Tuesday, February 23, 2010

The Color Wheel, Part 3

A color that holds a position directly across the wheel from another is known as a complement. In the world of pigments and color mixing, the color pairs are: yellow-violet, red-green, and blue-orange. When pigment complements are mixed together, they result in a neutral gray, that is, a gray with no hue identity. This is all pretty familiar to anyone who has fooled around with paints.

But in the realm of afterimages, light mixing, and visual perception, the complement pairings are slightly different. Blue is opposite yellow, not orange.

You can see this for yourself with the diagram above. Stare at the middle of the colorful circle for 20 seconds. Then shift your gaze to the middle of the white circle and relax your eyes. The complementary colors should emerge. Note that the afterimage of blue is yellow and vice versa.

But if you were to mix that yellow and that blue as paints, you wouldn’t get a gray, you’d get a green. In fact, pigments can behave unpredictably when mixed. Intermediate mixtures don’t always land on the straight line drawn between the two starting colors.

How does this affect the way we design a color wheel? First of all, you have to decide whether you want to try to represent the practicalities of pigments or the behavior of color in an optical or a mathematical realm. In other words, your color wheel must either represent the ideal world of optical color or the physical world of paints, but no single wheel can accurately represent both color universes.

Many color wheels include the dimension of grayness versus intensity, known as chroma, also commonly called saturation. Here’s the traditional artist’s color wheel I made years ago, which goes to zero chroma at the center.

(By the way, which term do you use? Please vote in the poll at left). "Chroma," a term invented by Albert Munsell, is the degree a color ranges between neutrality and vibrancy or purity.

Tomorrow I'll try to take a look at that question posed yesterday: Are some colors really more primary than others?

Reviewing the posts in this series:
Part 1: Wrapping the Spectrum
Part 2: Primaries and Secondaries
Part 3: Complements, Afterimages, and Chroma
Part 4: Problems with the Traditional Wheel
Part 5: The Munsell System
Part 6: Cyan, Magenta, and Yellow
Part 7: The Yurmby Wheel


Zeke said...

I'm really enjoying these posts about color theory! Thanks for the lessons.

BTW I grew up reading Dinotopia and I'm a huge fan, the art was inspirational

James Gurney said...

Zeke, glad you're enjoying the posts, and thanks for the kind words about Dinotopia.

John Larriva said...

You've really shifted my paradigm. Now I get to ponder what light really does, not just which pigments play well with other pigments.

Anonymous said...

I think most people who have looked at a sunset with that combination of despair and amazement have internalized that paint and light are distinct things, and that you're not going to be able to achieve God's palette. :)

I agree, these posts are wonderful. It's nice to see some of the tired lessons get a good kick in the tires to see if they're still useful.

Charley Parker said...

Great series of posts. I'm glad you're going into the question of what primaries actually are, and the idea of a 4 primary color system (which I've seen referred to as "artist's primaries").

Ray Tsang said...

love your blog james, its an inspiration as is your book- i cant wait till the new one comes out. i read here http://www.huevaluechroma.com/091.php that the terms chroma and saturation are not interchangable as chroma refers to value and saturation and is a more accurate representation of colorfulness than the term saturation. the first time i read through davids site it blew my mind. also i wanted to ask you about the color batches you mixed, ive tried this and i tend to waste alot of paint, does this happen to you and if not how do you avoid it.

Mark Heng said...

I think my eyes were opened when I realized, after messing about with Photoshop, that all colors can be described with 3 words: Hue, Value, and Saturation. I had previously read Joseph Itten's book, "The Elements of Color" and found it kind of confusing and impractical, especially with it's multiple words for saturation- "Brilliance, purity, quality, diluted, intense, etc..."

inkdestroyedmybrush said...

james - this is great stuff and works well with my initial forays in to "imaginative realism" (just got the book for my birthday so i'm not too far in right now). After years of art school dogma, i emerged more perplexed by color than understanding it. The adherence to a traditional color wheel as the ultimate "explain everything" was something that never worked for me. I'm really going to enjoy reading your investigations with color... and maybe coming to a better understanding myself.

no wonder most of my work is black and white.;-)

j. w. bjerk said...

Software generally uses the term "Saturation", so it is probably bound to dominate, though in my experience people without grounding in the digital, or physical sides of art wouldn't understand "saturation", or "chroma".

And my compliments on the explanation, it's certainly more clearly expressed than it was in my Freshman art classes.

Casey Klahn said...

Awesome lessons, and very well described. You are a great teacher and communicator, James.

Anonymous said...

Munsell, imho the best and most accurate system of color, states "the endless variations possible in color can be be defined in the Munsell system by three attributes: hue, value, and chroma." Where as Chroma "is the strength or intensity of color. Neutralized, dull or greyish colors have weak (low) chroma, while intense colors (cadmium yellow) have strong (high) chroma." While Munsell breaks hue and value into steps of ten, chroma in theory has no limit. A true, neutral grey would yield a chroma of 0-no matter the value, while cadmium yellow yields a value of somewhere around 16 or so depending on the brand, and neon colors go even higher. The term saturation is generally a term used for additive color (meaning light emitted color such as a computer monitor or television. My advice on color theory is to find one system that works for you and learn it in its entirety. Don't be bamboozled by all of the other systems out there because they will only cofuse you as one system tends to contradict another. I like Munsell because it's accurate, predictable, scientific, and flexible-it easily adapts to newer, modern pigments with higher and higher chroma-something most color wheels can't do. It's also flexible enough to use for all mediums.

