What is the difference between protanopia, deuteranopia and tritanopia?

Protanopia is a missing or shifted long-wavelength (red) cone, deuteranopia a missing or shifted medium-wavelength (green) cone. Both cause red-green confusion; protan vision additionally dims reds noticeably. Tritanopia involves the short-wavelength (blue) cone and causes blue-yellow confusion. It is far rarer and is more often acquired than inherited.

Can this simulator diagnose colorblindness?

No. It is a perception aid and a design tool, not a medical device. Diagnosis needs a clinical test such as Ishihara plates, the Farnsworth D-15 or an anomaloscope, administered by an optometrist under controlled lighting.

Is it free, and does it upload my camera feed?

It is free with no account, and nothing is uploaded. The simulation runs frame by frame in your browser; the video never leaves your device.

MyColorTools → Colorblind simulator

Colorblind simulator — and a preview of what colorblind glasses would do

Q: Does the filter simulate colorblindness, or correct for it?

Both, as two separate directions in the color-vision assist panel. They are opposite transforms of the same deficiency type. Simulate it reproduces the deficiency for a viewer who does not have it, collapsing confusable colors together — this is the one to use for an accessibility check. Correct for me daltonizes for a viewer who does have it, pulling those same colors apart, which is the idea colorblind glasses are sold on. The panel labels the active direction, and the default is Correct for me.

Q: Do colorblind glasses work?

They do not cure or correct color vision deficiency. Notch-filter lenses cut a band of wavelengths where the red and green cone responses overlap, which exaggerates the difference between colors that were already being confused. Some people with red-green deficiency find this genuinely useful; others notice little beyond an overall tint, and they do nothing for the total colorblindness of achromatopsia. Because the effect varies so much between individuals, trying an approximation of the effect first is worth the five minutes.

Point your camera at anything and watch it re-rendered in protan, deutan or tritan color vision. Then turn on the correction filter and see whether pulling those colors apart actually helps you — before spending two or three hundred dollars finding out.

Open the simulator →

Nothing is uploaded. The camera feed is transformed frame by frame in your browser. There is no server, no account and no analytics, on this page or in the tool.

What the simulator actually does

Human color vision starts with three cone types, sensitive to long (L), medium (M) and short (S) wavelengths. Color vision deficiency is what happens when one of those is missing, non-functional, or shifted so far that it overlaps a neighbour. Modelling that means taking each pixel into LMS cone-response space, altering the affected channel the way the shifted or missing cone would, and coming back out to sRGB. This tool uses the Machado, Oliveira & Fernandes (2009) matrices, which bake that whole round-trip into a single 3×3 color matrix per type — so it runs on every frame of live video on a phone, and the result looks like a photograph rather than a color filter dropped over one.

Three types are modelled:

Type	Cone affected	What gets confused	Roughly how common
Protan	L (long / “red”)	Red–green; reds also look noticeably darker	~1% of men
Deutan	M (medium / “green”)	Red–green, without the darkening	~6% of men — the most common by far
Tritan	S (short / “blue”)	Blue–yellow, blue–green	Rare, and usually acquired

Around 1 in 12 men and 1 in 200 women have some form of color vision deficiency. Most are not dichromats — they are anomalous trichromats, with a cone that works but is shifted. That is why the strength slider matters: dragging it to 100% shows the textbook dichromatic case, while 40–70% is closer to what most colorblind people see, and it is the setting worth using if you are checking a design.

Normal trichromatic vision — eight distinct hues:

The same eight, approximated in deuteranopia — the first four have largely collapsed into each other:

Two static strips only go so far, which is the point of doing this on a live camera. Traffic lights, ripe and unripe fruit, a heat map in a spreadsheet, wire colors behind a wall plate, a red pen on green paper — the confusions that matter are the ones in front of you, not the ones in a diagram.

