How does the color difference test work?
You see a grid of tiles that are all one color except for one, which is a slightly different shade. Tap the odd one out. Each time you find it, you move up a level: the difference gets smaller and, every few levels, the grid gets bigger. The test ends after 3 wrong taps, or at level 30. There is no clock.
How is your result calculated?
Your result is the number of levels you completed. With it you see the smallest color difference you spotted, as a number: the smaller it is, the subtler the difference you could see. After a wrong tap the different tile is outlined for a moment, and you get a new grid at the same level.
How the color difference is measured
Each round picks a random mid-tone color and a second color a set distance away from it in a random direction: lighter, darker, more or less vivid, or slightly shifted in hue. Both are measured in OKLab, a color space published by Björn Ottosson in 2020 and now part of the CSS color specification, built so that equal distances look about equally different to the eye. The number you see is the distance between the two tiles' colors in OKLab, multiplied by 100. Color scientists call a number like this a ΔE, or delta E.
The distance is measured on the colors actually sent to your screen, after rounding to the 256 levels per channel that screens use, so it is never smaller than the difference that was really shown. The targets for some levels:
| Level | Target difference (ΔE) | Grid |
|---|---|---|
| 1 | 10.0 | 2 × 2 |
| 5 | 5.0 | 3 × 3 |
| 10 | 2.1 | 5 × 5 |
| 15 | 0.9 | 6 × 6 |
| 20 | 0.8 | 6 × 6 |
From the level where the target reaches 0.8, it stays there: a smaller step could round away to nothing on a standard screen.
What can change your result
- Your screen. Brightness, contrast and how a screen is calibrated all change how different two colors look. Results from different screens aren't directly comparable.
- Night mode and filters. Blue-light filters, night modes and automatic color adjustments change every color on screen. Turn them off for a fair run.
- Viewing angle and light. Many screens shift color when seen from an angle, and glare washes out small differences. Look at the screen straight on, out of bright light.
- Color vision. Some differences are changes of hue, which people with color blindness may find much harder to see.
What this test does not measure
It isn't a color vision test. Clinical color vision tests use carefully controlled colors and lighting, and a phone or computer screen can't reproduce them. Your result describes what you could see on this screen, in this light. For another test of noticing small differences, try the Change Blindness Test.