a neutral grey by adding a little Hansa Yellow. (If I add so much Q. Violet to the Pthalo Blue that it looks violet, then adding yellow will tend to make brown, not grey.) Of course, there is no mystery about this, since combining the same colors (and proportions) in different orders will show that what I'm really doing is equivalent to mixing orange and blue or red and green.

Einion
03-20-2004, 01:18 PM
I'm willing to be proved wrong, but so far I remain unconvinced.
You have been and if you actually tried things practically the paint would prove it to you!

This discussion is beginning to remind me of one Patrick and I had with a member called Shimo, who had some equally idiosyncratic views on colour that had little bearing on practical paint experience, and if I remember correctly who also based his notions on the idea of 'pure' colour...

So there's no more reason to believe blue and yellow will make green any more than it will make red. This is of course assuming that the blue hue really is at 240 degrees.
You're consistently ignoring this isn't merely an intellectual exercise, it is subtractive paint mixing we're concerned with (or should be). What happens to the 'other' light in the mix? You can take an orange-yellow like Diarylide Yellow, mix it with a violet-biased blue and you still get green.

I'm going to knock this on the head for you once and for all. Go and buy a cheap tube of a yellow acrylic (pick the pigment by referring to Handprint's suggestions for primary yellow) and a tube of Ultramarine Violet - a colour obviously far more violet than anything one would consider blue. Now mix them together in steps and tell us what you get. Since both paints are single-pigment colours this simplifies any discussion about possible causes since there's no possibility of one pigment in a mixed starting colour dominating at one concentration and not at another.

I would like to try and test out that mix, but I can't trust my eyes (and haven't got the measuring equipment to confirm the blue hue would be at precisely 240 degrees, or that a yellow is at precisely 60 degrees).
Didn't I say it would be a sophistry? I am continually getting the impression this is only an intellectual diversion for you - please tell us now whether you're concerned only with the theoretical side of things or in practical applications. If the latter then concentrate on it and forget the theory for the time being - go and push some paint around, for your own sake if not for ours. If on the other hand you're just interested in debating the issues ad nauseam and not actually applying anything then I'm going to request to the moderators that this thread is moved to Debates where it should be.

With inks (such as the professional ink process), blue (which is cyan and magenta) and yellow will definitely make a hueless colour (near black in this case). I'm 90% convinced the same thing happens with paints too. Think of Y + (C + M).
I work in the graphics industry so thanks for bringing this up. First, CMY never gives a colour as achromatic as it's supposed to for various reasons (and a good example of why the black is needed). Second, this is a glazing effect and not directly applicable to physically mixing the colours together. And third, you can't model actual pigment interactions purely theoretically. Buy some process colours (System 3 acrylics have a decent set and they're cheap) and play with them.

I didn't quite say that. They can reflect green /and/ an equal amount of red too, and thus the resulting mix (of blue and yellow) will be a hue at a point from the continuum of blue to hueless to yellow.
I was commenting on a quote from Marc as you can see. And as for the next part you're forgetting again that mixing lines aren't always straight, especially for mixed greens. You must have seen how far off they can be from the theoretical to the practical.

Einion

Marc Sabatella
03-20-2004, 03:33 PM
And an important question: How does one define the hue of true blue?

-no apparent hue bias towards green or red?
-equal stimulation of R & G cones, and higher stimulation of B cones?
-0, 0, 255 or 51, 51, 255 (etc.) on a colour monitor?
-a paint which seems to mix greens and violets equally well?
-the average person's personal idea of true blue?

Well, that's definitely the issue now, as I see it. Both Einion and I have clearly been thinking of a blue that is less apparently violet than ultramarine, so now it seems especially relevant to consider how we decide what blue really is.

As near as I can figure it, the notion of a 240 degree blue comes about as a direct result of how the additive/light primaries are chosen. I say this because it is those colors that then define magenta, cyan, and yellow as the subtractive primaries, and the combination of these two is how the color wheel in which blue is placed at 240 degrees comes about.

The surprise to me is that there is nothing particularly magic about the wavelength of light chosen to represent the additive primaries. Other wavelengths could apparently work as well. According to my understanding from the Handprint site, the wavelengths typically chosen is selected because they are the wavelengths with the maximum *difference* in cone responses. In particular, additive primary blue is defined as the wavelength at which the difference between the B response and the R & G responses is as great as it can be. As it happens, defining your primaries this way doesn't really work - they don't add up to white. So you have to bias one of your colors a bit. But I guess it's usually the red that is fudged here, not the blue, according to handprint.

At the wavelength chosen to represent additive primary blue, it definitely does appear the G cones are triggered more than the R cones, which is consistent with what I suggested earlier - that a "pure" blue would likely trigger a higher G response than R, so seeing a cone response chart with more G than R doesn't actually demonstrate a "green bias" in the color itself.

This wavelength is not actually the one with the highest B cone response - the wavelength with the highest B cone response has quite a bit more G & R (and again, more G than R) so the difference isn't maximal. But if we look at the wavelength with maximal B response, I note it is more toward the green side than the wavelength normally chosen to represent the additive B.

