How to use the simulator
Set a color with the input or the sliders. Each column shows that color — and a fixed six-color test palette — as it appears under typical vision and four color-vision deficiencies. Scan across the columns: any colors that look distinct under “Normal” but merge under another column are a risk if you use them to convey meaning.
The color science
The sRGB color is decoded to linear light, a 3×3 Machado (2009) matrix is applied, and the result is re-encoded to sRGB. We work in linear light because it is the more physically-rigorous convention — and because the achromatopsia grayscale uses Rec. 709 relative luminance, which is defined on linear RGB. The original paper applied the matrices in gamma-encoded sRGB and considered the difference small, so other tools may vary slightly.
Frequently asked questions
What are the types of color blindness?
The common ones are red-green deficiencies: protanopia (missing red cones) and deuteranopia (missing green cones), which together affect roughly 8% of men and 0.5% of women. Tritanopia (missing blue cones) is rare, and achromatopsia (little or no color perception) is very rare. This tool simulates all four at full severity; the achromatopsia panel is a luminance-preserving grayscale and doesn't model the reduced acuity or light sensitivity that often accompany it.
How accurate is the simulation?
It uses the Machado, Oliveira & Fernandes (2009) model — a physiologically-based method, and the same model Chrome DevTools uses for its built-in color-vision emulation. We apply the transform in linear light (the sRGB color is decoded to linear, transformed, then re-encoded), which the more rigorous references treat as more accurate across the full brightness range. The original paper applied the matrices in gamma-encoded sRGB and found the difference small, so you may see minor variation between tools — it's a simulation, not a medical diagnosis.
Why do reds and greens look the same in some columns?
Protanopia and deuteranopia both impair the red-green axis, so colors that differ mainly along it (a red and a green of similar lightness) collapse toward the same muddy yellow-brown. That's exactly why color alone is an unreliable way to convey meaning — pair it with text, icons, or patterns.
Is my color sent anywhere?
No. The simulation runs entirely in your browser with a small matrix multiply. The color you're testing is encoded in the URL (so you can share a result), but nothing is uploaded.
How do I make my palette color-blind safe?
Don't rely on hue alone to distinguish important states — vary lightness too, and add a non-color cue. Use the contrast checker to keep text legible, and check that semantically-different colors stay distinguishable across the columns here (especially protanopia and deuteranopia).
Related tools
References
This tool’s color math is grounded in the following standards and primary sources.
- A physiologically-based model for simulation of color vision deficiency — Machado, Oliveira & Fernandes (2009), IEEE TVCG
- Understanding SC 1.4.1: Use of Color — W3C WAI — why color alone is unreliable
- culori — color library — Dan Burzo (MIT)
Spotted an error? Let us know — reader corrections are the best review this site gets.