Ishihara-style colour blindness test
Plates drawn for this page, for red–green and blue–yellow, and a set of shapes for children. A screen, not a diagnosis.
Before you start
Set the screen, then the room.
- Set the brightness to a normal indoor level. Not full, not dim.
- Turn off night mode and any blue-light filter.
- Sit in a well-lit room, and look straight at the screen.
- The result depends on this display. Another screen can disagree.
Nothing is uploaded. The plates are drawn on this device.
What the chat window cannot finish.
Colour vision deficiency is a difference in how the cones respond, most often along red and green, much more rarely along blue and yellow. Screening plates hide a figure in a field of dots so that the figure and the field sit on the same confusion line: easy to separate with typical colour vision, and easy to lose when that channel is weak. This test draws a new plate every time. It does not reproduce a printed clinical set.
An assistant can describe protanopia. It cannot put a plate in front of you, score what you actually see, or notice that a night filter has flattened the screen. The result lives in your eyes and on this display.
What it can do, afterwards, is help you redesign the thing that failed: a status colour, a chart, a map. The simulator and the contrast checker on this site are the next step. The test itself only sorts the plates.
How common it is
MedlinePlus Genetics, US National Library of Medicine. Figures vary by ancestry. These are not a diagnosis and not a claim about any one person.
- Red–green, Northern European ancestry
- About 1 in 12 men and 1 in 200 women
- Blue–yellow
- Fewer than 1 in 10,000 people, any gender
- Worldwide, any colour vision deficiency
- About 2.5% of people
How to read a result
- No pattern
- You read the plates a typical screen expects. That is not a clean bill of health.
- Red–green, protan-like
- The plates that fail for a red-weak pattern were missed, and the green-weak plates were not.
- Red–green, deutan-like
- The reverse. This is the common pattern.
- Red–green, not separated
- Both families failed, so the test will not guess which.
- Blue–yellow
- Uncommon, and a screen is a poor instrument for it.
- No reading
- The control plates failed, or the answers did not settle. Fix the room and try again.
- Mild / strong
- Mild means only the closest plates failed. Strong means widely separated plates failed too.
About this test
The plates are original. A circle is filled with packed dots of varying size, and the figure is a geometric digit or a simple shape. Nothing is traced from a published plate.
- Where it runs
- In the browser. The picture is not uploaded.
- What it cannot do
- It cannot calibrate your screen, and it is not a clinical instrument.
- Control plates
- Two plates differ in lightness, so they should be visible even in greyscale.
- Children
- A separate sitting uses a circle, square, triangle and star. No numbers.
These constants are published under CC BY 4.0. Quote them, cite them, put them in your own documentation. Attribution to Original Design Studio is the only condition.
The principles behind what this tool tells you.
Review
Reviewer: Not yet assigned
Review date: Not yet reviewed
An optometrist still needs to read the wording and the plate behaviour before this is treated as ready to rely on. Until that review, keep the disclaimer and do not describe the result as clinical.
Is this test accurate?
It is a screening exercise, checked against a colour-vision simulation so that each plate loses its figure under the deficiency it targets and keeps it under typical vision. It is not a clinical test, it is not calibrated to your display, and it must not be used to diagnose.
Can I pass the test with glasses?
Ordinary glasses and contact lenses correct focus, not cone response. Wear what you usually wear so the plate is sharp. A tinted lens or a blue-light filter can change the result, so take those off.
Why do I see different numbers on different screens?
The plates are made of colour differences, and screens do not share a calibration. Brightness, night shift, a wide-gamut panel and a cheap laptop can disagree. The setup screen is there because of that.
Is this the official Ishihara test?
No. Ishihara plates are a published clinical set. These plates are generated here: random dot packing, geometric figures, and colour pairs chosen along confusion lines. The method is the familiar one. The artwork is not.
What does a red–green result mean day to day?
Red and green stop being a reliable signal on their own. Status dots, error text and map keys that use only those hues are the usual failures. Lightness, position or a word fixes them for everyone.
Can a child take it?
Yes. Choose “Shapes, for a child”. The figures are a circle, square, triangle and star, the targets are large, and there are no numbers. Sit with them, and still treat the result as a screen rather than an answer.
Does a normal result rule out an eye condition?
No. The sitting only asks whether these plates separated on this screen. It says nothing about acuity, disease, or a deficiency too mild for this display to show.
Why is there no hidden number?
Some printed tests include a figure that typical vision misses and a colour-vision deficiency reveals. A linear simulation cannot honestly produce that, so this test does not pretend to. Every plate here should be visible with typical colour vision. The controls are the check that you are looking.
Do blue-light filters change the result?
Yes. So does night mode. Both shift the colours the plates depend on. Turn them off, set a normal brightness, and take the test in a lit room.
What should I do if the result is not typical?
Book a formal test with an eye-care professional. If you design interfaces, stop using red and green, or blue and yellow, as the only difference, and run the palette through the contrast checker and the simulator.
Next tool
Colour palette from a photo
Want this run across your whole product?
A written plan against the ten disciplines.