I have no evidence to back this up, but my guess it, this is the color closer to what most people identify as blue, and the color normally chosen to represent blue appears more violet in comparison. Ultramarine probably sits somewhere between these two colors, which is why I can look at it and say it *looks* to be on the violet side of what i consider a true blue, but a "real" color theorist can say "no, it's greener than the color I have chosen to represent additive primary blue". There is probably no rationale basis to resolve this discrepancy in defining blue.

It can also still make sense to discuss what would happen if a color the hue of this additive primary blue were mixed with a yellow. I think what I speculated before may still hold - a color that *appeared* to be at this hue would do so because it reflected that wavelength and the surrounding violet and green wavelengths, but not necessarily enough actual *red* wavelengths from the opposite end of the spectrum to prevent the mixture with yellow from being perceived as green. But it would probably depend on what pigments were used in creating that blue - presumably, a color that *did* reflect a significant amount of red could also appear to be the same hue as that additive primary blue if it reflected green and violet wavelengths in the right proportions.

Anyhow, I now see this whole discussion in a completely different light. It isn't really about Twinbee assuming all colors should behave in subtractive mixing as if they were monospectral, but rather, about the particular issue of how far apart yellow and blue are on the color wheel. Clearly, if one defines the colors and the color wheel in a way that makes them truly complementary, of course one is going to assume the mixture would lack hue.

drollere
03-20-2004, 05:26 PM
ok, let's shift gears for a moment and talk about a "true" blue or a "pure" blue.

anytime a color term ("blue," "red," "punk ass yellow") is modified by a moral, epistemological or evaluative adjective ("pure blue," "best red," "true punk ass yellow"), you know you're knee deep in color confusion, and sinking fast.

what does "pure" mean, anyway?

the unique blue can be found by asking people to cancel many monospectral colors to a pure white (gray) by adding the three remaining unique hues -- for blue, these would be red, yellow, green -- and in that situation "pure blue" is just the monospectral wavelength that requires only the addition of the yellow light to neutralize.

wait, the yellow light? *which* yellow light? the yellow at 570nm, or 575nm, or maybe 580nm, or 567nm? well, you're the boss, you get to choose. which one do you want?

but if you choose 570nm, you'll get a different "pure" blue than if you chose 575nm. so why did you choose one rather than the other?

well, if it's that arbitrary, why not just show people a lot of blues and let them choose the "pure" blue they like the best? no reason at all! will that chosen blue be the same as the experimentally defined blue? nope, probably not. so which "pure blue" of the two "pure blues" you have (or three, if you used more than one yellow) is the "true" "pure blue"?

wait, i want to try that hue cancellation thing. so i do it, and my hue cancellation results are different from yours. yikes, which is the "pure blue" now? my "pure blue," your "pure blue," the "true pure blue" that was the "best" of the two or three "pure blues" you came up with?

see where this is going?

it's paints, kids, its paints and clays and dyes and glasses and lights, lots of lights, and it all bounces around in a world of stuff and things and days and nights, and tired eyes and big eyes and people trained in color and people who think "blue is the sky" and the sky is the purest pure blue there is.

where did that "pure" blue get to, anyway? it must be way up in the sky.

Richard Saylor
03-20-2004, 06:40 PM
anytime a color term ("blue," "red," "punk ass yellow") is modified by a moral, epistemological or evaluative adjective ("pure blue," "best red," "true punk ass yellow"), you know you're knee deep in color confusion, and sinking fast.
Good taste seems to be sinking fast too.
it's paints, kids
Here's one 'kid' who plans to be scarce around here, at least until things settle down. Obviously all color speculation should be left to the professionals. Maybe 'Color Theory' should be deleted as one of the forum topics.

Twinbee
03-21-2004, 06:01 AM
First off, welcome to the forum drollere. If I'm not mistaken, you're the author of the excellent Handprint.com site. I'm only new here, but I think I speak for everyone when I say it's an honour to have you posting on this forum :)

To the quickest path leading to the resolution of 'True Blue' (tm) ;), I'll start off with Domer's post. Otherwise we could be debating semantics...



a: no apparent hue bias towards green or red?
b: equal stimulation of R & G cones, and higher stimulation of B cones?
c: 0, 0, 255 or 51, 51, 255 (etc.) on a colour monitor?
d: a paint which seems to mix greens and violets equally well?
e: the average person's personal idea of true blue?

To me, B seems to be the most accurate, scientific and rounded definition of a 100% blue hue, because the eye cones are really at the root of the issue. It's the one I've been using anyway. Because of this, I would say A equals B. As far as I know, this is also the exact definition of 240 degrees.
C is very close to A and B, but of course, in the real world, even monitors will never quite reach full colour saturation (let alone combined with certain pure hues such as blue or cyan).
E will hopefully veer towards A and B over time, but I understand that it's very convenient to say even near 240 degree paints are blue, so perhaps I shouldn't say this.

