♿ Game Accessibility Lab: Colour Vision, Contrast and Target Size

2D interactive simulation · educational model
The science

Colour-vision deficiency. Each pixel is converted to linear light and multiplied by a 3×3 matrix (Machado, Oliveira & Fernandes 2009, severity 1.0) for protanopia, deuteranopia or tritanopia, then converted back to sRGB. Every matrix row sums to 1 so greys and white are unchanged. Real people are often anomalous trichromats (partial loss), so full dichromacy is the strong case; the simulation cannot reproduce a person's actual experience.

Contrast. WCAG relative luminance L = 0.2126 R + 0.7152 G + 0.0722 B on linear channels; contrast ratio (L1 + 0.05)/(L2 + 0.05). WCAG 2.x asks for 4.5:1 (AA) or 7:1 (AAA) for normal text, 3:1 for large text (≥ 24 px or 18.7 px bold). The 4.5/7/3 thresholds are conventions, not a physical boundary; low-vision users may need more.

Colour distance. Colours are converted XYZ (D65) → CIELAB and compared with ΔE*ab = √(ΔL² + Δa² + Δb²) (CIE76). A just-noticeable difference is about 2.3; for game markers that must be told apart at a glance, ΔE ≥ ~30 is used here as an illustrative comfort threshold, not a standard.

Pointing. Fitts' law: ID = log2(D/W + 1) bits and MT = a + b·ID. The constants a = 0.20 s, b = 0.15 s/bit are typical mouse-order values chosen for illustration; your own result will differ (use the target test). With hand tremor modelled as 2-D Gaussian scatter σ around the aim point, a click on a square target of side W hits with probability P = erf(W/(2√2 σ))²; expected time with retries is MT/P.

Target size. WCAG 2.2 asks for at least 24×24 CSS px (2.5.8, AA) and recommends 44×44 (2.5.5, AAA). Redundant cues (shape, icon, number) mean information is never carried by colour alone (WCAG 1.4.1).

Honest simplifications. Static colour model without screen gamut or ambient light; CIE76 rather than CIEDE2000; Fitts constants not fitted to any user; tremor treated as isotropic noise; the screen is drawn in 1000×560 units that stand in for CSS pixels.