30

What Color Is and How Humans See It

Color is not a property of light, nor a property of matter. It is the experience your visual system constructs when photons strike a few square millimeters of living retina. This article is the interactive starting point for everything else on color science: physics, biology, perception, and the standards that connect them.

Foundations 13 Live Demos ~45 min read Beginner → Expert
380-780
nm Visible Range
3
Cone Types (L M S)
~120M
Rods per Retina
~7M
Cones per Retina
01

What is color, really?

Pose the question to a physicist, a biologist, and a painter, and you will get three answers that do not quite line up. The physicist will speak of wavelength and spectral power distribution. The biologist will describe photoreceptors, opponent channels, and cortical pathways. The painter will talk about hue, chroma, and value, and may insist that color is whatever you feel when you stand in front of a Rothko at the right time of day.

None of them is wrong. Color sits at the intersection of three independent layers: a physical stimulus, a biological response, and a perceptual experience. A complete understanding requires all three. This article walks the chain end to end.

Stimulus
A spectral power distribution (SPD): how much energy a beam of light carries at each wavelength between roughly 380 and 780 nanometers.
Response
Three cone types in the retina absorb photons and report L, M, and S signals. The brain combines these into red-green and blue-yellow opponent channels.
Experience
What you actually see: hue, brightness, saturation, the warmth of skin, the coolness of a winter sky. This is the part that varies by context, observer, and culture.
Working definition: Color is the perceptual consequence of how a spectral stimulus interacts with a particular visual system under particular viewing conditions. Change any of those three things - stimulus, observer, surround - and the perceived color can change too, even when the physical light is identical.
Advanced lab · Observer workbench

Run the whole color signal chain from stimulus to percept

This cockpit compresses the article into one live model. Change the spectral profile, adaptation white, surround, and observer type; the canvas redraws the stimulus, cone catches, opponent channels, and approximate final color.

Perceived color #ccff44
Hue and photon math yellow-green 2.235 eV, 540.2 THz
Cone catches L 0.99 / M 0.88 / S 0.00
Opponent channels Lum 0.94 / RG 0.11 / BY -0.94
Chromaticity estimate x 0.379 / y 0.578
A single wavelength is easy to name, but the observer is doing the real compression: many possible spectra collapse into three cone catches and then into opponent signals.
02

The visible spectrum

Visible light is a tiny window in the electromagnetic spectrum. The same family of waves that carries radio broadcasts at the slow end and gamma rays at the fast end produces, between roughly 380 nanometers and 780 nanometers, the band our cones can detect. The slow side is red; the fast side is violet. Everything we call "color" lives in that narrow octave.

Interactive 01 · Wavelength explorer

Slide a single wavelength across the visible band

Drag the slider to scan a monochromatic light source through the visible spectrum. The patch shows the approximate appearance of that single-wavelength beam, with the nanometer reading and rough perceived hue name underneath.

380 450 500 555 600 680 780
yellow-green
RGB approx: 204, 255, 68 Energy: 2.235 eV Frequency: 540.2 THz
Important caveat: almost no light in everyday life is monochromatic. Sunlight, an LED bulb, a phone screen, a printed page reflecting room light - all of these are mixtures of many wavelengths. The slider above shows the colors of ideal laser-like beams. Real color is what happens when whole distributions of wavelengths reach the eye at once.
Interactive 02 · SPD composer

Build a light source from three colored components

Real light sources have a spectral power distribution - energy spread across many wavelengths. Mix three component bands to see how their sum (the SPD curve) relates to the eye's response and the resulting perceived color.

Resulting color: #bba366
CCT estimate: 3400 K
03

Light meets matter

A photon leaves the sun. Eight minutes later it strikes an apple. What happens next decides the apple's color. Most of the visible-light photons hit the apple's skin and are absorbed. The few that survive - mostly in the long-wavelength range around 600-700 nm - are reflected diffusely, scatter into your eye, and you see "red." Change the surface chemistry, and the absorption profile changes. Change the light source, and the photons available for reflection change too.

Absorption
Photons transfer energy to the material's electrons and are lost (usually as heat).
Reflection
Photons bounce off a surface, either specularly (mirror-like) or diffusely (scattered in all directions).
Transmission
Photons pass through a translucent or transparent material, sometimes bending (refraction) along the way.
Scattering
Photons interact with small particles and change direction without being absorbed - the reason the sky is blue.
Emission
A material releases photons of its own, as in stars, displays, and fluorescent dyes.
Fluorescence
A material absorbs short-wavelength photons and re-emits longer ones, making whites and certain dyes appear unusually bright.
Interactive 03 · Surface × illuminant

Watch an object's color change with the light source

A surface has a fixed reflectance curve - the fraction of light it returns at each wavelength. But the light reaching your eye is reflectance × illuminant. Swap the illuminant and the same surface produces a different stimulus.