And D is tricky semantically, but I'll take a shot that theoretically speaking, it would be 'greens and reds' (rather than violets) equally well.




And of course Domer gave that diagram showing that Ultramarine Blue was biased towards green:
http://www.handprint.com/HP/WCL/IMG/coneblu.gif
Again, since these are cone responses, they don't really show any such bias. They may simply show what we already know to be true: that the G cones are more responsive to the blue wavelengths and the greenish wavelengths than R cones are, and neither cone responds particularly well to the more violet wavelengths.

But aren't the eye cones at the root of the issue? If the green bar is higher than the red, then this indicates to me that there's a green bias. As far as I know, you can use any combination of wavelengths, but as long as they lead to a situation where the red and green eye cones are stimulated equally (with blue higher), then you can always mix with yellow to make a hueless colour (ignoring real world hue distortion).

Perhaps someone else can confirm this. In other words, can someone confirm that equal stimulation of red, green and blue eye cones is what anybody would immediately call pure white?



You're consistently ignoring this isn't merely an intellectual exercise, it is subtractive paint mixing we're concerned with (or should be). What happens to the 'other' light in the mix? You can take an orange-yellow like Diarylide Yellow, mix it with a violet-biased blue and you still get green.

Again, I would attribute this to 'hue distortion', as things in the real world don't always behave exactly like what the theoretical colour wheel would predict. In the first page of this thread, WFMartin said how Flake White mixed with Cadmium Orange makes a surprisingly red hue. This is another case of real-world 'hue distortion'. It's worth mentioning that cmyguy also found that a blue (very close to ultramarine) when mixed with yellow created a neutral grey, so this time, it matches what the theoretical colour wheel would predict.

Marc made an interesting point how blue was almost right next to green in the visible spectrum (compared to red which is far away). What I am willing to consider is that the amount of 'hue distortion' (when it does occur) is apparently increased because they're so close.

But I hold by what I say - Without any 'hue distortion', blue at 240 degrees (or definition B of Domer's post), mixed with pure yellow (60 degrees) will make a hueless colour.



I'm going to knock this on the head for you once and for all. Go and buy a cheap tube of a yellow acrylic (pick the pigment by referring to Handprint's suggestions for primary yellow) and a tube of Ultramarine Violet - a colour obviously far more violet than anything one would consider blue. Now mix them together in steps and tell us what you get. Since both paints are single-pigment colours this simplifies any discussion about possible causes since there's no possibility of one pigment in a mixed starting colour dominating at one concentration and not at another.

No need for me to test - I take your word for it that this would make a green-ish hue. That's quite a surprise, so there must be a good deal of 'real world hue distortion' going on here. By the way, just for the record, Ultramarine Violet is very close to 240 degrees according to this Handprint site (http://www.handprint.com/HP/WCL/waterv.html).



Didn't I say it would be a sophistry? I am continually getting the impression this is only an intellectual diversion for you - please tell us now whether you're concerned only with the theoretical side of things or in practical applications.

In my first post, I said I was a graphic artist who loved discussing colour theory. The practical applications interest me too, as it provides insight into what's happening at a theoretical level - if you see what I mean :)



I was commenting on a quote from Marc as you can see.

Ooops, my mistake. It's easy to get lost when sorting through all the quotes.



And as for the next part you're forgetting again that mixing lines aren't always straight, especially for mixed greens. You must have seen how far off they can be from the theoretical to the practical.
I do agree. Mixing hues in the real world will often differ from the theoretical colour wheel thanks to 'hue distortion'. Nowhere did I say that real life will always follow the colour wheel exactly. If it did, then cyan and red would always make a hueless colour, as would green and magenta (they're opposites on the colour wheel of course).



wait, the yellow light? *which* yellow light? the yellow at 570nm, or 575nm, or maybe 580nm, or 567nm? well, you're the boss, you get to choose. which one do you want?

but if you choose 570nm, you'll get a different "pure" blue than if you chose 575nm. so why did you choose one rather than the other?

For this very reason, I avoid defining colour hue by wavelength, and prefer to judge it by eye cone response. As you obviously know, the visible spectrum misses out magenta as a monospectral wavelength completely! I wonder if there's a more fundamental level than even eye cone response. If there is, it's probably lodged back further than the eye, perhaps the 'subjective' mind itself.

Marc Sabatella
03-21-2004, 06:55 PM
> a: no apparent hue bias towards green or red?

> b: equal stimulation of R & G cones, and higher stimulation of B cones?

To me, B seems to be the most accurate, scientific and rounded definition of a 100% blue hue, because the eye cones are really at the root of the issue.


To me, this is by far the most artificial of the proposed definitions. Consider, this method would fail miserably at defining red or green - what wavelengths have equal stimulation of B & G but higher R? The fact that the eye cones have been given names "red", "green", and "blue" is more a generalization than a real statement about the exact nature of these colors.


Because of this, I would say A equals B.


