Colorful Structures

Added on by Doug Dowd.

90 x 90 Day 22. Returning to a post I wrote in the early days of my blog, when it began to dawn on me that this format might work as a teaching tool. Color is a dimension of artistic practice that causes more grief than necessary, although as I type that sentence–excuse me, as I keyboard it–I do wonder about contemporary passivity in explorations of color. What do I mean? I mean that mixing pigment or ink is quite different than selecting something from a color picker in Photoshop or some other digital environment. Like so many things in our ultra-convenient cultural moment (unless of course you live in a corner of the world where things are rather less convenient, like, say, Syria) alienation from stuff works against deep knowledge. Note to self: come back to this sometime soon and argue the opposite. We no longer churn butter; for many decades we have just bought it in the store. BUT ANYWAY I miss messy pigments, and you should, too.

I wrote 6 years ago:

Classes begin anew tomorrow. I will be presenting a color workshop for junior designers and illustrators in the morning. My approach to color is extremely practical. Because I was trained as a printmaker originally, I tend to see color as a question of choice, not bravura improvisation. As a student I recall being told that I had “bad color.” This condition was not defined, nor was a course of treatment prescribed. It was sort of like having an extra chromosome: that’s just the way it was. Nonsense, of course: over the years I have gotten pretty good at color, but not by impersonating a Fauvist: rather, by learning to simplify and concentrating on atmospheres and hierarchies. 

Color investigation, like many design practices, is an essentially empirical activity: you try some things, you evaluate the results. Becoming attuned to visual experience requires observation and patience. (Students in our program will recognize this idea. It is for them primarily I am writing this, to provide a modest written record of tomorrow’s talk.) 

So: color, observation, and manipulation. 

First, a little description and terminology. The color radio has three dials: hue, value, and the variously titled chroma, saturation or intensityHue is the color of the color: red, or blue, or brown. Value is the relative lightness or darkness of the color. And saturation (the most literal term, so probably the best; alternatively, chroma or intensity) refers to the to the density of pigmentation in the color. A color with low saturation is a grayed-out version of a given hue: that is, a dull red, or a reddish gray. 

But the very terms I have just used—dull red, reddish gray—are extremely imprecise.

About a century ago, a combination of the Positivist spirit and the emergent need for an industrial language for color gave rise to a variety of color naming systems. Such naming systems aspire to define colors quantifiably: not simply how you make it, but what it looks like. Among the most influential of these systems was the Munsell System, created by landscape painter, art teacher and inventor Albert Munsell. The system was first explicated in A Color Notation in 1905. It remains in use today in various fields, and is one of several such systems recognized by the U.S. Bureau of Standards. 

Munsell identified the three dimensions of color as articulated above, using the terms hue, value, and chroma. (Saturation is probably the better term, but I have a hard time letting go of chroma, simply because I have used it for a long time.) 

Munsell specified five primary colors and five secondary colors: red, yellow, green, blue, and purple, interspersed between red-yellow, yellow-green, blue-green, blue-purple, and red-purple. Each of these colors was given ten values divided by ten segments, yielding a language like Red 2.5 to describe hue. Munsell’s hue circle is shown at right.

But of course hue provides pretty gross chunks of information: blue is notoriously variable. While Munsell’s numerical system for hues improves on this, the more subtle problems of value and chroma remain. (When I used to teach a lot of 2-dimensional design to first year students, I felt that I had done a bang-up job at teaching color if they left the course really understanding the difference between value and chroma.) 

Value for Munsell runs vertically, like a core sample through the middle of the earth. The hue circle rings the neutral core like a hula hoop. The neutral colors are black (at the bottom) white (at the top) and a stack of grays between the two. Values also have numerical coefficients, which are expressed as the numerator of a fraction following the hue notation. Values also run from 0 up to 10 (literally), but both extremes are theoreticals, so only 1 through 9 are used. 

Chroma values run out from the neutral core to the hue circle to fill out the structure. A given hue has a leaf, or slice that runs from the dullest possible version of the color at a given value out to its most saturated version at that value.

The trickiest part of this turns on the idea of spectrum value. Spectrum value refers to the value at which a given color is at its brightest or most saturated. Consider yellow. Yellow is very bright at light values; a dark yellow is really brown. Yellow has a spectrum value of 8, meaning that most vibrant yellow theoretically possible must occur at a value of 8. Red and blue have spectrum values of 5 and 3 respectively.

As a result, the color solid constructed by Munsell is really sort of a lumpy globe, bulging as it does at varying values.

Our concern with Munsell ends here. I don’t really care about the industrial specificity of color x or y; the system is useful insofar as it captures the relationship between the three dimensions of color. It especially helps to isolate the relationship between value and chroma.