Appearance
Approx hex: #c12b2b
Stimulus curve
— surface — illuminant — product
Why this matters: the apple is not "red" by itself. The apple's surface is "the set of reflectance curves that, multiplied by typical illuminants, hit the eye in a way that most people call red." Change the illuminant enough and the apple really does look different. We have all seen this in a parking-lot sodium-vapor lamp, where every color collapses toward yellow.
04

The human eye

The eye is an optical sensor with biological glass. Light enters through the cornea, narrows through the pupil, focuses through the lens, and lands on the back wall - the retina - where photoreceptors absorb photons and start the signal that will eventually become color.

Interactive 04 · Eye anatomy

Hover any labeled part to see its role

Move over the diagram. Each labeled element of the eye contributes to forming a focused, correctly exposed image on the retina where color processing begins.

Cornea Iris Pupil Lens Retina Fovea Optic n.
Hover a labeled part to read about its role in vision.
The fovea is a depression about 1.5 mm wide at the back of the retina. It is packed almost exclusively with cones, contains no rods, and is where your sharpest, most color-accurate vision occurs. When you "look at" something, you are aiming its image at this tiny region - which is why peripheral vision is so poor at telling colors apart in low light.
05

Rods, cones, and the retina

The retina contains two families of photoreceptors. Rods are highly sensitive but color-blind - they support vision at extremely low light levels and do not contribute to color perception. Cones are less sensitive, but come in three types - L, M, S - tuned to different parts of the spectrum. Color vision arises from the comparisons between these three cone signals.

~120M
Rods / retina
~7M
Cones / retina
~64%
L cones
~32%
M cones
~4%
S cones

Each cone contains a photopigment whose absorption peaks at a specific wavelength. The L cone ("long") peaks near 564 nm, the M cone ("medium") near 534 nm, and the S cone ("short") near 420 nm. The names "red, green, blue cones" are useful shorthand but technically inaccurate - the L cone is most sensitive to yellow-green, not red.

Interactive 05 · Cone response

How three cone types respond to any monochromatic light

Drag the wavelength marker through the visible spectrum. The bars and the colored vertical line show how strongly each cone type fires for that single wavelength. Color vision is your brain reading the ratio between these three numbers, not any one of them.

L cone · 0.90
M cone · 0.85
S cone · 0.06

Cones are concentrated at the fovea, the tiny central pit on the retina that handles your sharpest vision. Rods dominate everywhere else, which is why peripheral vision is more sensitive in dim light but worse at distinguishing color and fine detail.

Photopic vision
Daytime, cone-dominant vision. Full color, sharp detail. Active above roughly 3 cd/m².
Mesopic vision
Twilight, mixed cone and rod activity. Color desaturates and reds darken first (the Purkinje shift).
Scotopic vision
Night, rod-only vision. No color, peak sensitivity around 507 nm (cyan-green).
06

Trichromatic theory

If you only have three types of color sensor, then every perceivable color can - in principle - be matched by mixing three appropriately chosen primaries. This is the trichromatic theory, proposed by Thomas Young in 1802 and later refined by Helmholtz. It is why a phone screen can show you almost any color using only red, green, and blue subpixels.

Mathematically: any color the eye can see can be written as C = r·R + g·G + b·B, where R, G, B are three independent primaries and r, g, b are weights. The constraint that any color falls into a three-dimensional space - rather than a continuous spectrum - is the most consequential fact in all of color science.

Interactive 06 · Additive RGB mixer

Mix three primaries to produce any visible color

Adjust the red, green, and blue lights. Three overlapping circles show additive mixing - light adds together to form brighter, lighter colors. Click any swatch below to jump to that mixture.

Mixed color: #ffc83c
Subtractive mixing is the opposite world. When paints or inks combine, each pigment subtracts wavelengths from the light hitting it. Mixing all pigments approaches black. This is why printing uses cyan, magenta, and yellow (the complements of red, green, and blue) plus black - a different mathematics for a different physical situation.
07

Opponent process theory

Trichromatic theory explains how three sensor types capture color information at the receptor level. But the experience of color does not feel three-dimensional - it feels like three opposing axes: red versus green, blue versus yellow, and black versus white. You can imagine a reddish-yellow (orange) or a bluish-green (teal), but you cannot imagine a reddish-green or a bluish-yellow. These pairs are opponents.