Previewing colorblind glasses without buying them

Colorblind glasses — the notch-filter kind — do not restore a missing cone. They remove a band of wavelengths where the L and M cone responses overlap most, which widens the gap between the signals that were being confused. Colors that used to sit on top of each other separate. Everything also picks up a tint, and low light gets worse, because the lens is throwing away light to do its job.

The hue shift slider in the assist panel approximates the useful half of that: it rotates hues so confusable pairs land further apart, without discarding any light. Combined with a deficiency simulation it answers the question people are really asking before they buy — would separating these colors help me, in the situations I actually care about?

Try it honestly:

Pick your own type, and set the direction to Correct for me — you already have the deficiency, so there is nothing to simulate. (Simulate it is for the other question: showing someone with normal color vision what you see. Leave it off for this test.)
Point the camera at something that genuinely trips you up day to day.
Rotate the hue shift and watch whether the thing you could not distinguish separates — and whether anything you could read stops being readable.
Repeat in different light. Lenses behave very differently indoors and outdoors.
Save a frame or two as you go — the download button in the tools row writes the picture out with the filter baked in, and names the file after the deficiency, the direction and the strength it was taken at, so you can put two settings side by side afterwards.

If the separation helps here, physical lenses are worth trying — most reputable sellers offer a return window, and you should use it. If nothing improves, you have saved yourself a few hundred dollars and an afternoon.

This is not a medical device. The simulator and the assist filters are perception aids and design tools. They cannot diagnose color vision deficiency and they are not a substitute for prescribed lenses or an optometrist. If you have never been tested, get tested — a clinical test takes minutes.

For designers and developers

The simulator works on uploaded images too, which makes it a quick accessibility check: drop in a screenshot of your dashboard, chart or game UI, set deutan at 100% with the direction on Simulate it, and look for anything that has stopped being distinguishable. Check the direction before you trust the result — on Correct for me the filter separates those colors instead of merging them, so a broken palette can look fine.

Common failures it catches in seconds:

Red/green status indicators with no shape, icon or label to back them up.
Chart series separated only by hue, with no pattern, direct label or varying lightness.
Required-field errors marked by color alone.
Link text that stops being distinguishable from body text once hue is removed.
Heat maps built on a red-to-green ramp instead of a lightness ramp.

The fix for nearly all of them is the same: never let hue be the only carrier of meaning, and keep enough lightness contrast that the design survives in grayscale. The color identifier reports a WCAG contrast ratio for any sampled color, which is the other half of that check.

Frequently asked

Do colorblind glasses work?

They do not cure or correct color vision deficiency. They exaggerate the difference between colors that were already being confused. Some people with red-green deficiency find this genuinely useful; others see little beyond a tint. They do nothing for achromatopsia. Because the outcome varies so much between individuals, previewing the effect first is worth the five minutes it takes.

Can this diagnose colorblindness?

No. Diagnosis needs a clinical test — Ishihara plates, the Farnsworth D-15 or an anomaloscope — under controlled lighting, administered by an optometrist. A phone screen's color accuracy and your ambient light are both uncontrolled here.

Does the filter simulate colorblindness, or correct for it?

Both — they are separate directions in the assist panel, and they are opposite transforms of the same deficiency type. Simulate it reproduces the deficiency for someone who does not have it, collapsing confusable colors together. Correct for me daltonizes for someone who does have it, pulling those same colors apart. Picking the wrong one gives you the reverse of what you wanted, so the panel labels the direction under the toggle. The default is Correct for me.

Which type should I choose?

If you are checking a design, use Simulate it on deutan first — it is the most common by a wide margin — then protan. If you are exploring your own vision, use whatever your optometrist told you and set the direction to Correct for me; if you have never been tested, try each type on Simulate it and see which one looks like an ordinary day.

Does it work on a phone?

Yes, on iOS and Android. It is a website, so there is nothing to install, though you can add it to your home screen. Live camera use needs an HTTPS connection, which this site uses.

The rest of the toolkit

Open MyColorTools →