What’s Your Colour-Eye Level?

weirdly accurate.

    What’s Your Colour-Eye Level?

    How good is my colour perception?

    You can measure it. Colour scientists put the gap between two colours in ΔE00 (CIEDE2000), where a difference of about 1 is close to the edge of what people can see. Eight trained observers picking the odd colour out on a calibrated monitor had thresholds around ΔE00 2.5 (Hong et al., 2026), and in dental shade matching half of observers notice a gap of 0.8. Where you land depends on your eyes, your age, your screen and the light.

    This test works like an eye chart for colour. In a grey colour-matching booth a paint fan deck opens with seven chips; one is a hair off, and you tap it. There are 15 rows: the difference starts at ΔE00 16 and halves every three rows, reaching 1 at row 13 and 0.63 at row 15. Every fan is a new colour from twelve paint families, and the odd chip is lighter or darker, more vivid or greyer, or turned a touch round the hue wheel, the three taking turns, so your level covers colour in general. Each right tap takes you down a row. The first miss gives you a second try at the same row in a new colour; the second miss ends the test. Your finest row becomes a threshold: the difference you would see half the time once lucky guesses (1 in 7) are taken out, the same criterion Hong et al. used.

    It is not a colour-blindness test and it can’t diagnose anything; if you are worried about your colour vision, see an optometrist. It does not rank you either: nobody has measured a large crowd with this task on everyday screens, so instead of inventing a percentile it places your number among differences professionals actually work to. Brightness, Night Shift or eye-comfort modes, True Tone, colour profiles, viewing angle and a bright room all change what reaches your eye, and one run is noisy.

    What the test measures

    One skill, measured like an eye chart: the smallest colour difference you can still pick out.

    The task

    • Seven paint chips on a fan deck in a neutral grey light booth: tap the one that is different
    • 15 rows measured in CIEDE2000 (ΔE00), from 16 down to 0.63, the difference halving every three rows
    • Twelve paint families round the colour wheel; the odd chip changes in lightness, colourfulness or hue, the three taking turns
    • Two misses in all: the first is a second try at the same row in a new colour, the second ends the test

    Your result

    • Your colour-eye level (1–15) and your threshold in ΔE00
    • Your number among four reference points: the dental difference half of observers notice (0.8), the one half still accept as a match (1.8), trained observers on a calibrated lab monitor (about 2.5) and the print tolerance for cyan, magenta and yellow (3.5)
    • Everyday terms: whether you would spot a poster printed within tolerance or the seam between two neighbouring teeth, and two chips you can slide together to feel your own threshold
    • Row-by-row details: which kind of change you saw best, which colours your misses fell on, and how long the hard rows took
    • A shareable card and a challenge link; the link carries your level in the part after the #, and no account is needed

    Before you start

    • Screen brightness up; Night Shift, eye-comfort modes and True Tone off
    • Look at the screen straight on, away from a bright window or lamp
    • Works on phones and computers in a browser with 3D (WebGL) switched on
    • About two minutes

    Colour perception FAQ

    Is this test free? Do I need to sign up to see my result?

    Yes, it’s free, and there’s nothing to sign up for: no account, no email, no payment. Your result appears on this page the moment you finish, and you can share it or play again straight away.

    How can I test my colour perception online?

    With a task that hides one slightly different colour among identical ones and shrinks the difference every row. This test uses seven paint chips and 15 rows measured in ΔE00, the colour-difference unit printers and dentists use, and reports the finest difference you picked out. Because it runs on your own screen, turn the brightness up and switch off Night Shift and True Tone first.

    Is this a colour blindness test?

    No. Colour blindness tests, such as the Ishihara plates an optometrist uses, look for specific kinds of colour-vision deficiency. This test measures how small a colour difference you can pick out on your screen, and it can’t diagnose anything. If you think you may be colour blind, see an optometrist or eye doctor.

    What does ΔE mean in colour?

    ΔE (delta E) is a number for how different two colours look. This test uses CIEDE2000, written ΔE00, the version used in printing, paint and dentistry, where about 1 is close to the edge of what people can see. In dental research half of observers notice a difference of 0.8 between ceramic samples, and offset printing allows 3.5 for cyan, magenta and yellow solids.

    What is a good score on a colour perception test?

    There is no published average for this task on everyday screens, so this test gives no percentile. It gives reference points instead: trained observers on a calibrated monitor had thresholds around ΔE00 2.5 (Hong et al., 2026), and a threshold finer than 1.8 means you would see a tooth-shade difference half of observers still accept as a match (Paravina et al., 2015).

    Does colour discrimination get worse with age?

    On average, yes. In Reinecke, Flatla and Brooks’s 2016 online study of 29,044 people on their own screens, telling colours apart was best between ages 20 and 30 and worse with age. On the Farnsworth-Munsell 100 hue test the average error score is lowest around 19 and roughly triples by the 70s (Kinnear & Sahraie, 2002). Neither used this task, so their numbers don’t convert to a level here.

    Why did I get a different colour-eye level on another screen?

    Because the screen and the room are part of the measurement. Reinecke and colleagues’ 2016 online study found people did worse in brighter surroundings and on dimmer screens, and settings such as Night Shift, True Tone and colour profiles shift the colours you are shown. One run is also noisy; a few runs on the same screen tell you more than one.

    Where the numbers come from

    There is no ranking: no study has measured a large population with this task on everyday screens, so your threshold is placed among reference points instead. Your level is converted to the guess-corrected half-way criterion Hong et al. (2026) used, but they used a calibrated monitor and this is your screen. The print tolerance (3.5) is the informative value for cyan, magenta and yellow solids in ISO 12647-2:2013, as summarised by Yu, Chung & Myers (2015, TAGA Proceedings); neither has an open page to link. The site does not publish its own players’ data.

    How the level is scored

    ΔE00 (CIEDE2000)

    The colour-difference formula used in printing, paint and dentistry. Around 1 is close to the edge of visibility; the smaller the number, the harder the pair.

    Threshold

    The difference you would pick out half the time after removing lucky guesses. With seven chips a blind tap is right 1 time in 7, so the threshold sits between chance and certainty, as in Hong et al. (2026).

    Two-miss chart

    Every right tap takes you down a row, so a run is short. Your level is the last row you passed; the threshold is that row’s difference, corrected for how a chart that stops at the second miss behaves.