Color analysis when you are color blind

Accessibility · which tests survive · and which do not

Every color-analysis guide on the internet, ours included, quietly assumes you can see the difference between a blue-based red and an orange-based red. Nobody mentions this. So you work through the tests, get nowhere, and reasonably conclude you did it wrong. You didn't. Here's the version with the assumption written down: of the six standard tests, one works perfectly for you, two are usable, and two are a straight waste of an afternoon. Below is which is which, plus what to do about the half you can't get to on your own.

Analyze my photo – free

A dot mosaic shown twice: as normal color vision sees it, with the figure legible, and simulated for deuteranopia, where it nearly disappears

First, the useful distinction

Color vision deficiency comes in types, and the type decides which tests are worth your time. The common ones are red-green: protanopia and protanomaly (the red cone missing or shifted) and deuteranopia and deuteranomaly (the green cone). Blue-yellow, tritanopia and tritanomaly, is much rarer. Complete absence of color vision, achromatopsia, is rarer still. The Color Blind Awareness organisation puts red-green CVD at roughly 1 in 12 men and 1 in 200 women.

If you don't know which type you have, an optometrist will tell you in a few minutes. We're not going to offer you a test for it, and you should side-eye any website that does. The standard screening plates are copyrighted, and more to the point they're only valid in calibrated print – so running one on a phone whose brightness you adjusted with your thumb about nine seconds ago is a confident answer to a question nobody actually asked properly. A site that ships one anyway has told you something useful about its standards, just not about your eyes.

What this page can do is tell you which parts of color analysis depend on the cones you're missing, because that's a question about the method rather than about your eyes.

The six standard tests, honestly rated

Rated for red-green CVD, which is the common case. Each one is judged on the single axis it actually asks your eye to discriminate.

Value contrast, the greyscale test – works fully: This one is pure luminance. It asks how light or dark things are, which has nothing to do with the red-green axis. It arguably works better for you than for everyone else, because you aren't being distracted by hue while you judge it. Start here, and it's the interactive test below.
White against off-white – marginal: Mostly a luminance and warmth judgement, so some of it survives. Doable in strong daylight, but treat the result as a lean rather than a finding.
Gold against silver – usually readable: Gold is bright and yellow, silver is close to neutral, and yellow is generally intact in red-green CVD. This one often works for protan and deutan types. If you have a blue-yellow deficiency instead, skip it.
The vein test – unreliable: Blue against green is precisely the discrimination most deuteran readers can't make. Plenty of people with normal color vision also find this test ambiguous, so if it gives you nothing, that's the test failing rather than you.
Blue-based against orange-based red – useless: This is the red-green axis itself, so it's asking for the one comparison your vision doesn't supply. Better lighting won't help. Squinting won't help. You have our permission to never think about it again.
Pink against golden undertone – useless: Same axis, same answer. This is the test that decides warm versus cool in most guides, which is why the whole framework tends to stall out for color-blind readers at exactly the same place.

Why the red test cannot work for you

Six lipsticks in a row, running deep wine to warm coral – every shade the lipstick guide asks you to choose between. Below it, the same photograph transformed for deuteranopia and protanopia. This is a real transform of real pixels, not an illustration of the idea.

Six lipsticks running from deep wine to warm coral, shown as normal color vision sees them and then simulated for deuteranopia and protanopia, where all six converge to near-identical olive shades — Six clearly different reds at the top. Six near-identical olives underneath. This is why "hold a blue-based red next to an orange-based red" is not a test you can run, and why no amount of better lighting changes that.

And why the metals test usually still works

Same transform, applied to gold against silver. Gold is bright and yellow, silver is close to neutral, and yellow largely survives red-green deficiency – so the two stay distinguishable down all three rows. That is what makes this one worth your time when the red test is not.

A gold bangle and a silver bangle against a wrist, shown as normal color vision sees them and then simulated for deuteranopia and protanopia, where the two metals remain clearly distinguishable — The gold stays warm and the silver stays neutral in all three rows. If you have a blue-yellow deficiency rather than a red-green one, this is the test to skip instead.

So what can you actually establish?

Your value contrast, on your own, reliably. That's the relationship between how light your hair, skin and eyes are, and it's a genuinely large part of what makes an outfit work or not. High-contrast coloring wants clear separations; low-contrast coloring wants shades that sit close together. You can settle that today with the test below and it will be correct.

Your temperature, not on your own. Warm versus cool is a hue judgement on the red-green axis, so the standard tests can't deliver it to you. There's no clever workaround with better lighting, and it's worth knowing that rather than assuming you'll get there with more effort.

That leaves two honest routes for the second half. Ask someone whose color vision you trust to run the drape tests and tell you what they see, which works fine and costs nothing. Or use something that measures hue numerically off the pixels rather than asking an eye to judge it – which is the one place on this site where reaching for an instrument isn't a shortcut, it's the only method available to you.

The test that works: your value contrast

Luminance only, so your color vision is irrelevant to it. Answer for what you see rather than what you think you're supposed to see, and take the greyscale photo if you can – it removes hue from the judgement entirely, for everybody.

