The Left Half Looks Lighter. The Right Half Looks Darker. They are the Same.
Which patch is lighter?
You are looking at the Cornsweet illusion, described by Tom Cornsweet in his 1970 book Visual Perception (sometimes called the Craik-O’Brien-Cornsweet illusion after its three independent discoverers across the 1940s-1960s). Two regions of grey meet at a central boundary where the grey briefly ramps up on one side and ramps down on the other. Far from the boundary, the two regions are pixel-identical. At the boundary, they differ. Your brain propagates the boundary difference across the entire region · and reads one side as uniformly brighter than the other.
What you are about to learn. What the Cornsweet illusion actually is, why it is the single cleanest demonstration of edge-based brightness perception, what it says about how your visual system reconstructs “filled-in” regions, why it has become the flagship example in neural coding of brightness, and how to prove the illusion exists with a finger covering the centre.
What the Illusion Looks Like
Draw a wide horizontal rectangle of a mid-grey. At the centre, on either side of the vertical midline, introduce a soft luminance gradient: the left half ramps up in luminance toward the midline; the right half ramps down away from the midline. Far from the midline, the two halves are the same grey.
Perception: the left half reads as a lighter uniform grey; the right half reads as a darker uniform grey. The lightness difference is dramatic · you would swear they are two different paints. Measure them with a pixel sampler and the far-from-edge pixels are identical.
The minimal recipe. A uniform-luminance surface with a local edge transition · one side ramping up just before the edge, the other ramping down just after. The edge transition itself is narrow, maybe a few pixels wide. The uniformity far from the edge does not matter · your brain uses the edge to “fill in” a brightness difference across the entire span.
Why It Works: Edges are Currency
Your visual system does not encode absolute brightness. It encodes differences · edges, transitions, gradients. The brain then reconstructs the interior of regions by filling in from the edges.
Your retina responds strongly to luminance edges. The centre-surround ganglion cells fire when light falls on the centre but not the surround, or vice versa · they are essentially edge detectors. Uniform regions of light produce almost no ganglion response.
The brain integrates edges to reconstruct regions. Starting from the edges and propagating inward, your visual system fills in the “surface” representation of each region. If the edge says “brighter on the left, darker on the right”, that is how the fill-in proceeds.
The Cornsweet edge is misleading. The local edge ramp falsely implies a much larger brightness difference between the two sides than actually exists. The fill-in mechanism runs with the edge’s implied difference and paints the whole region accordingly. The result: a large perceived brightness difference from a vanishingly small physical one.
You are seeing your own edge-to-surface reconstruction. The Cornsweet illusion is not a trick of the eye · it is your visual system showing its work. The mechanism it uses to build surfaces from edges is a legitimate computation that makes your perception more robust to shading, shadows, and illumination gradients. Cornsweet just found a stimulus where that mechanism produces a reliably wrong answer.
Proving the Illusion
The standard proof is the finger-cover test.
Cover the centre. Place a finger vertically over the centre of the figure, covering the soft edge at the midline. Now the left and right regions appear as a single uniform grey · they are the same. Lift your finger and the illusion returns instantly. The edge was doing all the work. Without it, the two regions have no inferred brightness difference, because there is no physical difference to infer from.
Craik, O’Brien, and Cornsweet
The illusion is sometimes called the “Craik-O’Brien-Cornsweet” illusion to honour its three independent discoverers:
- Kenneth Craik described a related effect in his 1940s Cambridge work on luminance perception
- Vivian O’Brien, in 1958, published a clean figure
- Tom Cornsweet popularised it widely in his 1970 textbook, and the name stuck
Cornsweet gets the modern branding because his book put the effect into the common vocabulary of psychology undergraduates. But the mechanism was well-understood by the 1950s, predating any of the three.
Why it is called the Cornsweet. A good illusion-naming convention favours the person who most effectively communicates the demonstration to the broader community. Cornsweet’s 1970 book was widely adopted in psychophysics and vision courses, so the name spread. Craik and O’Brien published their versions in more specialised outlets. This is a common pattern in illusion history: discovery and naming are not always the same act.
The Mach-Band Connection
The Cornsweet illusion is closely related to Mach bands · the phantom bright and dark stripes you see at the edges of luminance ramps (described by Ernst Mach in 1865). Both arise from lateral inhibition, both exaggerate edges, both are manifestations of the visual system’s tendency to sharpen transitions.
The difference: Mach bands add extra bright and dark strips right at the edge of a gradient. Cornsweet takes the same edge-amplification machinery and uses it to reassign the brightness of whole regions far from the edge. Same family, different consequence.
Common misconception: “the regions are not really the same.” They are. Take a screenshot, open it in any image editor, and sample pixel colours in the far-from-centre regions of both halves. The RGB values are identical. If that is not convincing, print the figure and hold a piece of tracing paper next to it to compare · same grey, confirmed independently by every measurement tool you can bring to bear. The difference exists only in your perception.
A Harder Variant
Below is a Cornsweet figure at difficulty 3 · with a sharper edge gradient and larger regions. The two uniform areas are still identical.
Which patch is lighter?
The Cornsweet in reverse. Flip the direction of the edge ramp · left side ramps down toward the midline, right side ramps up away from it · and the illusion reverses. The left now looks darker, the right lighter. The edge has a direction, and so does the fill-in. Watching this reversal in real time makes the edge-to-surface machinery vivid: you are literally seeing your visual system paint the surface in response to an edge cue.
Cornsweet in the Modern World
- Display engineering. HDR and edge-enhancement algorithms in televisions and monitors implicitly use Cornsweet-style tricks · exaggerating edges subtly to make images look crisper while preserving the average luminance of regions. Some image processing directly implements the Cornsweet as a sharpening filter.
- Printing. Laser and inkjet printing benefits from Cornsweet-type unsharp masking in image pipelines. The printed result looks sharper and more vibrant than the raw photograph, but the mean luminance is preserved · a trick your visual system invented first.
- Medical imaging. Radiologists’ image-viewing software often enhances edges slightly, which (beyond making pathology easier to spot) recruits the Cornsweet mechanism to improve perceived contrast across regions that would otherwise read as uniform.
- Art. Representational painters routinely sharpen edges for figures against backgrounds, knowing that the viewer’s visual system will take the local edge cue and extend it into a general figure-ground brightness separation. Vermeer and the Dutch masters were unknowingly virtuoso Cornsweet exploiters.
Test Yourself on 50 More Illusions
The Cornsweet illusion is one of more than 50 classical illusions on PlayMemorize. Each round draws a deterministic SVG scene and asks one grounded question: which is larger, which is brighter, which is actually parallel. The reveal overlay shows the true geometry plus a one-line “why it works” caption.
- Keep playing Cornsweet → · the standalone game, pinned to this one figure with fresh seeds each round
- Play Illusions → · spot the tricks across size, colour, orientation, and impossible figures
- Play Spatial → · train mental rotation and area estimation
- Play Matrix → · abstract pattern reasoning under time pressure
The takeaway. The Cornsweet illusion is a live demonstration of the fact that your visual system does not see surfaces · it computes them from edges. Every uniform region of brightness you perceive is the output of an edge-to-surface reconstruction running on raw retinal edge signals. Cornsweet found the stimulus where that reconstruction goes wrong in a particularly dramatic way. Understanding it is understanding one of the deep computational strategies your brain uses to turn a 2D retinal image into a perceived 3D world of surfaces.
Illusions
Your eyes lie - the math knows the truth. Spot equal lengths, identical greys, and truly parallel lines across 57 classic optical illusions
שחקו עכשיו - בחינםללא חשבון. עובד בכל מכשיר.