Munsell principal and intermediate hues, digital illustration by Fred Hatt
I am a person of serial obsessions. Every few years I feel compelled to learn everything I can about some topic, usually something esoteric or scientific. Around 2003-2005, my obsession was color: the science of light and spectra, the biology and psychology of color perception, the technology of color reproduction, ways of naming colors and dividing color space, and philosophical ideas about color. When I had the idea of writing a blog post about color, I started looking through my notes and collections of digital images, making a list of interesting things I’d learned. There was enough there for a book or a semester course! Perhaps in the future there will be more posts on color. For now, I’ve selected a few interesting or lovely images from my collection, and here present them with interesting related factoids. Even if you don’t share my hunger for knowledge about color, I hope you’ll appreciate the beauty of these diagrams.
I’m titling this post “The Full Gamut” – we’ve all heard that phrase meaning the complete range of something that has varieties. The word gamut originally meant a range of musical notes. It’s used in color science to indicate the limited range of colors that can be described or reproduced given a certain technological context. A computer monitor, for example, can simulate many colors by combining various intensities of red, green, and blue “primary” colors. The surface colors of most naturally occurring objects can be reproduced, but there remain many colors outside the gamut of the monitor. You can see pure spectral colors by looking at the reflections on a CD or DVD. The colors in the image at the top of this post approach the limits of saturation achievable on a monitor, but compared to pure spectral colors they’re surprisingly dull. Even Newton’s prismatic spectrum does not contain the full range of vivid colors – magentas and purples cannot be represented by single wavelengths, but only exist as the blending of the opposite ends of the spectrum.
Color is a three-dimensional phenomenon. Every model for describing colors requires three variables: three primaries, or three polarities. For a general understanding of color independent of any particular medium or technology, the clearest dimensions are hue, value (lightness or luminance), and chroma (saturation or intensity). Albert Munsell’s model of color space is one of the most illuminating systems, based on rigorous study of human color perceptions rather than on physical or technological variables. In Munsell’s system, value is the vertical dimension, hue is the angular dimension, and chroma is shown as the distance from the center. The resulting arrangement of colors is called a color solid, or a color tree.
The Munsell colors are produced in rigorously accurate sets as books and charts to be used to describe colors by visual reference to standard samples. They come very close to representing the full range (gamut) of colors that can exist in the form of physical objects.
Munsell Color Tree, illustration by limaorian@hotmail.com
The “color wheel” most people are taught in basic art classes is a rigid and simplistic model compared to Munsell’s color solid. The color wheel doesn’t account for the fact that different hues have different ranges of chroma or intensity, and that some hues (e.g. yellow) achieve their highest chroma at high values, while other colors (e.g. bluish purple) are more intense at a darker value. Munsell’s system defines the hues by letters and numbers, starting with five fundamental hues (red, yellow, green, blue, and purple), and five secondary or intermediate hues (yellow-red, green-yellow, blue-green, purple-blue, and red-purple). The diagram below shows five cross-sections of the Munsell color solid, with the principal hues on the right and the complementary intermediate hues on the left.
Five cross sections of Munsell Color Solid, digital illustration by Fred Hatt
Here are the most saturated colors around the perimeter of the Munsell Solid. Here, the hue circle is repeated twice along the horizontal axis with the values arranged on the vertical axis.
Munsell hues at maximum chroma, digital illustration by Fred Hatt
If we consider the color solid as a kind of globe, with the neutral grays as the axis, we can look at the irregular shape from a point of view centered above the north (white) pole or the south (black) pole. The colors with maximum chroma are at the outer bound of these polar views, whether they are on the “equator” (middle value perimeter) or not. Please note that the gamut of the computer monitor is considerably smaller than the gamut of the physical samples included in the Munsell standard, so the colors closer to the outside edge of the figures below are not really accurate. You can see that the colors yellow and green achieve high chroma at the higher values, while deep blues and purples are most intense at low values.
Light and dark hemispheres of the Munsell color solid, digital illustration by Fred Hatt
Some of the transitions between adjacent colors in the illustration above may seem abrupt, but that’s because of variations in the maximum achievable value or chroma. If we look at the full range of hues at a uniform value and chroma level, as in the circle below, the transitions are very smooth.
40 Munsell hues at value 7, chroma 8, digital illustration by Fred Hatt
This circle is at value 7 and chroma 8, the maximum chroma level achievable all the way around the hue circle at any value in the Munsell solid. We probably all learned in school that Newton proved that light is a waveform, and that different colors are different wavelengths of light. The diagram below charts the level at which the Munsell samples, at the same chroma and value seen in the above illustration, reflect various wavelengths of the spectrum. The horizontal axis goes from short wavelengths (violet blue) at the left, to long wavelengths (red) at the right. You will notice that even these samples, which appear quite vividly colored, are all reflecting almost half the spectrum at over half their average reflectivity. These colors are not “pure”, but they do look intense!