01
Hair against skin
The big one. In the greyscale photo, how far apart do your hair and your face sit?
02
Brows against skin
Often the real driver, and the one people forget. Judge your natural brows, not a filled-in shape.
03
Eyes against the skin around them
Iris and lash line together, against the surrounding face.
04
A black liner and a deep lip, together
optional
How does that actually look on you? This is the sanity check on everything above.

Your reading

Answer all three gaps to get a reading. 0/3 so far.

Reading your own gaps is genuinely hard, which is why the cross-check above exists. siOsi reports your contrast level directly from one photo, next to your undertone and season – so you get the measured version without judging your own face.

Analyze my photo

First look free · one photo · fifteen labs

One test that works, one that does not

Left is the greyscale comparison, which is a luminance judgement and available to you. Right is the vein test, which asks for a blue-green discrimination and is the one to skip.

A portrait shown in color beside the same portrait in greyscale, used to judge how much the hair, skin and eyes differ in lightness — Works. Convert a photo of yourself to black and white and judge how far apart your hair, skin and eyes sit.

Close-up of an inner wrist in daylight where the veins read as a mix of blue, green and grey — Skip it. Blue against green is the discrimination red-green CVD does not provide.

What a season palette looks like with red-green CVD

Two of our twelve palettes, each shown as normal color vision sees it and then transformed for deuteranopia and protanopia. Computed from the actual palette data, so it shows the practical problem: shades a guide treats as opposites can arrive nearly identical.

Two season palettes shown three times each: as normal color vision sees them, then simulated for deuteranopia and for protanopia, where warm and cool shades converge — Deep Autumn above, Deep Winter below. The rust-versus-berry distinction the vs-page turns on largely disappears in the lower rows.

Practical workarounds that hold up

None of these are about seeing color better. They are about not needing to.

Shop by label, not by eye: Shade names carry the information you need surprisingly often. "Cool rose", "warm terracotta", "blue-based red" and "neutral beige" are all doing the work of a test you cannot run. Brands that only name shades poetically are the ones to avoid.
Build around your value contrast: This is yours, so lean on it. Getting the lightness relationships right accounts for a great deal of whether an outfit reads as intentional, independent of temperature.
Use one trusted second opinion, once: You need the temperature answer once, not continuously. Get someone to run the drape tests with you a single time, write the result down, and shop from it for years.
Prefer numbers over impressions: Any tool that reports an undertone or a hue reading is doing the discrimination arithmetically rather than perceptually. That is exactly the right division of labour here.
Greyscale your own photos: Your phone can do this in two taps, and it turns a color question into a lightness question. Useful for checking whether an outfit has enough separation in it.

“As they rely heavily on coping strategies, colour blind people often think they have correctly identified a colour because it appears to them as the same colour as other things which they know to be a specific colour.”

Color Blind Awareness · PROTECTED1

Where a photo genuinely beats an eye

Most of this site cheerfully tells you to go and check things against a window, because your own eyes in daylight usually beat any tool. This page is the exception, and only for one specific half of the problem.

siOsi judges hue, value and chroma from a photo rather than from a pair of eyes. It returns your undertone as warm, cool, neutral or olive, your season, and your contrast level, and none of that depends on whose cone response is doing the looking. For the temperature question specifically, that isn't a convenience – it's a route to an answer that the at-home tests genuinely cannot give you.

The rest of what it does is about makeup rather than color, and quite a lot of it is also independent of color vision: whether your base oxidises warm as the day goes on, whether it throws flashback in photos, how your blending and transitions read at normal distance, how long things are lasting on you. Fifteen labs in total, and most of them are judging placement, finish and wear rather than shade.

One photo, and your first look is free without an account. And do still run the greyscale contrast test above yourself, because that one is fully yours.

Free color analysis The 12 seasons

Watch it done

Draping these two side by side, from color analysts on YouTube. Some use the “Dark” naming instead of “Deep” – same seasons.

Red-green color blindness: do you have it?

Demystifying Medicine McMaster

How color vision works: types of color blindness, genetics and disease

Doctor Eye Health

Answered

Can siOsi tell me if I am color blind?

No, and it doesn't try to. It analyses a photo of your face, which tells you nothing about your own color vision. Color vision deficiency is diagnosed by an optometrist with calibrated plates. Any website offering to screen you on-screen is overreaching, because uncalibrated displays make those tests invalid.

Which color-analysis test should I start with?

The greyscale value-contrast test, which is on this page. It depends only on lightness, so your color vision has no bearing on the result. It will give you a genuinely reliable answer for one of the two main axes.

Is color analysis worth doing at all if I am color blind?

Yes, for two reasons. Value contrast is a big part of it and it's fully available to you. And the temperature half, once someone or something establishes it for you, is a fact you can shop from for the rest of your life – arguably it is *more* useful to you than to someone who can eyeball it in a shop.

Do you use ColorADD or a similar symbol code?

No. Every color swatch on this site carries its name as text instead, in nine languages, which is more specific than a symbol and needs no legend to read. A symbol code works well on a garment label you see repeatedly; it works poorly on a page you visit once.

Does the type of color blindness change the advice?

A little. The ratings on this page are for red-green deficiency, the common case. If you have a blue-yellow deficiency, also skip the gold-versus-silver test, since yellow is the axis affected. If you have no color vision at all, the value-contrast test is still fully valid and the temperature question needs a second opinion or a measurement.