Huespan

What Huespan trains

Huespan works your interval reproduction in color: seeing how large a lightness step is, then recreating that same step from a different starting shade. What you match is the gap itself, not either endpoint. That is subtler than picking a shade out of a line-up, because your pair never shares a starting point with the reference. Each day fixes one hue and one saturation for the whole puzzle, leaving lightness as the only dimension carrying the signal.

How to play

Every round shows a reference pair, A → B. Read its size and direction: the step may run lighter or darker, and both appear during a session. Below it sits a new anchor shade, C, and an empty slot, D. Move the slider to set the signed step from C, watch D repaint live, and press Lock in D. The slider covers both directions and resets to zero each round, so you build every step from scratch. Keyboard users tab to it and adjust with the arrow keys; a live region announces each value and result. There are four rounds: the interval shrinks and the tolerance tightens as you go.

How scoring works

Each round is scored on the absolute difference between the step you produced and the reference step. That error is divided by the round's tolerance and turned into an accuracy from 0 to 100, so the same error costs more in a tight late round. Your daily result is the plain 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. Miss the spot-on threshold and the answer is revealed: D repaints to the true value, the slider jumps to the target step, and the game waits for you to continue.

What your score means

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

One day is noisy: screen brightness, ambient light and the hue you were dealt all move the number. The trend across many days is the honest signal.

Common mistakes

The commonest error is matching an endpoint instead of a gap: making D look like B, which only works when C happens to sit near A. The game keeps C away from A, so the shortcut fails. A second trap is losing the direction: the reference steps darker while you build a lighter step of the right size, which scores as a large error. Third, many players compress large intervals and stretch small ones, pulling every answer toward a middling step. Fourth, the surround changes what you see: a step near the background lightness looks bigger than the same step further from it.

Practice progression

(5 min, beginner) Play today's puzzle and, before touching the slider, name the step out loud as a size and a direction, such as "a bit lighter" or "clearly darker". Naming it first stops you sliding around hunting for a match. (10 min, intermediate) In the practice tab, shrink the interval width and lock in your first estimate without fine-tuning. (weekly) Pair it with Huetween, which asks for a midpoint rather than a copied step, and read color midpoints and interval perception.

The science behind it

Lightness perception is not linear in physical luminance, which is why copying a step is harder than it sounds. Discrimination is sharpest near the background level and degrades away from it, the effect Whittle mapped by having people build equal-interval brightness series, finding a sharp minimum in the luminance steps at the background luminance. The perceptual scales behind hue, saturation and lightness try to flatten that curve so equal numeric steps look equal; they only approximately succeed, and this game exposes the residual error. Context adds a second layer, the family of effects covered in simultaneous contrast.

Limitations

Huespan is a game and an informal self-check, not a clinical or diagnostic tool. It runs on an uncalibrated screen in an uncontrolled room, so results from two devices are not comparable. It tests lightness intervals only, and says nothing about hue discrimination or color vision more broadly. A single low day is more likely tiredness or glare than anything about your eyes; if you have a real concern, see an optometrist.

Where it shows up

Copying an interval is quiet, constant work in anything that has to look evenly graded. A designer building a tonal ramp for an interface needs each step, from surface to raised surface to hover state, to feel the same size, or one row of a table jumps forward while the rest recede. A photographer lifting shadows in two frames of one set applies the same perceived change to both, though the frames start at different exposures, or the pair will not sit together on a page. A painter mixing a shadow for a second object shifts it as far from its local color as the first shadow was from its own. A print operator spots at once when one band of a gradient is wider than its neighbours. Data visualisation makes the same demand: a sequential scale is only readable if a step near the light end means roughly what a step near the dark end means, which is what color difference metrics exist to formalise. In every case no reference pair sits beside the work. You carry the size of the step in your head and reproduce it somewhere else, and that is what Huespan drills.

Related reading & games

• Color midpoints and interval perception: why the middle of two colors is not where you expect it.
• Value and saturation in composition: even tonal steps as a design tool.
• Huetween (the midpoint) and Huedrift (hold a shade steady): neighbouring interval games.

Source: Whittle, P. (1992), Brightness, discriminability and the "crispening effect", Vision Research 32(8), 1493-1507.