This doesn't follow either. We aren't conscious of cone response, and there is no reason to assume that a color that triggered more G than R would actually be *perceived* as greenish. The fact that no one in the world calls ultramarine a greenish blue should make this obvious.


If the green bar is higher than the red, then this indicates to me that there's a green bias


No, it doesn't. It simply indicates that the cones we, for the sake of convenience call "green" cones are more stimulated. This happens at many wavelengths that no one would reasonably call "green", and doesn't happen at many wavelengths that most people *would* consider green. I think you are reading way, WAY too much into the naming of these cones. Pretend the cones were just "Roger", "George", and "Bartholomew", and see how much sense these types of arguments make.


Marc made an interesting point how blue was almost right next to green in the visible spectrum (compared to red which is far away). What I am willing to consider is that the amount of 'hue distortion' (when it does occur) is apparently increased because they're so close.


That wasn't the main point of my observation, though - it was that R cone response in bluish colors would not necessarily be from actual reflectance in the red region of the spectrum, whereas G cone response would much more likely be. Thus, even a blue that had equal G and R cone response could still make perfectly viable greens, because there is significant green reflectance but not red reflectance. This isn't any sort of "distortion", this is a plain and simple application of substractive mixing using the actual spectrum and not imaging that thinking in terms of cone response in the way you have been will tell you everything you need to know.

{QUOTE]
For this very reason, I avoid defining colour hue by wavelength, and prefer to judge it by eye cone response. As you obviously know, the visible spectrum misses out magenta as a monospectral wavelength completely! I wonder if there's a more fundamental level than even eye cone response. If there is, it's probably lodged back further than the eye, perhaps the 'subjective' mind itself.[/QUOTE]

And I'd claim this is much *more* relevant than cone response as a means of gauging color. If everyone in the world is saying that a given blue exhibits no bias toward green or red even though the George cone is firing more than the Roger cone, why not take that at face value? Why define the color blue in terms of equal stimulation of two cones, when that in general doesn't work as way of defining other colors, either?

Twinbee
03-25-2004, 08:43 AM
> a: no apparent hue bias towards green or red?

> b: equal stimulation of R & G cones, and higher stimulation of B cones?


To me, B seems to be the most accurate, scientific and rounded definition of a 100% blue hue, because the eye cones are really at the root of the issue.

To me, this is by far the most artificial of the proposed definitions. Consider, this method would fail miserably at defining red or green - what wavelengths have equal stimulation of B & G but higher R?
This is it though - it avoids the complications of wavelengths altogether. A 100% red hue for example, one can simply say "equal stimulation of G & B cones, and higher stimulation of the R cone".



The fact that the eye cones have been given names "red", "green", and "blue" is more a generalization than a real statement about the exact nature of these colors.

As far as I know, if the green cone is stimulated on its own, this as as green as it gets, and everybody would most likely see this 'green' hue as the same. It's almost impossible to prove this, but I think that's the general consensus. Same goes for red and blue too.



The fact that no one in the world calls ultramarine a greenish blue should make this obvious.

It is so close to real blue that I'm not surprised so many call it blue. Only once you put it side by side with real blue would you notice the difference.

As an aside, I remember having a shock when I saw how little saturation the cyan colour on monitors had. Only when I compared it to true cyan (through the use of an optical illusion) did it become apparent that I was seeing a watered down version. See the 'Eclipse of Mars' at:
http://www.skytopia.com/project/illusion/illusion.html




For this very reason, I avoid defining colour hue by wavelength, and prefer to judge it by eye cone response. As you obviously know, the visible spectrum misses out magenta as a monospectral wavelength completely! I wonder if there's a more fundamental level than even eye cone response. If there is, it's probably lodged back further than the eye, perhaps the 'subjective' mind itself.
And I'd claim this is much *more* relevant than cone response as a means of gauging color. If everyone in the world is saying that a given blue exhibits no bias toward green or red even though the George cone is firing more than the Roger cone, why not take that at face value?

Because someone could very easily give a different name to the same colour. For example, many people call bright cyan, "blue" - when we all know that they are 60 degrees apart. The only objective way of quantifying colour so we can be fairly sure we are talking about the same colour is through eye cone response.

JamieWG
03-25-2004, 09:34 AM
As an aside, I remember having a shock when I saw how little saturation the cyan colour on monitors had. Only when I compared it to true cyan (through the use of an optical illusion) did it become apparent that I was seeing a watered down version. See the 'Eclipse of Mars' at:
http://www.skytopia.com/project/illusion/illusion.html

OMG! That site has the most amazing illusions I've ever seen.

Jamie

Richard Saylor
03-25-2004, 09:41 AM
If you define color in terms of cone cell response, then color becomes a subjective phenomenon, since the spectral sensitivities of cone cells are bound to vary significantly from one person to another, just there are wide variations in the taste buds, sense of touch/pain, rod response to photons, etc. This may be interesting from a biological/psychological perspective, but...... :confused: :confused: :confused:

Marc Sabatella
03-25-2004, 07:50 PM
This is it though - it avoids the complications of wavelengths altogether. A 100% red hue for example, one can simply say "equal stimulation of G & B cones, and higher stimulation of the R cone".