Ewald Hering proposed this in the 1870s, in apparent opposition to Young-Helmholtz. Modern neuroscience reconciles them: trichromatic encoding happens at the cones, and the retinal ganglion cells immediately recombine those three signals into opponent channels: L − M for red/green, (L + M) − S for blue/yellow, and L + M + S for luminance.

Interactive 07 · Afterimage demo

Watch your opponent channels exhaust themselves

Stare at the central dot for 20 seconds without moving your eyes. Then press the button: the colored circle vanishes onto a neutral background. The afterimage you see is roughly the opponent color of what you were just staring at - direct evidence that opponent channels in the retina are temporarily adapted.

Stare at the small black dot · do not move your eyes

Red-green channel
L − M comparison. Discriminates reds from greens. A defect here is the most common form of color vision deficiency.
Blue-yellow channel
(L + M) − S comparison. Separates warm yellows from cool blues.
Luminance channel
L + M (and a touch of S). Carries brightness information at high spatial frequencies - the part of vision that resolves text, edges, and detail.
Unique hues
The four hues that look "pure": red, green, blue, yellow - the endpoints of the two chromatic opponent axes.
08

Color constancy and adaptation

A white shirt looks white in noon sunlight, under tungsten bulbs, and at sunset - even though the actual spectrum reaching your eye is wildly different in each case. The visual system discounts the illuminant and tries to recover the underlying surface property. This is color constancy, and it is imperfect by design: getting it exactly right would require knowing the illuminant, which the eye cannot directly measure.

Interactive 08 · Simultaneous contrast

The same gray looks different depending on its surround

The center patches are identical. The only thing that changes is the background. Your visual system rescales lightness and chroma against the local surround, producing two different perceived grays.

On dark surround
On light surround
Interactive 09 · Chromatic adaptation

The same object under different illuminants

Hover or tap the swatches to scrub through a sequence of illuminants. Notice how the perceived "color of the apple" stays stable even though the raw pixels shift dramatically. This is your brain attempting (imperfectly) to keep object color constant across lighting conditions.

Apple under Tungsten 2856 K
Adaptation is automatic and selective. When you walk from a sunlit garden into a warm living room, your visual system rebalances its chromatic gains over a few seconds. The world stops looking yellow. But cameras do not do this for free - which is why "white balance" has to be set explicitly. It is a software emulation of biological chromatic adaptation.
09

Color vision deficiency

Roughly 8% of men and 0.5% of women have an inherited form of color vision deficiency, almost always X-linked. The most common forms affect the red-green channel and arise when one cone type is missing, shifted in sensitivity, or non-functional. For designers, accessibility engineers, and educators, simulating these conditions is the fastest path to inclusive color choices.

Interactive 10 · CVD simulator

See common color palettes through a deficient observer

Pick a deficiency model. Each row shows the same palette as it would appear to a typical observer of that type. Tips: deutan and protan losses collapse reds and greens; tritan collapses blues and yellows; monochromacy collapses everything to brightness.

12-step rainbow + accessibility hues
Normal vision
Simulated Protanopia

Protanopia

L cone absent. Reds appear dark and muddy; reds and greens confuse easily.

Deuteranopia

M cone absent. Most common form. Greens shift toward beige; reds shift toward yellow-brown.

Tritanopia

S cone absent. Very rare. Blue/yellow distinctions collapse; greens look bluish.

Achromatopsia

All cones missing or broken. Only rod vision; world appears in shades of gray.

Anomalous trichromacy

One cone present but spectrally shifted. Color discrimination reduced but not absent.

Tetrachromacy

A theoretical fourth cone type in some women carrying CVD genes. Evidence remains contested.

Designer's rule: never encode information by color alone. Pair color with shape, label, position, or pattern. WCAG contrast requirements exist because the difference between two hues that look obvious to a trichromat may be invisible to roughly one in twelve men.
10

The full perceptual pipeline

We have now walked each link of the chain. Putting them together produces a remarkably consistent picture: photons leave a source, interact with matter, reach the retina, become electrochemical signals, and finally arrive at the visual cortex as the experience of color.

01

Source

A light source emits a spectral power distribution (SPD) - photons spread across wavelengths. Sun, bulb, screen, candle - each has its own SPD.

02

Surface or medium

The SPD interacts with matter. Some photons are absorbed; others are reflected, transmitted, or scattered. The surviving spectrum carries the surface's reflectance signature.

03

Optical entry

Light enters the eye through the cornea, passes through the pupil and lens, and forms a focused image on the retina.

