Pick a color to see full summary.
Click "Find Closest Names" to search.
Heuristic naming tuned for designers. Switch styles on the left for different flavors.
Lightness-based tint/shade ramp. Click chips to copy HEX.
CIE L*a*b* Color Space (CIE 1976)
The CIE L*a*b* (CIELAB) space represents colour in three dimensions: L* (lightness, 0–100), a* (green–red axis), and b* (blue–yellow axis). It is derived from CIE XYZ via a cube-root transfer function and is designed to be approximately perceptually uniform.
This tool uses Lab as the basis for computing colour distances (ΔE76 and CIEDE2000), making name lookups perceptually meaningful rather than simple RGB Euclidean distance.
Berlin & Kay: Basic Color Terms (1969)
Berlin and Kay’s landmark study identified 11 basic color terms that appear across languages in a universal hierarchy: black, white, red, green, yellow, blue, brown, purple, pink, orange, gray. This work underpins modern colour naming research and the family classification system used in this tool.
Languages with fewer basic terms always follow the hierarchy — e.g., a language with only 3 terms will have black, white, and red.
Munsell Color System (1905)
Albert Munsell defined colour in Hue, Value (lightness), and Chroma (saturation). Munsell names like “5R 4/14” encode hue position, value step, and chroma level. The system remains a reference for perceptual spacing and is used in soil science, dental ceramics, and art education.
ΔE76 vs CIEDE2000
ΔE76 is simple Euclidean distance in Lab space. It is fast but not perceptually uniform for all regions of Lab.
CIEDE2000 (ΔE₀₀) adds corrections for lightness, chroma, and hue non-uniformities, plus a rotation term for blue. It is the industry standard for colour difference evaluation.
ΔE 1–2 barely perceptible
ΔE 2–3 perceptible on close inspection
ΔE 3–5 clearly noticeable
ΔE > 5 distinctly different colours
AI-Style Naming Methodology
This tool generates names using deterministic heuristics, not machine learning. Three styles are available:
- Practical UI — maps hue/lightness/saturation to design tokens (e.g., “Steel Blue 500”, “Warm Gray 200”)
- Poetic / Brand — evocative metaphors based on colour temperature and saturation (e.g., “Arctic Glow”, “Ember Whisper”)
- Playful — associates colours with familiar objects/scenes (e.g., “Fizzy Sunrise”, “Cosmic Punch”)
All approaches are deterministic — same input always produces the same output. No ML, no network.
Color Family Classification
Colors are classified into families based on HSL hue ranges:
Yellow: 45°–70° Green: 70°–170°
Cyan: 170°–200° Blue: 200°–265°
Purple: 265°–300° Pink: 300°–345°
Gray: S < 8% Black: L < 15% White: L > 85% (with S < 8%)
The dictionary uses author-assigned categories (cat field) which may differ from hue-based classification.
NCS, RAL & Pantone Systems
NCS (Natural Colour System) is based on the opponent-colour model with six elementary colours. RAL provides standardised industrial colour codes across 2530+ shades. Pantone defines spot colours for print and fashion. This tool’s dictionary includes approximations from all three systems.
sRGB → XYZ → CIE L*a*b*
C_lin = C/12.92 if C ≤ 0.04045
C_lin = ((C+0.055)/1.055)^2.4 otherwise
Step 2: Linear RGB → XYZ (D65)
X = 0.4125·R + 0.3576·G + 0.1805·B
Y = 0.2127·R + 0.7152·G + 0.0722·B
Z = 0.0193·R + 0.1192·G + 0.9505·B
Step 3: XYZ → L*a*b* (D65: Xn=0.9505, Yn=1, Zn=1.0889)
f(t) = t^(1/3) if t > ε, else (κ·t + 16)/116
ε = 216/24389, κ = 24389/27
L* = 116·f(Y/Yn) − 16
a* = 500·[f(X/Xn) − f(Y/Yn)]
b* = 200·[f(Y/Yn) − f(Z/Zn)]
ΔE76 — Euclidean Distance in Lab
Simple, fast, but NOT perceptually uniform
for large distances or in the blue region.