But wavelengths are one of the few things we *can* measure objectively! Why do you call them "complications"? As far as I am concerned, it is cone responses that are complications. You seem to be suggesting we define "true red" as a color that does not exist in the spectrum. Why would you want to do that? We've got these perfectly good colors that exist universally in nature; why not use them?

In any case, colors that reflect red wavelengths - necessary if we are to perceive them as red - also trigger a lot of G cone response. In order to get a color that behaves as you describe - equal stimulation of G & B cones, we'd have to add quite a bit of blue reflectance. What makes you think such a color would look red at all? Can you find one of those cone response diagrams for a pigment or mixture that has these properties? What does the color look like? We've already seen that a color with equal G & R response is more violet than ultramarine, something msot people would not identify as "true blue" at all. I'm guessing a color with equal G & B resposne but higher red would also look very violet; not red at all.


As far as I know, if the green cone is stimulated on its own, this as as green as it gets, and everybody would most likely see this 'green' hue as the same.


You are right, this would be impossible to test - no human being has ever had their G cones stimulated on their own, nor is that ever likely to happen in our lifetimes. But in any case, even if it were true, it would not necessarily follow that a real life color that produced equal B & R response but higher G would be perceived as the same hue as this mythical pure green.


It is so close to real blue that I'm not surprised so many call it blue. Only once you put it side by side with real blue would you notice the difference.


It's not that most people can't see the difference - they *can* see the difference, and say that ultramarine is more *violet* (not more *green*) than what they consider real blue. So-called "240 degree" blue is more violet than that even, and the color you are describing, that actually manages to get the R response up to the level of the G response, would almost certainly be even more violet still.

The point I'm making is that while it is *possible* to define colors in the way you suggest, why not invent new names for these colors? Red, green, and blue are already taken. You can still roughly describe the resultant colors as being basically bluish, reddish, and greenish (although I still have my doubts whether or not that would actually be the case for the latter two), but thinking that these are "true" versions of the colors strikes me as going against centuries of usage of these terms.


The only objective way of quantifying colour so we can be fairly sure we are talking about the same colour is through eye cone response.

No, that is most likely subjective as well, not to mention difficult to measure. The best objective way of quantifying color is through reflectance diagrams.

Richard Saylor
03-25-2004, 09:55 PM
This is it though - it avoids the complications of wavelengths altogether. A 100% red hue for example, one can simply say "equal stimulation of G & B cones, and higher stimulation of the R cone".

Aside from the difficulty of measuring this, doesn't this assume that everyone's cone cells exhibit precisely the same spectral sensitivity?

drollere
03-26-2004, 12:13 AM
what, really, is the point of specifying a subjective state through a neural response, or specifying a perception in terms that forcibly eliminate any stimulus?

translation: what, really, is the point of specifying a color through a cone response, or specifying a color perception without permitting any reference to a color stimulus?

why do you have to declare a specific point in order to talk about color? because color is a highly contextualized experience. without the point, you can't specify the context, and without the context, all color terminology, all color relations, all color dogmas or theorems or models, become meaningless.

the concept of a blue that contains neither red nor green was launched in the 19th century and forms the basis of several color models, including the swedish natural color system. the unique blue in that system (the scientists who developed it know better than to use moral adjectives such as "true blue" or "pure blue" or "best blue") has been carefully calibrated by experiment to appear to most people to contain neither green nor blue. it is approximately the hue of cobalt blue or phthalo blue red shade. it is "greener" than ultramarine blue.

so, what have you learned? if you don't embrace the peculiar perceptual theory upon which the concept of "unique blue" rests, what good is unique blue to you?

color is a highly contextualized experience. cone responses have a useful relationship to reflectance profiles in the narrow case of an unrelated color viewed under limited conditions. it seems to fly under the radar of everyone posting here that, in fact, in vivo human cone responses have never been measured directly, that the retinal elaboration of color stimuli is entirely uncharted, and that the relationship between human cone responses and general, everyday color perception of light stimuli from natural objects is still beyond the reach of even the most sophisticated "theoretical" model of color vision. and we still haven't left the eyeball.

it seems unnoticed also that the formulation "R and G cones equal, and B cone greater than both" can describe a pure black, a slightly violet tinted white, or any gray in between; or a color that most people would report to be a purple or a blue violet or a middle blue, depending on the luminance of the color and the surround, and the adaptation state of the visual system; or cool colors of widely varying saturation, in lights and in paints; or a color that is undefined to most of us, if viewed by a dichromat. in short, it's a formulation that says nothing useful.

what's worth looking at: the word "blue" seems to float along anyway, tantalizing with its specific linguistic denotation, functioning in the same apparent way as any other noun, like "dog" or "tapioca," and certainly it *seems* like you should be able to slap "blue" into in a sentence with other words, such as "true" or "cone" or "scientific" ... but you're merely playing with words. as einion said: sophistry.

to address color perceptions as such, you need to have a specific perceptual problem you are trying to solve -- otherwise, the ambiguity about the context sucks the meaning out of the words. you're required to have a specific viewing situation and viewable stimulus before you can talk about the "meaning" of cone responses, or wavelengths, or put color labels on stimuli.

sure, you can still talk about "blue" in the absence of specific problems or settings, but then you're debating language, not perception, and the definitions or truth operations you conventionally accept or reject as relevant to color labels. but that debate spins you into epistemology (empiricism, solipsism, platonism, relativism, absolutism), not perception, and certainly not anything to do with art.

