Hueseam

A wide color bar hides a faint vertical seam. Slide the marker to the edge you can just make out. 4 rounds.

Learn the theory →

Where does the shade step? Slide the marker to the seam.

About Hueseam — find the hidden seam

Hueseam trains edge detection — finding the faint boundary where one shade gives way to another inside a wide color bar. Because the two halves are so close, the seam almost dissolves, and surrounding color pulls your judgement around through simultaneous contrast. Locating soft edges like this is what lets you catch banding, spot where a mask feathers wrong, or see a subtle join in a printed gradient.

A broad bar hides a barely visible vertical seam between two near-identical shades. Slide the marker to the exact edge you can just make out; four rounds make the seam ever fainter.

Your surroundings bias the boundary through simultaneous contrast. Read that guide, or browse all our Learn about color articles on the science of color perception.

What Hueseam trains

Hueseam works your edge detection at low contrast: finding the boundary where one shade gives way to another when the difference between them is barely there. The two halves of the bar are close enough in lightness that the join almost dissolves, and the whole surface is dithered with fine luminance noise, so you are not looking for a line so much as pulling a weak signal out of a grainy field. Nothing is hidden: the seam is on screen the whole time. The difficulty is deciding where a gradual-looking change becomes a step.

How to play

Each round shows one wide bar in the day's hue. Somewhere between fifteen and eighty-five per cent of the way across, the lightness steps up or down, and either side may be the lighter one. Move the marker to where you think the step falls: drag on the bar, use the slider underneath, or tab to it and use the arrow keys. The marker resets to the middle every round, so the starting position tells you nothing. Press Lock in seam to commit. The bar is then repainted without any noise and the true seam flashes at its real position for about two seconds, before the next round begins. There are four rounds, and all three difficulty knobs move against you: the lightness step shrinks, the noise grows, and the tolerance tightens.

How scoring works

Your error is the distance between your marker and the true seam, measured as a percentage of the bar's width. That error is divided by the round's positional tolerance and turned into an accuracy from 0 to 100. The tolerance tightens gently as you go, from 13 per cent of the bar's width in round one to about 9 per cent in round four, roughly thirty per cent narrower, so the same misplacement scores a little lower late than early. It is the fainter step and the heavier grain that make the later rounds hard, not the tolerance. Your daily result is the average of the four rounds. At or above 85 the round is marked as spot-on on the result strip, the middle band as a near miss, and anything lower as a miss. Score 60 or more and your streak continues. Because the reveal is automatic and unconditional, you always learn the answer, whether you nailed it or missed it entirely.

What your score means

Read the daily number as a band, not a measurement:

OutcomeRough interpretation
85-100 accuracyYou are localising the edge, not guessing
45-84 accuracyRight region, imprecise placement
Below 45The noise won, or you stopped scanning

One day is noisy in its own right: screen brightness, viewing angle and room light all change how visible a small step is. The trend across days is the honest signal.

Common mistakes

The commonest error is staring at one spot. A low-contrast edge is easier to catch in peripheral vision and while the eye moves, so sweeping your gaze along the bar beats fixating on a candidate. Second, chasing the noise: the dither throws up light and dark cells that look like edges, and the giveaway is that a real seam runs the full height of the bar while a noise cell does not. Third, defaulting to the middle: the marker starts at fifty per cent and the seam almost never lands there. Fourth, trusting the brighter half: a step looks better defined from the lighter side, which biases placement toward the darker one. Finally, tilting a laptop screen changes low-contrast visibility more than anything you will do with your eyes.

Practice progression

(5 min, beginner) Play today's four rounds and, before you place the marker, sweep your eyes along the bar twice rather than hunting cell by cell. Watch the reveal each time, because that is the calibration. (10 min, intermediate) Sit back further from the screen. Distance shrinks the noise grain faster than it shrinks the seam, which is a genuinely useful technique rather than a cheat. (weekly) Read simultaneous contrast and color contrast and accessibility together, then check a real gradient or a photograph for banding you had not noticed.

The science behind it

The visual system is built to find edges, not absolute levels: early processing emphasises local differences, which is why a step of a few units can be visible at all when the two sides are almost identical. That same machinery is why context distorts what you see, the territory Kingdom surveys in his review of lightness and brightness: a patch shifts in apparent lightness depending on what surrounds it, and gradual changes get filled in perceptually rather than measured. This game turns the problem into signal detection. The seam is the signal, the seeded dither is the noise, and each round lowers the one while raising the other, so you are practising a decision under uncertainty as much as the seeing. The underlying axis is described in hue, saturation and lightness.

Limitations

Hueseam is a game and an informal self-check, not a clinical or diagnostic tool. It runs on an uncalibrated screen at an unknown brightness and viewing distance, so results from two devices are not comparable, and a phone in sunlight will score worse than the same phone indoors. It measures the localisation of one kind of low-contrast lightness edge, and says nothing about hue discrimination, acuity, or contrast sensitivity as a clinician would assess it. If a low score comes with a real change in your vision, see an optometrist rather than play another round.

Where it shows up

Finding a faint edge is a working skill wherever a gradual thing is supposed to look seamless. A retoucher blending two exposures checks for the line where the blend stops being invisible, because that seam will surface later on someone else's brighter monitor. A designer exporting a large background gradient looks for banding, the flat stripes that appear when eight bits per channel run out of steps, and the only way to catch them is to see edges that are near your threshold. A print operator watching a long tonal ramp comes off press is looking for exactly the same defect in ink. A microscopist reading faint boundaries in a grainy field does it as the core of the job, with the grain at the same scale as the thing being looked for. Anyone laying floor or fitting fabric checks two batches under real light for the join that will announce itself as a stripe once the room is finished. In each case the difference is small, the surface is textured, and the edge either resolves or it does not. Hueseam gives you that decision four times a day, with the answer revealed straight afterwards.

Ready to find today's seam? Four bars, a fainter step and more grain each time, in about a minute, with a fresh seam tomorrow.

Play today's Hueseam ↑

Frequently asked questions

Is the seam always in the same place?

No. It is drawn afresh between fifteen and eighty-five per cent across each round, and either side may be the lighter one.

Why does the bar look grainy?

Deliberately. A fine luminance dither is laid over the whole bar, so the step has to be found in noise rather than spotted as a clean line.

Do I get to see the answer?

Yes, every round. After you lock in, the bar is repainted without the noise and the true seam is marked, so the step becomes obvious in hindsight.

Does it get harder as I go?

Yes, in three ways at once across the four rounds: the lightness step shrinks, the noise amplitude grows, and the tolerance for your placement tightens.

Can I use the keyboard?

Yes. Tab to the slider and use the arrow keys, or drag on the bar if you prefer a pointer. Both drive the same marker.

Related reading & games

Source: Kingdom, F. A. A. (2011), Lightness, brightness and transparency: A quarter century of new ideas, captivating demonstrations and unrelenting controversy, Vision Research 51(7), 652-673.

Understand it, now train it. The daily games stay free forever. A Personal Pass adds progress tracking (trends + export); Teacher / School tools let you assign it and follow a whole class. For teachers & self-learners →

Privacy · Cookies · Terms · About · Contact