How the Eye Sorts Color

Three receptors, two arguments

Most people see color through three kinds of cone cells in the retina, sensitive to long, medium and short wavelengths of light. This trichromatic idea, developed by Thomas Young and Hermann von Helmholtz in the nineteenth century, explains why three channels — red, green and the third primary of a screen — can mimic most visible colors.

Ewald Hering added a second, complementary insight: the signals from those cones are compared in opposing pairs. Red is weighed against green, yellow against its opposite, light against dark. That is why we can imagine a reddish yellow but not a reddish green. For photographers, the opponent model is quietly practical. It explains why complementary pairs feel charged and why afterimages appear in the opposite hue after staring at a saturated patch.

Value is the skeleton

Neuroscientist Margaret Livingstone, in Vision and Art: The Biology of Seeing (2002), describes how much of our perception of depth, motion and spatial layout depends on luminance — lightness contrast — rather than hue. Two colors of equal lightness can be vividly different and still seem to vibrate or float, because the part of the visual system responsible for edges and depth sees them as nearly the same.

This is an established principle, not a stylistic opinion. It is why the grayscale test works. Convert an image to monochrome and look again: if the subject still separates from the background, the structure is sound. If it dissolves, the image is relying on hue alone, and that reliance will fail on poor screens, in certain prints and for many viewers.

Neighbors repaint each other

In 1839 the chemist Michel Eugène Chevreul, working on dyes at the Gobelins tapestry works in Paris, published his law of simultaneous contrast. He observed that colors placed side by side appear to shift away from each other: a grey beside red takes on a greenish cast; the same grey beside green looks faintly warm. Josef Albers turned this into a classroom method at Yale, showing students how one color can be made to look like two, and two like one.

For image-makers, the lesson is humbling. You cannot judge a swatch in isolation. A skin tone that looks right in a neutral room may look sallow beside a saturated wall. A background grade can shift the perceived color of a subject without touching the subject at all.

Constancy and its failures

Our brains try to keep object colors stable under changing light. A white shirt looks white at noon and at sunset, even though the light reflected from it changes dramatically. Edwin Land, founder of Polaroid, explored this through his retinex theory in the 1970s. Constancy usually helps us; it is also why cameras need white balance and our eyes do not seem to.

The viral photograph of “the dress” in 2015 showed what happens when the cues fail. Viewers disagreed sharply about its colors because they made different unconscious assumptions about the light falling on it. Ambiguous illumination produces ambiguous color — something a photographer can avoid, or deliberately exploit.

Designing for every viewer

Color vision deficiency affects roughly 8 percent of men of Northern European descent and around 0.5 percent of women. The Ishihara plates, first published in 1917, remain a familiar screening test. The most common forms reduce the ability to distinguish reds and greens.

• Make sure critical separations survive in grayscale.
• Pair hue differences with value or texture differences.
• Check charts, infographics and captions where color carries information.

A short field practice

Before your next edit, look at the frame for three seconds and note where your eye went first, second and third. Then ask whether that sequence tells the story you want. If not, change the relationships — value, saturation, proportion — rather than reaching for a new hue. Perception is the grammar; the story is what you write with it.