Spectral reflection curves for five principal Munsell hues at value 7, chroma 8, illustration from article by A. Kimball Romney and Tarow Indow
Munsell’s model arranges colors by measures of equal perceptual distance, but what does that have to do with how we learn to identify and name colors? One of the most cited academic papers of all time is Berlin and Kay’s cross-cultural survey of color names. Berlin and Kay used a study of color terms to address the question of linguistic relativity, that is, whether linguistic categories define perceptions, or vice versa. They used the highest-chroma Munsell samples of colors at the full range of hues and values, asking participants of various linguistic and cultural backgrounds to choose the “best examples” of their basic color words, and the range these words would cover. The “best examples” were called “focal colors”. In the diagram below, the focal colors are marked as chosen by speakers of American English.
American English focal colors in a Munsell grid, based on data from Basic Color Terms: Their Universality and Evolution, 1969, by Brent Berlin and Paul Kay, digital illustration by Fred Hatt
Berlin and Kay found a high degree of uniformity in the specific colors chosen as focal colors between speakers of different languages. They also found evidence that color terms evolve in a given language in a predictable order. First, a distinction is made between dark/cool and light/warm. Red is the first individual color to be given a name. Next, green or yellow are distinguished, followed by blue. More complex languages separate brown, purple, pink, orange, and gray. Berlin and Kay’s findings have been challenged and reproduced by many subsequent researchers, using the same Munsell grid. The chart below shows interesting variations on how the color range can be divided, with eight divisions in English and five in a language called Berinmo.
Distribution of English and Berinmo color names, illustration from the article "Colour categories in a stone-age tribe", by Jules Davidoff, Ian Davies and Debi Roberson, Nature 398, 1999
Below are pretty close representations of the “focal colors” chosen by English speakers. There are eleven basic color terms in English, the eight easily identifiable ones shown below, plus black, gray, and white. Chosen samples of focal colors would be very similar for nearly every language in the industrial world. Why are these colors seen as basic? They are not evenly distributed on the grid of colors, and no one, as far as I know, has been able to show any fundamental relation between these specific colors and any measurable aspect of color vision or color physics.
Focal colors, digital illustration by Fred Hatt
You’ll notice that people make finer distinctions in the colors around the red/yellow portion of the range. Human skin color and the colors of most animals are in this area, so perhaps we are more attuned to fine differences there than we are in the blue and green areas associated with the landscape.
Randall Munroe, author of the classic geek webcomic XKCD, conducted an online color-naming experiment, with a random color generator that asks random web participants to name the colors they see. His report on the results of the survey is hilarious as well as interesting. Here’s his map of how thousands of participants intuitively divided up the color space.
Dominant color names mapped to RGB cube, illustration from XKCD Color Survey Results, from Randall Munroe's XKCD blog
Color naming experiments are usually done by showing subjects one color at a time. When the colors are shown together, as in the chart above, or in the Munsell grid illustrating the Berlin and Kay survey, we notice the arbitrariness of the lines we draw to distinguish colors.
Color perception is a relativistic phenomenon. The book Interaction of Color, by the painter and teacher Josef Albers, shows by example how colors are seen differently according to their surroundings. In the illustration below, the double-x line looks very different depending on its background, but where the line joins we can see its continutiy.
Illustration from Interaction of Color, 1975, by Josef Albers
As an artist, I find it most useful to look at colors as polarities tending one way or another. Many real-life colors are very muted and subtle, but if you can look at a shadow and see, for example, that it has a bluer tone compared to the adjacent highlight’s yellower tone, you can begin to capture those subtleties.
In photography, it is common to analyze and correct colors using such polarities. The most important one is the color temperature axis, what most painters would describe as the warm/cool distinction. In the study of light, it was observed that the temperature of any incandescent substance, such as a heated piece of metal, could be determined by the color of its glow. White hot is hotter than red hot, and blue hot is hotter still. Color temperature is a scientifically defined scale for describing the color of light on a red/orange/white/blue scale. Typical incandescent lights glow at 2500-3200 degrees kelvin, while daylight is 5000-7500 degrees. The temperature-color correspondence is exactly the opposite of what is taught to artists as warm and cool colors.
Color temperature illustration, from a webpage by W. A. Steer, PhD
Of course, fluorescent lights, neon lights, high-intensity discharge lamps, LEDs, and other non-incandescent sources aren’t defined by the color-temperature scale, so correcting colors from those lights involves a second scale, which photographers call “tint” or “plus green” and “minus green”. Minus green is magenta or pink. A minus green filter, for example, can overcome the tendency of fluorescent lights to photograph as greenish. These two axes, orange-blue and green-magenta, are used in filtering for lenses or light sources while shooting, and in digital post-processing of photographs and video recordings.
In figurative art, I’m always looking at the variations in flesh tones. I find it useful to look at these very subtle differences as tendencies along axes of complementary colors: orange/blue, magenta/green, red/green blue, yellow/deep blue.
Eight part color arrangement, digital illustration by Fred Hatt
Some of the illustrations in this post are my own, and others are found on the web. Clicking on found images links to the site where I found them. For the Munsell colors used in some of the digital illustrations I am indebted to Wallkill Color for their Munsell Conversion Software.