Colour Blindness Simulator
See a design through the four colour vision deficiencies — simulated in linear light.
Original
Deuteranopia
How this processes your file
Each pixel is converted from sRGB to linear light, multiplied by the standard Viénot–Brettel–Mollon matrix for the deficiency, and converted back. Doing it in linear light is the part most implementations skip: sRGB values are gamma-encoded, so multiplying the stored bytes directly is quicker, looks roughly plausible and understates how much contrast is actually lost. A useful property of those matrices is that every row sums to one, which is why neutral greys stay neutral — colour vision deficiency does not affect achromatic perception, and greys drifting toward a colour cast is the visible symptom of a matrix copied down wrongly.
What the results mean
- Protanopia
- The long-wavelength cone absent. Red appears dark and is confused with green and dark grey.
- Deuteranopia
- The medium-wavelength cone absent. Red and green confused without red darkening — why red-and-green indicators fail.
- Tritanopia
- The short-wavelength cone absent. Blue confused with green. Not sex-linked, and often acquired.
- Achromatopsia
- No colour discrimination. The honest test of whether a design carries information in luminance.
Common problems and fixes
- Two chart series became the same colour
- Then colour was carrying the information alone. Add a second cue — a shape, a label, a line style — rather than only changing the palette, since the next palette has the same failure mode.
- My design looks fine in all four
- Good, and check contrast separately. Colour deficiency and low contrast are different problems, and passing one says nothing about the other.
- Is this what colour blind people see?
- No. It approximates what information is lost. Nobody perceives the world as a filtered version of someone else's, and the value here is spotting lost distinctions rather than depicting an experience.
Frequently asked questions
Why simulate in linear light?
Because the transformation is defined on light rather than on encoded values. sRGB is gamma-encoded, so applying the matrix to the stored bytes is a different calculation — it looks approximately right and understates how much separation is lost.
How common is colour vision deficiency?
Around 8% of men and 0.5% of women, but most of that is anomalous trichromacy, which is milder than the dichromacy simulated here. Roughly 1% of men have protanopia and 1% deuteranopia.
What should I change if a design fails?
Add a cue that is not colour. Shapes, labels, patterns and position all survive every deficiency, and picking a different palette only moves the problem to a different pair of colours.