Thursday, March 18, 2010

Charting Pigments

Every pigment occupies a specific position on a color wheel. The position is a combination of its hue and chroma.

The chart above shows where many familiar pigments would appear. Earth colors, since they’re low in chroma, appear close to the center in the red and yellow sectors, and black and white sit together at the center.

Since maximum known chroma varies from one color to another, even the some of the strongest tube colors often fall short of the perimeter.

This chart doesn’t include every pigment. But even if it did, there would gaps where pigments are unavailable.

Further info:
This chart is hand-painted in oil, using dots of pigment over a high-key gradated Yurmby wheel. Note that the wheel is set up according to the way our eyes actually perceive color relationships. The complements are additive, not paint-mixing complements. So ultramarine blue is opposite cadmium yellow, not burnt sienna. The reason I set it up that way is because I want to know how the colors will relate and vibrate on my picture from the perspective of the viewer’s experience. I’m not as worried which paints I need to mix a gray.

This chart was based on data compiled by Bruce MacEvoy,


Sara Light-Waller said...

Thanks for the great post. I've never quite understood the plotting of the chroma color chart before and you've made it very clear. Thank you!

Sakievich said...

This complicates color a bit...

On a separate note, during the Hudson River Fellowship, Jacob Collins gave a fascinating lecture on color and light and how they function in the atmosphere and how that relates to painting landscapes from life and the imagination.

Scale said...

Thank you for the precise distinction of viaul vs. mixing complementaries, I didn't have it so clear.

As a beginner I tend to worry a lot about which color should be mixed to a given color to go towards gray. It would be useful to have a more or less universal color wheel for mixing too. But I wonder wether it's possible to have one or it's all strictly dependant on the brand and quality of pigments in use. In fact in the brand of acrylics I use most often, ultramarine seems to mix to the most neutral gray with Van Dyck brown rather than burnt sienna.

António Araújo said...
This comment has been removed by the author.
António Araújo said...

the behaviour of pigment mixing cannot be predicted accurately by any single colour wheel. Unlike mixing lights, where a colour has a well-defined complement (or rather, a set of them along a straight line on the chart), and any colour arising from mixing always lies on that line, the colours you get from mixing paints do not lie on a line but on a curve whose shape cannot be predicted from the colour of the end points (it depends on a lot of physical characteristics of the paints and not just on their colour).

So a specific paint can have many mixing complements (that is, other paints that will mix with it to a near neutral), and those complements can, and will generally have different hues than the ones you would expect from light mixing.

Light mixing is well behaved and allows for a straithforward mathematical theory , and so is the basis of our understanding of color. But mixing paints requires not only knowledge of the theory, but knowledge of the actual materials too - meaning, the need for lenghty practical experience cannot be avoided.

Don Cox said...

"As a beginner I tend to worry a lot about which color should be mixed to a given color to go towards gray."

Why not mix gray with it?

António Araújo said...


because there is this big myth about not using black (because supposedly it murders the colors or something :)).

Actually, myths aside, using gray is a great way to lower chroma. You just mix your color with the gray of the same value and in so doing you lose chroma without touching the value or hue (hopefully).

On the subject of "mixing with black", one can find a very good explanation of how to mix a shading series by using black (without killing or maiming or otherwise emotionaly distressing any colors :)) at David Brigg's site ( ).

Roberto said...

I don’t like mixing-down my colors with grays or blacks, unless it is absolutely necessary. I tend to wind up with mud to fast.
The key to successful color mixing is creating the widest variety and range of colors from the fewest number of paints, and still maintaining clarity and saturation of hue. To achieve this I use an opaque(warm) value and a transparent(cool) value for each primary(pigment):

Yellows = Arylide (cool/trans), Cadmium (warm/opaque)

Reds = Quinacridone/Magenta (cool/trans), Cadmium Red (warm/opaque)

Blues = Ultramarine (cool/trans)*, Cobalt or Cerulean/Cyan (warm/opaque)

*(I have seen ultramarine referred to as warm because it leans towards the red, but when I mix it with white and compare it to cobalt or cerulean it sure looks cool to me. I think its because it leans towards a cool/transparent or Magenta red).

Each Secondary(pigment) color can then be mixed to four values:
(cool/trans) to (cool/trans);
(cool/trans) to (warm/opaque);
(warm/opaque) to (cool/trans);
and (warm/opaque) to (warm/opaque).