CIEDE2000 (ΔE₀₀)
+ Rᵀᵀ·(ΔC′/k₂·S₂)·(ΔH′/k₃·S₃)]
Weighting functions:
Sₗ = 1 + 0.015·(L̅−50)²/√(20+(L̅−50)²)
Sₐ = 1 + 0.045·C̅′
Sₕ = 1 + 0.015·C̅′·T
T = 1 − 0.17·cos(H̅′−30°) + 0.24·cos(2H̅′)
+ 0.32·cos(3H̅′+6°) − 0.20·cos(4H̅′−63°)
Rᵀ = hue-rotation correction for blue (~275°)
Sharma, Wu, Dalal (2005). Industry standard for colour matching.
OKLab & OKLCh (Ottosson 2020)
l = 0.4122·R + 0.5363·G + 0.0514·B
m = 0.2119·R + 0.6807·G + 0.1074·B
s = 0.0883·R + 0.2817·G + 0.6300·B
Cube root: l′ = ∛l, m′ = ∛m, s′ = ∛s
OKLab:
L = 0.2105·l′ + 0.7936·m′ − 0.0041·s′
a = 1.9780·l′ − 2.4286·m′ + 0.4506·s′
b = 0.0259·l′ + 0.7828·m′ − 0.8087·s′
The summary also displays OKLab/OKLCh values for each color.
Tone Ramp Generation
For i = 1..N steps:
L_i = (i+1)/(N+1) × 100
S_i = S × (1 − |L_i − 50|/80)
S_i = max(S_i, S × 0.3)
HEX_i = HSL(H, clamp(S_i), clamp(L_i))
Result: N tint/shade chips from dark to light
with saturation tapered at extremes.
Naming Heuristic Rules
hueIndex = floor(H / 22.5) mod 16 → HUE_NAMES[i]
shade = L<25 "Dark" | L<45 "Deep" | L>80 "Light" | L>65 "Pale"
vivacity = S>80 "Vivid" | S<25 "Muted"
tier = round(L/10) × 100
Poetic:
adjective = POETIC_ADJ[hueIndex]
noun = POETIC_NOUN[lightnessIndex]
Playful:
adjective = FUN_ADJ[hueIndex]
noun = FUN_NOUN[lightnessIndex]
Relative Luminance & Contrast (WCAG)
Contrast Ratio = (L_lighter + 0.05) / (L_darker + 0.05)
WCAG AA: text ≥ 4.5:1, large text ≥ 3:1
WCAG AAA: text ≥ 7:1, large text ≥ 4.5:1
CIE Colorimetry
[2] CIE Publication 15, 1976 — Recommendations on Uniform Colour Spaces. Original CIELAB.
[3] CIE 142:2001 — Improvement to Industrial Colour-Difference Evaluation.
Color Difference Metrics
[5] Ottosson, B. (2020) — A perceptual color space for image processing. OKLab/OKLCh.
Color Naming Research
[7] Lindsey, D.T. & Brown, A.M. (2014) — The color lexicon of American English. JOSA A, 31(7), 1537–1548. Modern naming survey, 330+ terms.
Color Systems & Standards
[9] NCS — Natural Colour System, Scandinavian Colour Institute, 1981. Opponent-colour-based notation system.
[10] RAL colour standards, RAL gGmbH. Industrial colour system (2530+ shades).
[11] CSS Color Module Level 4, W3C, 2024. 148 named CSS colours specification.
General Resources
[13] Joblove, G.H. & Greenberg, D. (1978) — Color spaces for computer graphics. SIGGRAPH ’78, 20–25. HSL/HSV cylinder colour model.
About this tool
This tool implements research-grade HEX-to-name lookup via ΔE76 and CIEDE2000 in CIE L*a*b* space, AI-style heuristic naming (practical, poetic, playful), a 1050+ entry searchable color dictionary, tone-ramp palette generation, family classification, Lab a*b* scatter plot visualization, ΔE76 vs CIEDE2000 metric comparison, palette management, and batch analysis with CSV export — entirely client-side (zero network). Color approximations from NCS, RAL, and Pantone are not official matches.
Bar chart showing the ΔE distance of the 10 closest dictionary matches. Shorter bars = closer perceptual match.
Closest matches plotted on the CIE Lab a*–b* plane. The base color is shown as a larger circle with a dark outline.
Side-by-side top 5 results from both metrics to see how they differ.
Click any entry to set it as the base color. Dictionary sourced from CSS4, X11, Crayola, Pantone approximations, NCS, RAL.
Count and percentage of dictionary entries per color family (cat field).
No saved palettes yet.
Enter one HEX per line. Returns closest name, ΔE76, ΔE₀₀, family, and L*a*b* for each.