Twinbee
04-02-2004, 04:30 PM
OMG! That site has the most amazing illusions I've ever seen.

Glad you liked them :-)



If you define color in terms of cone cell response, then color becomes a subjective phenomenon, since the spectral sensitivities of cone cells are bound to vary significantly from one person to another,

It's true that wavelengths of light will trigger different proportions of RGB cone response for different people, but if we measure the actual cone response voltages, then we can (or should) be able to determine with a fair degree of precision the colour that the person sees. This way, we can avoid the added variable of spectral sensitivity.




This is it though - it avoids the complications of wavelengths altogether. A 100% red hue for example, one can simply say "equal stimulation of G & B cones, and higher stimulation of the R cone".

But wavelengths are one of the few things we *can* measure objectively! Why do you call them "complications"?

'Complex' as in people will see a different colour from the same wavelength/s. Measuring by cone response avoids this issue.



You seem to be suggesting we define "true red" as a color that does not exist in the spectrum. Why would you want to do that? We've got these perfectly good colors that exist universally in nature; why not use them?

Well for one, the spectrum misses out the fundamental colour magenta (unless you combine wavelengths of course).



In any case, colors that reflect red wavelengths - necessary if we are to perceive them as red - also trigger a lot of G cone response. In order to get a color that behaves as you describe - equal stimulation of G & B cones, we'd have to add quite a bit of blue reflectance. What makes you think such a color would look red at all?

Again, I'm ignoring wavelengths completely. Imagine a hypothetical way of precisely stimulating the desired eye cone/s without affecting the others.




As far as I know, if the green cone is stimulated on its own, this as as green as it gets, and everybody would most likely see this 'green' hue as the same.

But in any case, even if it were true, it would not necessarily follow that a real life color that produced equal B & R response but higher G would be perceived as the same hue as this mythical pure green.

I'm almost certain that it would. But I accept I could be wrong, especially in light of Drollere's post.



No, that is most likely subjective as well, not to mention difficult to measure. The best objective way of quantifying color is through reflectance diagrams.

As far as I know, the 3 bar (RGB) graphs that Domer posted are based on eye cone response. Perhaps drollere can confirm this.



the concept of a blue that contains neither red nor green was launched in the 19th century and forms the basis of several color models, including the swedish natural color system. the unique blue in that system (the scientists who developed it know better than to use moral adjectives such as "true blue" or "pure blue" or "best blue") has been carefully calibrated by experiment to appear to most people to contain neither green nor blue. it is approximately the hue of cobalt blue or phthalo blue red shade. it is "greener" than ultramarine blue.

Interesting. Have you any reference for this? Are you sure they were stimulating someone's blue cones entirely with the R and G cones left untouched?

If this is true, and that they really saw a more green hue than what theory would say, then I concede that there must be a level even more fundamental than eye cone response. Perhaps we need to go right down to the red/green and yellow/blue ganglion cells. My argument would then be based on the precise stimulation of these cells - presumably quite difficult to test by experiment though.

Assuming you're right, and the response of individual eye cones is not what I expected. Even then, I would still say there's an easy definition for 'true blue'. It would be the colour that could not be mixed from any other colours. Something like cyan and magenta light would only make a half-saturated blue (additive light mixing). However, we could get fully saturated cyan by mixing green and blue of course. Additionally, the truest red, green and blue would likely produce the largest gamut of hues.



it seems unnoticed also that the formulation "R and G cones equal, and B cone greater than both" can describe a pure black, a slightly violet tinted white, or any gray in between; or a color that most people would report to be a purple or a blue violet or a middle blue, depending on the luminance of the color and the surround, and the adaptation state of the visual system; or cool colors of widely varying saturation, in lights and in paints; or a color that is undefined to most of us, if viewed by a dichromat. in short, it's a formulation that says nothing useful.

Yes, but assuming the background colour was constant, and the 'adaption' state of the eye beforehand was also constant. In other words, assume someone was seeing a black backdrop for half an hour, and then saw the desired hue for inspection against this backdrop of black.

Marc Sabatella
04-03-2004, 04:44 PM
'Complex' as in people will see a different colour from the same wavelength/s. Measuring by cone response avoids this issue.


But creates the equally problematic issue - what assurance do you have that people with the same cone response are "seeing" the same color in any meaningful sense?