04

Photoreceptor absorption

L, M, and S cones (plus rods) absorb photons. Each cone reports an integrated response based on its own spectral sensitivity curve.

05

Retinal recoding

Bipolar and horizontal cells recombine the three cone signals into red-green, blue-yellow, and luminance opponent channels before the data even leaves the eye.

06

Optic nerve transit

Signals travel through the optic nerve to the lateral geniculate nucleus (LGN) in the thalamus, where each channel takes its own pathway.

07

Cortical processing

Primary visual cortex (V1) and area V4 in the temporal lobe map color, integrate with shape and motion, and apply context-dependent adjustments.

08

Adaptation & context

Chromatic adaptation, simultaneous contrast, memory color, and attention modulate the final percept. The same spectrum can look like two different colors depending on what else is present.

09

Conscious percept

You experience hue, brightness, and saturation as a unified, named color. This is the only step you can introspect on - and the one all the others were designed to produce.

11

Why we needed standards

If color is observer-dependent, then how can a paint factory in Milan match a swatch designed in Tokyo? The answer is the system of colorimetric standards built up by the CIE (Commission internationale de l'éclairage) starting in 1931. By defining a standard observer - an averaged spectral sensitivity curve representing "normal" human vision - and a set of standard illuminants, the CIE created a shared coordinate system for color.

CIE 1931 2° standard observer
Color matching functions x̄(λ), ȳ(λ), z̄(λ) representing average vision in a small central field. Still used everywhere today.
CIE 1964 10° observer
A second set of matching functions for larger fields - more typical of paint or fabric viewing.
D65 illuminant
A standardized "average daylight" spectrum used as the default reference white in computing, displays, and the sRGB color space.
D50 illuminant
A warmer daylight reference used for print and ICC color management.
CIE XYZ
Tristimulus space derived from the standard observer. Every other modern color space - sRGB, Lab, LCh, Rec.2020 - is defined as a transform of XYZ.
ΔE (Delta E)
A family of metrics that quantify how different two colors look. The first numerical answer to "how close is close enough?"

These standards turn color from a private experience into a measurable, communicable, and manufacturable quantity. The remaining 112 articles on this site assume that bridge is in place - and start exploring the territory on the other side.

12

Test your understanding

Six questions, instant feedback, no scores recorded. Answer one wrong and the explanation tells you where to look. Answer them all and you have the foundation needed for the rest of the color science library.

Quick check

Loading…
 
Question 1 of 6
13

Continue your journey

Each article below dives into a specific part of the chain you just walked. Read them in any order. The numbers reflect their position in the editorial roadmap of 113 articles, but nothing here strictly depends on what came before.

Foundations · 49

The Vocabulary of Color: Hue, Saturation, Value, Tint, Tone, Shade

The precise words for the color attributes introduced here.

Foundations · 38

Color Naming and Categorical Perception

How language carves the continuous spectrum you just learned into named buckets.

Foundations · 02

The History of Color Science from Newton to Hering

How Newton's prism, Young's three receptors, and Hering's opponents built modern color thinking.

Physics · 03

The Physics of Light, Wavelength, and Spectrum

What visible light is, where it sits in the electromagnetic spectrum, and how energy relates to color.

Physics · 04

Spectral Power Distributions and Why RGB Is Not Enough

How real light sources actually look as energy-per-wavelength curves, and the limits of tristimulus.

Vision · 05

Human Color Vision: Cones, Opponent Signals, and the Brain

A deeper biological tour of receptors, retinal recoding, LGN pathways, and cortical color processing.

Vision · 06

Color Blindness and Color Vision Deficiency

Inherited and acquired forms, screening tests, and the design implications of CVD.

Vision · 07

Color Constancy, Adaptation, and Why Colors Change with Context

How the brain holds object color stable across illuminants - and the elegant ways it fails.

Colorimetry · 08

CIE XYZ Explained

The universal tristimulus reference space and how every other modern color space derives from it.

Colorimetry · 09

CIELAB and LCH Explained

Perceptually uniform spaces, hue angle and chroma, and the foundation of ΔE.

Colorimetry · 10

Chromaticity Diagrams Without the Intimidation

The famous horseshoe: spectral locus, white point, gamut boundaries, and the line of purples.

Colorimetry · 11

Metamerism Explained

Why two different spectra can match perfectly under one light and fail under another.

Digital · 12

RGB, sRGB, Adobe RGB, ProPhoto, Display P3, Rec.2020

The canonical guide to digital RGB spaces, gamut, and when each one actually matters.

Design · 13

Accessible Color Design and WCAG Contrast

Contrast thresholds, exceptions, and how to apply them without breaking your brand.