Vision · screening indicator, not a diagnosis
Color Blind Test
This color blind test shows fourteen plates of colored dots, each containing a number. Read out what you see. The plates are generated live from the cone confusion lines that define each type of color vision deficiency, so the digits change every visit and cannot be memorized. It screens the red-green and blue-yellow axes separately — and it is a screening indicator, never a diagnosis.
What number do you see?
Answer with the digits you can read. If you cannot make out a number, say so — that is a real answer here and some plates are meant to be hard.
Turn off Night Shift, f.lux, dark-room modes and any blue-light filter before you start — every one of them moves the colors these plates depend on. Screen brightness at least half way up, and view the plate straight on.
What does this color blind test actually measure?
Normal human color vision uses three cone types, sensitive to long, medium and short wavelengths — L, M and S. Color is the ratio between their three responses. If one cone type is missing or shifted, whole families of colors that a normal observer separates collapse into one. Those families are called confusion lines, and they are the entire basis of this test.
Each plate here is built from two colors that differ only along one such line. A figure drawn in one color on a field of the other is a pure signal on a single cone axis: obvious to anyone with that cone working normally, invisible to anyone without it. Every dot also gets an independent brightness jitter drawn from the same range whichever role it has, so there is no lightness edge to find the digit by instead.
The types of color blindness
| Type | Affected cone | Prevalence (men) | What it looks like |
|---|---|---|---|
| Deuteranomaly | Medium-wavelength (green), shifted | ≈ 5% | The most common form by a wide margin. Greens and reds are distinguishable but muddier; the difference between them shrinks rather than vanishing. |
| Deuteranopia | Medium-wavelength (green), absent | ≈ 1% | Red, orange, yellow and green become variations on one hue separated only by lightness. |
| Protanomaly | Long-wavelength (red), shifted | ≈ 1% | Similar confusions to deuteranomaly, plus reds appear noticeably darker than they should. |
| Protanopia | Long-wavelength (red), absent | ≈ 1% | Deep reds can appear almost black. This is the type most likely to cause trouble with brake lights and warning indicators. |
| Tritanomaly / tritanopia | Short-wavelength (blue) | ≈ 0.01% | Blue and green confuse, as do yellow and pink. Rare, inherited on an autosome rather than the X chromosome, so it affects both sexes equally. |
| Achromatopsia | All cones | ≈ 0.003% | True total color blindness. Very rare, and usually accompanied by poor acuity and severe light sensitivity. |
What to do with the result
- If nothing was flagged — that is the expected pattern for normal trichromatic vision. It does not certify anything, but there is nothing here to act on.
- If a red-green axis was flagged — this is by far the most common outcome, affecting roughly one man in twelve. It is not degenerative and it does not worsen. An optometrist can confirm the type and degree in a few minutes.
- If the blue-yellow axis was flagged alone — inherited tritan deficiency is genuinely rare, so the more likely explanations are a display problem, a blue-light filter you forgot to switch off, or an acquired change. Retake it on a different screen before drawing any conclusion, and see a clinician if it persists.
- If anything has changed recently — acquired color vision changes, particularly in one eye, can be a symptom of an underlying condition. That warrants an appointment regardless of what this page said.
For the clinical vocabulary — anomalous trichromacy versus dichromacy, what an anomaloscope does, and the exact geometry these plates are generated from — see the color vision deficiency test. To see what a colorblind viewer actually sees rather than screen your own vision, the color blindness simulator at hextorgb.io runs the same cone-space transforms in the opposite direction.
Color blindness questions
How accurate is an online color blind test?
Good enough to flag a deficiency worth investigating, and not good enough to diagnose one. The obstacle is the display: no web page knows your screen's color profile, its brightness or whether a blue-light filter is running, and all three shift the exact colors the plates depend on. Treat a flagged result as a reason to see an optometrist, not as a finding.
How common is color blindness?
Roughly 1 in 12 men and 1 in 200 women of Northern European descent have some form of red-green color vision deficiency. The genes for the long and medium wavelength cone photopigments sit on the X chromosome, so men — who have only one copy — are affected far more often. Blue-yellow deficiency is far rarer, around 1 in 10,000, and affects both sexes about equally.
Why are these plates different every time?
Because they are generated, not scanned. Each plate is built from two colors that differ along a single cone confusion axis, with fresh digits and dot positions on every visit. That keeps the test honest — you cannot memorize the answers — and it avoids reproducing the copyrighted Ishihara plates, whose web scans have usually had their color accuracy destroyed by JPEG compression anyway.
Can color blindness be cured?
Inherited color vision deficiency cannot currently be cured. Tinted lenses marketed for color blindness do not restore missing cone function; they exaggerate the difference between colors that were already confusable, which some people find useful and others find disorienting. Acquired color vision changes are different, and can point to an underlying condition — which is exactly why a sudden change should be seen by a clinician.
What jobs require normal color vision?
Commercial and military piloting, some maritime and rail roles, parts of the electrical trades and certain armed forces positions all set color vision standards, usually assessed with calibrated plates and sometimes a lantern test. If your career depends on it, book a proper examination — no browser test will be accepted as evidence either way.
Does this test work on a phone?
Yes, but turn off Night Shift, True Tone, dark-room modes and any blue-light filter first, and raise the brightness. Those features exist specifically to shift the color balance of the display, which is precisely what this test cannot tolerate. Take it near a window rather than in a dark room, and view the screen straight on.
Other instruments on this bench
- color vision deficiency testThe same plates, graded by confusion-line strength, so protan, deutan and tritan are separated rather than lumped.protan / deutan / tritan
- reaction time testFive trials on performance.now() with false-start detection, then median, mean and best against published norms.median ms · percentile
- aim trainerThirty targets, measured time-to-target and click accuracy — the pointing half of reaction speed.ms/target · accuracy %
- hearing range testWeb Audio pure-tone sweep from 20 Hz to 20 kHz, per ear, with calibration first. Screening indicator, not a diagnosis.highest audible Hz
Testing the hardware rather than the person? device tests — scroll, touch, refresh rate and input latency lives on a sister site. And if you want to see through someone else’s eyes rather than screen your own, use the simulate how an image looks to a protan, deutan or tritan viewer.