Well for one, the spectrum misses out the fundamental colour magenta (unless you combine wavelengths of course).


Precisely. And combining wavelengths happens all the time in the real world, is easily measurable, and is completely objective. Why you insist on introducing the difficult to meausre and completely subjective notion of cone response as a measure of color is beyond me. It might have use for you own thought experiments, but recall, this discussion started out when you were explaining why we should teach school children why blue and yellow don't make green. They do in the real world, and they do in the theoretical realm that best models the real world. Why make up a theoretical realm that *doesn't* model reality nearly as well as the model we already have, and then insist this is more "real" than reality?


Imagine a hypothetical way of precisely stimulating the desired eye cone/s without affecting the others.


Given that color can be explained quite well and in a relatively straightforward manner without resorting to inventing alternate realities that will never exist, I see no incentive to imagine any such thing.


As far as I know, the 3 bar (RGB) graphs that Domer posted are based on eye cone response.


They are definitely supposed to indicate that, sure. But to the extent they do, they are mostly likely simply inferred from the reflectance diagrams based only on available information regarding regarding cone sensitivity to different wavelengths. Or else, they represent some sort of composition average of how the cones of different people, who all perceive color slightly differently, might respond. The point being, this stuff is subjective. Reflectance diagrams are not.


Interesting. Have you any reference for this? Are you sure they were stimulating someone's blue cones entirely with the R and G cones left untouched?


Of course they weren't - this would be impossible. The experiment in question was regarding *perception* - what color do real people, who exist in the real world, actually claim to perceive as true blue. The color you are proposing as true blue simply looks more violet than what real people call true blue.


If this is true, and that they really saw a more green hue than what theory would say


They don't see something more green than theory says - they see exactly the color that theory says. At least, not the theory that I have read about. It doesn't match what *your* hypothesis predicts, but this simply demonstrates why I am arguing that your hypotheses aree simply wrong - they don't match reality nearly as well as the existing theory does. The existing theory works just fine; we don't need a new one.


Perhaps we need to go right down to the red/green and yellow/blue ganglion cells.


If so, it would would suffer exactly the same flaw as your current false hypothesis - assuming that because we attach the conveninece label "blue" to these cones, and by extension all the cells involves in the neural pathways they trigger - that this has some significance regarding what is actually perceived as blue. As I said before, you are much better off thinking of these as Bartholomew cones, and stop assuming that they relate to what we perceive as blue in precisely the way you seem to be expecting.


Even then, I would still say there's an easy definition for 'true blue'. It would be the colour that could not be mixed from any other colours.


Using additive or subtractive mixing? And how are you defining "colour" here? In terms of eye cone response? If so, you are likely to come back to the same color - something rather more violet than what real people in the real world call blue. I see no reaosn to call this "true blue"; why not just call it "the color that happens not to be mixable" and be done with it? The name "blue" is already taken, and it refers, through centuries of historical usage, to something different.

Twinbee
04-06-2004, 10:02 AM
Precisely. And combining wavelengths happens all the time in the real world, is easily measurable, and is completely objective. Why you insist on introducing the difficult to meausre and completely subjective notion of cone response as a measure of color is beyond me. It might have use for you own thought experiments, but recall, this discussion started out when you were explaining why we should teach school children why blue and yellow don't make green. They do in the real world, and they do in the theoretical realm that best models the real world. Why make up a theoretical realm that *doesn't* model reality nearly as well as the model we already have, and then insist this is more "real" than reality?
I suppose one reason why it might be good to teach to school children that they don't make green is because any green that is produced will mostly always be muddy. It would be far better to teach that /cyan/ and yellow make green. Also, in the 'real world' on monitors, mixing blue and yellow makes a hueless color, and as you probably know, mixing either subtractively or additively should in theory always produce the same hue angle (providing one isn't mixing a primary colour and its nearest secondary colour - such as red and yellow). So in other words:
Yellow and cyan in subtractive mixing is green (120°), and in additive mixing it is also green (120°), but half saturated.
Yellow and magenta in subtractive mixing is red (0°), and in additive mixing, it is also red, but half saturated.
Red and blue in subtractive mixing is black in theory (or dark magenta because of the ink impurities), and in additive mixing, it is also the same hue angle - magenta (300°).

This is just one reason why I think it is neat to say (teach) yellow and blue in both subtractive /and/ additive mixing is a hueless colour. Okay, so sometimes, this isn't what happens in terms of paints, but it really helps to get a unified idea of additive and subtractive mixing.


Of course they weren't - this would be impossible. The experiment in question was regarding *perception* - what color do real people, who exist in the real world, actually claim to perceive as true blue. The color you are proposing as true blue simply looks more violet than what real people call true blue.
The same angle of blue (240 degrees) that I'm on about is used on monitors, and I'm sure most people would agree that it can be classed as 'blue'.


