Simultaneous Contrast Illusion

The same grey, twice

In the demo, two squares of exactly the same grey sit on two backgrounds, one dark and one light. The square on the dark background looks noticeably lighter than the one on the light background. A colour picker confirms they're identical, and if you cover the backgrounds so only the squares show, the difference vanishes. Uncover them and it's back.

Why context wins

The visual system almost never judges brightness or colour on its own. Every patch is read relative to its surroundings, because relative judgements are far more useful for recognising real objects. The amount of light a surface reflects into your eye depends on the lighting, from midday sun to a dim lamp, so a system that reported raw brightness would see the same white shirt as a different shade every hour. Comparing a patch with its neighbours cancels much of that out. A patch surrounded by darker things reads as relatively light, and the same patch surrounded by lighter things reads as relatively dark.

This comparison is automatic, which is why knowing the squares are identical doesn't stop them looking different.

Wiring, inference, or both

Part of the effect is thought to come from lateral inhibition, where neurons responding to one spot partly suppress their neighbours. A grey square on a dark background gets little suppression from its dimly lit surroundings and so produces a stronger signal; on a bright background it's suppressed more and reads darker. That sharpens edges, which is useful, and the side effect is exaggerated contrast.

Lateral inhibition doesn't explain everything, though. The same grey can look very different depending on whether the brain interprets the surroundings as a different surface or as different lighting, which local wiring can't tell apart. Edward Adelson's review describes demonstrations where the local contrast stays the same but the perceived lightness changes with how the scene is grouped (Adelson, 2000). The checker shadow illusion is the best-known example of that higher-level inference.

Colour does it too

The same thing happens with colour. A grey patch on a strongly coloured background takes on a faint tint of the opposite colour: grey on red looks slightly green, grey on blue looks slightly yellow. The bigger the difference between the two backgrounds, the stronger the effect, whether the difference is in brightness or in colour.

Artists have used this deliberately for centuries, choosing neighbouring colours to make a colour look brighter, duller, warmer or cooler than its pigment would on its own. Josef Albers built a whole course around it; his book Interaction of Color opens with exercises in making one colour look like two (Albers, 1963).

A dye-works mystery

The French chemist Michel Eugène Chevreul set out the principle in 1839. As director of dyeing at the Gobelins tapestry works in Paris, he was investigating complaints that some dyes looked wrong, particularly blacks that seemed to lose their depth. He found that the dyes were fine; the problem was the colours woven next to them. His law of simultaneous contrast became essential reading for painters and textile designers, and later for vision scientists.