Anonymous said...

Here come the fundaMunsellists.

Gregory Becker said...

Hello James, I was wondering if you could describe how you made this wheel to achieve those grays and what colors you used.

jeff said...

OK I think we should try to keep this civil. To say things such as "Here come the fundaMunsellists" is uncalled for.

I use it to help me mix neutral grays, which if you have ever tried without a guide is a lot of work as your eyes will play tricks on you. The Munsell gray scales have helped me to do it faster and to train my eyes. I've used it to mix strings for still life paintings and I've also not used it. I'm not a fanatic about it but it has a long history in the art field. Frank Reilly based his palette on it. Given the track record of Reilly and his students I would say there is something to it. The beauty of it is it's very flexible and it works well. Art history is full of artist trying to organize their palettes and Munsell is only a tool for that end.

It's a very good system to describe color space. Bunky was not saying it's what people should use, he/she was only stating that it works for them.

Sally Dean said...

I stumbled upon this and love it- thanks for sharing such good information@
I gave you the sunshine award ( if you want it!)

Anonymous said...

Wonderful theory posts, extremely helpful!
Could you shine light in your later posts on the importance of making such a colour wheel yourself? I see some artists doing all sorts of colour wheels and different shades of colour, but I haven't found the reason. Why would painting it be helpful? And more importantly, how can we do it accurately?

Anonymous said...

As for the "third dimension" and the terms used to describe it, there is an interesting little read here: http://rourkevisualart.com/wordpress/2008/02/22/the-difference-between-chroma-and-saturation/

Don Cox said...

"Munsell, imho the best and most accurate system of color"

It's a good system, especially considering how long ago it was developed, but I think it has two main flaws.

First, the division of a colour circle into ten, rather than twelve or some other number divisible by four and three, is highly inconvenient.

Second, the spacing of the colours is not as even as it might be.

There is I think a lot to be said for forgetting about circles and using a simple rectilinear grid, like scaffolding. This is done in the CIE Luv and Lab spaces, and in the OSA Uniform Color Scales. In both, great care has been taken to try to get even spacing of the colours.

The vertical axis is always lightness, the two horizontal axes are basically blue-to-yellow and red-to-green, which matches the four main colours.

Don Cox said...

Liquitex used to make a range of paints based on Munsell, but it seems to be discontinued.

Thomas Jefferson Kitts said...


I allow my students us the term 'Chroma', 'Saturation', and 'Intensity' interchangeably once they understand that all three words mean the same thing. Each directly refers to the presence or amount of Hue in a mixture. Each are identical ways to express how strong or weak that Hue is.

What I don't allow my students to use in my classroom are terms such as "Shade" or "Tint" because both words can be confused with Value, Hue, or Chroma. Or confused with a combination of all three.

For example: a cad red mixed with white titanium does not produce a "Tint" (What do you mean sir, this 'Tint' is correct?) adding titanium white produces the same red Hue, but shifts that Hue towards a lighter Value (assuming the Hue isn't also being biased towards the blue -- which is a cooler Hue by the addition of that white. Which, truthfully, any white does to some extent.)

However, besides lightening the red Hue, the addition of white will also reduce the Hue's Chroma. A fact often overlooked by the casual or inattentive painter.

And we aren't even talking about modifying the base Hue, Value, and Chroma all at once, which involves considering the H/V/C model in its entirety.

I find the H/V/C system (or the Munsell Color Solid -- not that he invented it -- look to Chevreul for that) is an efficient way to mix paint colors with a high degree of predictability. Especially so if you use a Spectral Palette to mix with. The H/V/C model offers you a way to control 1 or 2 characteristics of your mixtures without affecting the other. Or ways to change only two, or all three.

Often, the novice painter struggles to deal with Chroma, whether he is aware the characteristic exists or not. Value and Hue are usually easy to differentiate, since that is what we have been exposed to from grade school on. But the idea that Chroma exists, and that it can be manipulated in helpful ways, is usually a new idea to explain, especially when teaching that novice how to mix tertiaries (low Chroma Hues) from Saturated palette of color (high Chroma primaries and secondaries, i.e., the Spectral Palette).

As for whether or not one system of color mixing is superior to another, I've seen it go every which way, with the critical difference usually having more to do the artist's understanding of how malleable color is. In the end, the H/V/C system is a fine way to anticipate the color you want, and it also provides a reliable road map for how you can mix it. It can also tell you if you can get that color using the tubes of paint you just bought at the store, or if you have to get back in the car and get some more. (Ha!)

For as most of us know, there is one thing you cannot do with paint - any paint at all -- and this is, increase the Chroma of a color by mixing two or more colors together. Chroma always goes downhill, which is why, if you need a certain intensity then you must start with intense colors.

Thomas Kitts


James Gurney said...

These are all very helpful comments, and I can hardly add to them, though some of the areas you've discussed, esp. Munsell, will come up in future posts in this series.

Sorry, I have limited access to the Internet (and electricity) today, and maybe tomorrow--we'll see--but there are three or four more posts to come in this series. Thanks, all!