How Human Vision Works: Why We See in RGB

The Photoreceptors of the Retina

The human retina contains two types of light-sensitive photoreceptor cells: rods and cones. Rods are more numerous (about 120 million per eye) and are responsible for low-light, monochromatic vision — they detect light intensity but not color. Cones (about 6–7 million per eye) are concentrated in the fovea at the centre of the retina and are responsible for color vision and high-acuity detail.

There are three types of cone cells in the human eye, classified by the range of wavelengths they respond to:

• S-cones (Short wavelength): Most sensitive to blue-violet light around 420 nm. Approximately 2% of all cones.
• M-cones (Medium wavelength): Most sensitive to green light around 534 nm. Approximately 32% of all cones.
• L-cones (Long wavelength): Most sensitive to yellow-green to red light, peaking around 564 nm. Approximately 64% of all cones.

The term "trichromat" — which describes normal human color vision — comes from this three-receptor architecture. Any color we see is encoded as a specific combination of activation levels across these three cone types.

How Cone Responses Encode Color

No single cone type uniquely identifies a wavelength — each cone responds to a broad range of wavelengths with varying intensity. A single cone firing more strongly tells the brain only that the light was somewhere in its response range, not which specific wavelength it was. Color is perceived through the comparison of signals from different cone types.

For example, light at 580 nm (yellow) strongly activates both L-cones and M-cones, with L slightly more than M, and barely activates S-cones at all. The combination of a high L response, a high M response, and a near-zero S response is what the brain interprets as "yellow." No single cone reports "yellow" — it emerges from the ratio of all three signals.

This is why RGB color mixing works: an RGB screen producing red + green light at equal intensities generates approximately the same L and M cone response as real yellow light at 580 nm, tricking the brain into perceiving yellow even though no 580 nm photons are present.

Opponent Process Theory

The cone signals do not travel directly to the visual cortex. They first undergo opponent-process coding in the retinal ganglion cells and lateral geniculate nucleus. In this system, colors are encoded as opposing pairs:

• Red vs. Green: A signal encoding how much more L-cone activation there is than M-cone activation (and vice versa)
• Blue vs. Yellow: A signal encoding how much S-cone activation differs from L+M combined
• Light vs. Dark: A luminance signal reflecting overall cone activation

This opponent-process system was proposed by Ewald Hering in the 19th century, in opposition to the Young-Helmholtz trichromatic theory. Both turned out to be correct: trichromacy describes the cone level, opponent processing describes what happens afterward. The visual system uses both mechanisms in sequence.

Opponent processing explains several perceptual phenomena:

• Afterimages: Staring at a red image then looking at a white surface produces a green afterimage — because the red channel is fatigued, the opponent green channel dominates temporarily
• Simultaneous contrast: A grey square looks greenish on a red background because the red background suppresses the red opponent channel, shifting the grey toward green
• Why there is no "reddish-green": Red and green are opponents — they cannot be simultaneously activated in the same channel, which is why we have no intuitive concept of a color that is both

Color Blindness: Missing a Cone Type

Color blindness occurs when one or more cone types are absent, have reduced sensitivity, or are shifted in wavelength sensitivity. The most common forms affect the L and M cones, which are both encoded on the X chromosome — explaining why color blindness is more common in males (1 in 12) than females (1 in 200).

• Deuteranopia: Missing M-cones (green cones). Affects approximately 6% of males.
• Protanopia: Missing L-cones (red cones). Affects approximately 2% of males.
• Tritanopia: Missing S-cones (blue cones). Rare, affects fewer than 1 in 10,000 people.
• Achromatopsia: All cones absent or non-functional. Very rare; results in monochromatic (black and white) vision.

"Color blind" people typically have two functioning cone types rather than none — they are dichromats rather than monochromats. They can still distinguish many colors, but cannot distinguish colors that differ only in the wavelength range of their missing cone type. Red-green confusion is the most common result.

Tetrachromacy: Four Cone Types

Some women carry an X-linked genetic variant that gives them four distinct cone types rather than three — a condition called tetrachromacy. Because the additional cone type is typically a variant of the L or M cone tuned to a slightly different wavelength, tetrachromats can in theory distinguish colors that appear identical to trichromats.

Research by Gabriele Jordan and colleagues at Newcastle University has documented functional tetrachromacy in a small number of subjects. Whether tetrachromats actually perceive a richer color experience in everyday life remains under study — the brain must develop the neural machinery to decode the fourth channel, which may require specific visual experience to establish.

The Visual Cortex and Color Perception

The opponent-process signals from the retina travel through the optic nerve to the lateral geniculate nucleus (LGN) in the thalamus, then to the primary visual cortex (V1) and onward to specialized color-processing areas (V4 and the inferior temporal cortex). The visual cortex does not simply record the raw cone signals — it interprets them in context, applying color constancy corrections that allow the brain to perceive an apple as "red" whether in direct sunlight or dim indoor lighting, even though the wavelengths reaching the eye are quite different.

Color perception, in short, is not a simple detection of wavelengths — it is a complex inference process that takes into account the illumination, the surrounding colors, and the viewer's experience and expectations. This is why visual illusions like the famous "dress" photograph (seen as blue-black or gold-white by different people) can exist: the visual cortex is making different assumptions about the illumination, and color perception follows those assumptions rather than the raw light data.

Generate colors →