How to improve your colour perception

What can change, and what cannot

Colour vision begins with three kinds of cone cell and ends with a judgement. The cones are genetics; their number, their sensitivity and whether all three types are present cannot be altered by looking harder, and someone with a red-green deficiency will have it after a year of practice exactly as before. This is worth saying plainly because some products imply otherwise.

Everything after the cones is learned, and learning is what perceptual learning research has documented for sixty years: adults improve at fine sensory discriminations with practice, the improvement lasts, and it is strongest for the specific judgement trained. Colour is no exception. People who work in colour-critical trades score better on hue-arrangement tests than people who do not, and it is commonly reported that scores on the Farnsworth-Munsell 100 Hue Test improve with repeated attempts — although part of that is learning the test, which is why a training method needs new targets every time.

Two separate skills are trainable. Discrimination is telling apart two colours that are both in front of you. Colour memory is holding a colour in mind after it is gone and reproducing it — the skill a designer uses when choosing a swatch to match a wall seen that morning, or a colourist uses to keep a grade consistent across a cut. Memory is the one that is most obviously poor in most people and most obviously improves.

First, find out which dimension you misjudge

A colour has three independent dimensions in the model most tools use: hue (which colour), saturation (how much of it) and brightness (how light). People almost never misjudge all three equally. A common pattern, seen by anyone who has watched people play a colour-memory game, is a hue that is nearly right and a saturation that is far too high — the remembered colour is more vivid than the real one was. Others hold saturation well and drift in brightness; a few are steady on both and lose the hue.

You cannot fix the error you do not know you make. The first week of any colour practice is diagnostic: rebuild colours from memory with three separate controls, and look at the per-dimension error rather than the total score. Hue, saturation and brightness explained covers the model, and why rebuilding from three sliders is more informative than picking from a set of swatches.

A working method

1. Fix the viewing conditions. A calibrated display, consistent brightness, a dim room without a coloured wall behind the screen. Colour judgement is relative: a target seen next to orange looks cooler than the same target next to blue, and a screen at full brightness at noon and half brightness at night is two different screens. Every serious colour test warns about this; the warning applies to practice too.
2. Judge before you measure. When you need a value, say it first — "about 210°, a third saturated, fairly light" — and then open the picker. The gap between the guess and the reading is the feedback, and it is free. A picker used without a prior guess teaches nothing.
3. Name the three dimensions separately. Hue, then saturation, then brightness, as three answers. Over a few weeks the one you keep getting wrong becomes obvious, and once it is obvious you start correcting for it before you answer.
4. Reproduce from memory, daily, with a score. Look at a colour, let it disappear, rebuild it, and get a number with a breakdown. Five rounds is enough; the targets must be new each time, or you learn the targets rather than the skill. Do it at the same time of day, in the same light.
5. Compare in context, then in isolation. Colours in a layout interact; a grey that reads neutral alone reads warm beside a blue. Practise both: the isolated judgement and the one in company. The second is the one the work needs.
6. Rest the eye. Staring at a saturated colour for more than a few seconds adapts the cones and shifts the next judgement — the afterimage effect. Look at a neutral surface between judgements. This is also why a memorisation phase of a few seconds, not a minute, is the right length for practice.

For designers, photographers and colourists

The method is the same in every trade; the dimension that matters differs. Interface and brand designers mostly need hue and saturation memory — holding a palette in mind across a day and spotting when a shade has drifted. Photographers and retouchers need brightness and saturation judgement above all, because a quarter-stop and a touch of saturation are exactly what separates a print that looks right from one that looks processed. Colourists grading video need all three, in context, across many shots, which is the hardest version of the task and the reason colour-grading suites use reference monitors and controlled rooms.

If you work in one of these, practise the dimension your work rests on, and measure it against your own earlier results rather than against a population. The number that matters is not whether you are better than average; it is whether you are better than you were in March.

If you have a colour-vision deficiency

Training does not change it, and nothing here should be read as suggesting it can. What practice can do is sharpen the judgements that do not depend on the affected cones — brightness and, often, saturation — and make you more precise about where your discrimination fails, which is itself useful knowledge in colour work. A good tool should say when an exercise cannot be meaningful for you, rather than scoring it and staying quiet about why; a bad score on a colour exercise is not a diagnosis of anything, and a diagnosis comes only from an eye-care professional.

Where Percore fits

Percore's Colour session is the daily reproduction step above, built in. A colour fills the screen for a few seconds with nothing else on it; it disappears; you rebuild it with hue, saturation and brightness sliders; and you see the two side by side with the error on each dimension and an accuracy from 0 to 100. The targets are new every day and the same for everyone, so your series is comparable with yesterday's. It is free, and takes about half a minute. What it does not do is test discrimination with the target in view — for that, use the hue-arrangement test on the list of colour perception tests, which also covers the other reproduction games.