Especially in a communication design context, hierarchy really matters. How is contrast used to manipulate emphasis? I’d argue that the single most basic issue for creating hierarchy in color turns on value relationships. This Jim Flora album cover from the forties demonstrates how value can be used to provide structure to a palette and to establish hierarchies. At right, a statement of the hues.

A straightforward primary triad from primary school days: red, yellow, blue. Plus white and black. Ho hum. But it turns out that the best part of this palette is really its value structure. (See analysis.)

Color schemes are used to establish sets of colors that may be used together successfully. I have suggested that an empirical approach will tend to work best;  I do not advocate a recipe-book attitude toward building palettes. But I raise the question for two reasons: 

1)I think that many people leave out chroma and value when they think about schemes, focusing exclusively on hues, often automatically resulting in garishly high-keyed color at close to spectrum value. Visual responsiveness continues to be advisable. Get a pile of acrylics. Generate a large sample set of colors to produce a library of colors with notations for how to reproduce them. Make sure you’ve got wide varieties of chroma and value, not just hue. [In class we get to these broad sample sets by mixing “ugly colors,” then “pretty colors.”] Trim them out to get chips. Then build some palettes: choose one color, then choose a second to go with the first, then a third to bridge the pair. 

2) When you get stuck, color schemes can help you out. If you need a three-color palette and you have two that you like, you can use schematic thinking to produce a new set of potential mates, assuming a triad or an analogous grouping or a split complement.

Above, find a animation still from Disney’s Sleeping Beauty. The color ranges widely across the wheel, but feels resolved and controlled. How does it work?

Here are a set of colors that play prominent roles in the image. They are arrayed on the hue circle. Value and chroma are as seen in the original image.

Here are the same colors adjusted to spectrum values and high chroma.

Here is an attempt to describe the basic structure of the palette. The red purples, blue purples, and yellow greens provide the most basic structural elements to the image. The blue greens tend to be very low chroma, and so do not play a significant role in the hue structure. But the low chroma colors provide an excellent low contrast setting for the other colors, and also help, via the use of atmospheric contrast, to create a sense of space.

Animation still from Gerald McBoing-Boing (1950), a landmark modernist short by the highly influential studio UPA, or United Productions of America. Gerald won the Oscar for best animated short in 1951.  

Alfred Munsell (circa 1905) a landscape painter and gifted standardizer of color notation in the early days of industrial color. 

My home-made translation of the Munsell Hue Circle. 

Wedges from a Munsell Color Solid. 

A crude drawing of chroma sections–more usefully, a diagram of how spectrum values work. 

A wicked lumpy grapefruit: the Munsell Color Solid (made somewhat confusing, in my view, by sudden accelerations of value in the blue/purple groupings below the equator.) 

Jim Flora, jacket design for Louis Armstrong's Hot Five, a 1928 jazz reissue by Columbia records (1947). 

Flora's palette arrayed on the hue circle sans adjustment of value or saturation. 

Schematic rendering of Flora's palette according to values, lightest to darkest, rotating clockwise. 

Procession scene, Sleeping Beauty, Walt Disney Studios, 1959. Eyvind Earle was responsible for background paintings and the palettes on this film. 

Eyvind Earle's Sleeping Beauty procession palette arrayed on the hue circle sans adjustment of value or saturation. 

The same palette, adjusted to spectrum values. 

Analysis of Earle's palette as a rough triad. Color schemes like triads, tetrads and analogous wedges are most useful as 1) descriptors and 2) get-out-jail-free cards. A little such analysis can get you out of trouble. Using schemes as recipes can yield somewhat tedious results.  

At some point I am going to write an extended post on the crappiness of contemporary commercial design (as compared to its mid-20th century equivalents, which were vastly better). When I do, the elements of decline will include the vector drawing versus the photostat, easily automated dumbass effects, and cheap, plentiful but totally undisciplined color. This example–an identity for a chain of convenience stores in Ohio–embodies all three regrettable developments: an irritating key drawing devoid of touch or dynamic contrast; a stupid drop shadow; a thoughtless palette of high-chroma color poorly differentiated in value. Just look at the earlier Marathon mark, which relies on simple, well-designed form and two hues at plainly different values. Specifically on the question of palette development: this was obviously, flippantly "picked" onscreen, not explored though mixing and patient comparison. Here's the thing: I would offer to redesign this goddamn thing for free, if I never had to gaze upon that stupid dog again. But of course everybody loves it, and it will never get redesigned. When passing through I will have to buy gas elsewhere, lest I fly into a rage.