Paste a palette and instantly see how well its colors separate under 8 vision channels — normal, red/green and blue/yellow deficiencies, achromatopsia and grayscale. Get a pairwise ΔE matrix, red-flagged confusion pairs, a lightness-ladder score and one-click fixes. All computed locally in your browser.
模拟采用 Machado 等人(2009)的线性 RGB 变换矩阵,作用于线性光(linear-light)空间,色弱按严重度在「单位矩阵 ↔ 完全型矩阵」之间线性插值;全色盲与灰度按 Rec.709 相对亮度直接去色。色差走 sRGB → CIE XYZ(D65) → CIE Lab,提供 CIE76 与 CIE2000 两种度量,并可叠加 OKLCH 感知距离。判读阈值:ΔE<5 几乎不可分(严重混淆)、5–10 难以区分、10–20 可区分、>20 清晰。图表配色建议任意两色在全部通道下 ΔE≥10。