How turning colour into numbers lets billions of monitors agree on a shade

How billions of screens agree on one shade

A new essay titled "Quantifying Colour", published on ekunazanu.foo and surfaced on the Hacker News front page, opens with a fact that is easy to take for granted: the billions of monitors in use worldwide can reproduce exactly the same colour on command. The author calls this a remarkable feat of engineering, but one that is only possible because colours were first given a precise, standardised numerical definition — something language alone could never supply.

Words are too coarse for the job

Most languages, the essay notes, contain at most twelve words for colour. That vocabulary suffices in everyday conversation but collapses when precision matters. A single label such as "green" covers an entire family of visibly different shades, and patching the gap with compound names — lime-green, olive-green, light-green, dark-green — produces a list that keeps growing while still failing to capture fine differences between similar-looking colours.

The alternative the essay points to is to treat colour like any other measurable physical quantity, such as distance or temperature: assign it numbered units, and gain or lose precision simply by using more or fewer digits. The catch is that colour, as a perceptual experience, first has to be tied to something physically measurable.

Light, measured wavelength by wavelength

That anchor, according to the author, is light. Photons carry energy determined by their wavelength, and the full range of possible wavelengths forms the electromagnetic spectrum. Photon energy can be measured directly, and those measurements are normalised — per unit of time as power, and per unit of the radiating body's area as intensity.

Rather than collapsing a light source into a single intensity figure, the essay explains, light is characterised by many intensity values: one per wavelength. The resulting intensity-per-wavelength profile is called the spectral power distribution. Because a photon's energy is inversely proportional to its wavelength, two bands emitted with the same number of photons can still register different intensities — short-wavelength photons each carry more energy.

The spectral power distribution quantifies light, but it says nothing yet about colour. Bridging that gap requires the other half of the phenomenon: the eye.

The eye as the reference instrument

Human eyes contain two broad classes of photoreceptors — rod cells and three types of cone cells — each with its own sensitivity across wavelengths. The essay points out that these sensitivities track the sun's output, roughly 380nm to 750nm, which is no coincidence: vision evolved under sunlight.

When light reaches the eye, each cone type responds according to its own sensitivity curve, so any single wavelength produces a distinct combination of responses across the three cone types. Because the curves are distributed the way they are, every wavelength maps to a unique response set. The brain has evolved to interpret each unique set as a distinct colour, while nearby wavelengths, which trigger similar responses, are perceived as similar colours.

The essay cautions that commonly shown sensitivity curves are normalised approximations. Real cone responses are less smooth, the S-cones are considerably less sensitive than the other cones, and rods outperform all of the cones in low-light sensitivity.

The questions quantification unlocks

With that machinery in place, the author frames the follow-on puzzles the rest of the piece tackles: what adding or multiplying colours would even mean, why the familiar 16,777,216 values of colour hexcodes still cannot represent enough colours, how colour blindness fits into the cone-response model, and how a single photographed dress became an internet-wide argument about its colour.

Why it matters

Hexcodes are the colour currency of the web, and every pixel a browser paints rests on this chain: photons measured as spectral power distributions, standardised reference data for human cone responses, and numeric encodings handed between devices. Understanding the chain explains why identical colour reproduction across billions of displays works at all, where it breaks down — for colour-blind viewers or under limited encodings — and why graphics and display engineering keeps refining the standards that carry colour from physics to screen.