Using additive or subtractive mixing? And how are you defining "colour" here? In terms of eye cone response? If so, you are likely to come back to the same color - something rather more violet than what real people in the real world call blue. I see no reason to call this "true blue"; why not just call it "the color that happens not to be mixable" and be done with it? The name "blue" is already taken, and it refers, through centuries of historical usage, to something different.

Definitions do change throughout time though. Perhaps everyone should define one hue as blue and be done with it. At the moment, people call 240 degrees at least three different names: "blue", "blue/violet", and even "violet". I don't particularly mind what's it's called as long as everyone agrees to one definition :-) Naturally, I think just calling it 'blue' is best for many reasons, but I understand that many artists tend to think of this hue as "blue/violet", (and would instead call something closer to 210 degrees "blue").

Marc Sabatella
04-06-2004, 05:57 PM
I suppose one reason why it might be good to teach to school children that they don't make green is because any green that is produced will mostly always be muddy.


Muddy is a relative term, but most blue pigments mixed with most yellow pigments produce perfectly acceptable greens, and this result is predicted perfectly by subtractive mixing based on reflectance patterns. Again, I see no reason to try to teach children a new theory that fails miserably in reality the first time they try actually mixing blue and yellow paints. Why not teach the existing theory, which works?


Red and blue in subtractive mixing is black in theory (or dark magenta because of the ink impurities)


Again, ink impurities have nothing to do with it. It is simply a matter of understanding how the theory does indeed predict exactly this result. The specifc color of the mixture will depend on the specific reflectance patterns of the blue and red pigments used. Many if not most red pigments combined with many if not most blue pigments will produce a clearly recognizable violet both in theory and in fact.


as you probably know, mixing either subtractively or additively should in theory always produce the same hue angle (providing one isn't mixing a primary colour and its nearest secondary colour - such as red and yellow).


There is a basic problem here in trying to do your color calculations using the artificial construction of "hue angle" to classify different colors and guess as to their behaviors in subtractive mixing. It isn't hue angles that get combined in subtractive mixing - it is reflectance patterns. It makes no sense to talk about what subtractive mixture two colors of given "hue angles" would produce in theory without talking about the specifc reflectance patterns that led to the perception of the given "hue angles". Two colors of the same "hue angle" might behave entirely differently in mixtures, and this isn't because of "impurities"; it is simply how subtractive mixing works. Trying to predict results of subtractive mixing without looking at reflectance patterns across the whole spectrum is like trying to guess how heavy an object is based only on its volume without knowing its density. Sure, you might guess close to right some times, but you're going to be wrong a lot too, and it's not due to impurities in the object, but simply because you weren't considering all the information you need to consider.


This is just one reason why I think it is neat to say (teach) yellow and blue in both subtractive /and/ additive mixing is a hueless colour. Okay, so sometimes, this isn't what happens in terms of paints


If it were only "sometimes" that blue and yellow happened to accidentally make green in the real world, I wouldn't be objecting. It's always. 10 times out of 10. 1000 times out of 1000. 1,000,000,000 times out of 1,000,000,000.


The same angle of blue (240 degrees) that I'm on about is used on monitors, and I'm sure most people would agree that it can be classed as 'blue'.


Sure - most people would classify any color within shooting distance of that color as some kind of blue. But asked to identify "true blue" - one with no apparent leaning toward violet or green - people regularly identify something less violet as being truer.

Still, I think this is the real source of the disagreement. A yet-to-be-discovered pigment with a hue angle of 240 might indeed have a reflectance pattern such that it failed to make green when mixed with yellow, although I think it equally likely, for reasons I've stated before, that it would still have enough green reflectance, and lack of red reflectance, to make adequate greens. But most of the colors that people call "blue" are much less violet than this, which is why, unless you redefine "blue" to mean something other than what people already understand blue to be, all known or likely to become known blue pigments will continue to make green when mixed with yellow.

Richard Saylor
04-06-2004, 08:21 PM
It's okay to teach kids that blue and yellow make green. It is not okay to teach them that red, blue, and yellow are the subtractive primaries. Blue and yellow make green because all known blues are cyan-like in that they reflect both blue and green. Red and blue usually make violet because most (but not all) reds are magenta-like in that they reflect wavelengths from both ends of the spectrum. I don't see anything at all wrong with teaching them that cyan, magenta, and yellow are the primaries which yield the largest gamut, and that other choices of primaries work because of certain similarities with these. If the CMY primaries are not going to be taught, then the word 'primary' should not even be introduced.

There is a minimal palette thread over on the watercolor forum. Most everybody was saying that they used a split primary RBY system, until I asked how they knew which yellow to mix with which blue to give the brightest green. One person gave the correct answer using the color-wheel system she had been taught. (It's interesting that her 'cool' RBY colors are exactly my choice for CMY primaries.) Another person gave the correct answer from basic theoretical principles. I suspect that most people use split primaries in immitation of some painter whom they admire without even understanding the principles involved. That system would be too complicated for me unless I always arranged the six colors in a circle in the same order, and that's too much like work. :(