In order to maintain clarity as I mix down my tones I first add an ‘earthtone’ to each Hue
(instead of mixing complaments or grays):

Red and/or Magenta = Burnt Sienna
Orange = Raw Sienna
Yellow = Ochre
Green = Oxide, or Raw umber
Blue and/or Cyan = (I have not found a satisfactory ‘earthtone’ for blue, other than a lower-toned blue pigment, such as Prussian or Indigo)
Violet = Burnt Umber

Next , I will mix the Complament (pigment), either cool or warm, to lower the chroma.
Only after mixing down with hues (or color) do I resort to mixing down with grays; and my final and last resort to lower chroma is to mix with the dreaded blacks.
Usually, if I’m trying to ‘match’ a given color, mixing down with grays or black is necessary, as most of the commercial paint systems are based on Munsell.
This approach can always be expanded to introduce new and various pigments, (paying attention to the opaque/transparency relationships is very helpful).

I rarely use a straight black. I prefer to mix a cool black with ultramarine, or a warm black with burnt umber or burnt sienna.

I am constantly learning and improving my approach and I have learned a lot from this excellent blog and the outstanding contributors, and the resources and links so generously shared. Thanx to all and especially to Jimmy G. -RQ

Jose Romero said...

Roberto, your system of mixing down tones adding an ‘earthtone’ to each Hue seems pretty logical to me. Regarding blue, have you tried using black? Blue and black make a nice dark blue; I remember Dan Gerhartz recommending it in one of his videos.

Another way of making a subdued blue would be mixing it from the adjacent colors in the color wheel (violet plus viridian, for example).

Scale said...

Thank you for the great tips!
Antònio: I see, I guess I'll have to do a lot more practice. I'll stick to the current brand for now, to avoid confusion.

Don Cox: mostly because I have seen only a couple values of premixed grays, and mixing my own gray value means using at least four paints for intermediate hues (white + black + 2 colors) which makes it very hard to control values and saturation. I realize using acrylics for this is part of the issue because they dry so fast. Maybe I'll try using a delaying medium to mix gray values more accurately.

Roberto: mixing a pure hue with an earth of similar hue seems a great way to control saturation, I'll have to try that too.

António Araújo said...

that is fascinating, because (am I mistaken?) then you won't be able to divide the fixed-value space (not plane anymore) into hue and chroma. With four receptors, you'll need three dimensions to describe the space of all colors of a given value. That's trivial to do in a CIE 4-stimuli sense, but I wonder what happens to the Munsell perspective? I wonder if it can be (some sort of ) hue, chroma, plus something else (what kind of something?), or if it would be some completely different dimensions. The dimensions are determined by the tests you do in order to define them. I wonder what those tests must be, in order to isolate each dimension.

Or am I just speaking nonsense?

For further complications, check out february issue of scientific american (the "impossible colors" article). It turns out you can (in a specific experimental setup) see new colors that arise as a mix between complementary hues. For instance, turns out you can see a reddish green, or a greenish red, instead of just a gray.

António Araújo said...

Actually, I was thinking about it and I think I get it. It's not all that different: You just use spherical coordinates instead of polar coordinates on the constant value space. Instead of the horseshoe of CIE plane you get a sort of ellipsoid truncated by a plane (on the purple side). Then you still get a measure of saturation along the line that is centered at white and describes the ellipsoid, the difference being that you need two angles and not just one to describe all the hues.
Then you just re-scale according to perceptual differences, and get a munsell renotation in this news space. It will be...hem... a bit harder to do a munsell "sphere" ( or catalog) though :D

António Araújo said...

>the difference being that you need >two angles

meaning that if you are tetrachromat, you have to consider a color "sphere" to describe hues, not a color "circle". The rest should be about the same.

And, if the fourth type of cone is not very important perceptually, you might get a spheroid that is so short in the new dimension that it looks almost like a disc on a plane (like the head of the USS enterprise :)). In that case tetrachromats would only very rarely notice that they have the ability to see some colours we don't.

Or I may just be seeing this whole thing wrong. Anybody has a better view of this or has actually read anything about it from informed sources?

Anonymous said...

Fascinating post, James. Colour is such an interesting topic and one I've thought about often in the past. I've noticed a correlation between the Golden Ratio and appealing colour schemes, I elaborate more on this in this post:

Intriguing how art and science are so often intertwined.

artistguy said...

Love it

Roberto said...

Jose Romero-
Thanx for your suggestions. This series has really stimulated me to revue and re-examine my approach to color theory. I am trying to follow up on all the links and ideas presented. Stapleton Kearns’ excellent blog has a section on color-mixing that supports your use of black (ivory black) as a substitute for earth-blue:

Who says you can’t teach an old dog and old trick! -RQ

groperofeuropa said...

Not sure if this has been posted before, but the fellow from xkcd did a review of the naming of colours between genders as a bit of a joke, but with some interesting results.
Essentially - guys and girls care equally about colour differentiation and humans are incapable of spelling the word "fuchsia" without the aid of a spellchecker, as well as Google..

Jim Moyer said...

The easiest way that I have found is as follow. If I want to gray a yellow-green I ask my self what is the comp to yellow (violet) then what is the comp for green (red) so I will use a red-violet. If a blue-green than orange for the blue and red for the green or
red-orange will do the job.