CIE76 (ΔE*ab) is the simplest formula — a straightforward Euclidean distance in LAB color space. It's easy to compute but doesn't account for human eye sensitivity variations across different color regions.

CIE94 (ΔE*94) introduced weighting factors (SL, SC, SH) to compensate for the eye's varying sensitivity to lightness, chroma, and hue differences. Two versions exist: one for graphic arts (K₁=0.045, K₂=0.015) and one for textiles (K₁=0.048, K₂=0.014).

CIEDE2000 (ΔE*00) is the most accurate and current ISO/CIE standard. It adds corrections for chroma-hue interaction (RT term), neutral-color behavior, and blue-region sensitivity. It's considered the gold standard for industrial color difference evaluation.

The interpretation depends on your application:

• ΔE < 1.0 — Imperceptible; colors appear identical to most observers
• ΔE 1.0–2.0 — Only detectable by trained observers under controlled conditions
• ΔE 2.0–3.5 — Barely perceptible; acceptable for most consumer-grade products
• ΔE 3.5–6.0 — Noticeable difference; may be acceptable for less critical applications
• ΔE 6.0–12.0 — Clearly distinct colors; generally unacceptable for brand-matching
• ΔE > 12.0 — Completely different colors; obvious mismatch

In professional printing, a ΔE below 3–5 is often targeted. In automotive paint matching, ΔE < 1.5 is typical. For consumer displays, ΔE < 2 is considered excellent.

CIE76 treats all regions of color space equally, but the human eye does not. We are more sensitive to certain hues (like blues and grays) and less sensitive to others (like saturated yellows). CIEDE2000 corrects for these perceptual non-uniformities, making its predictions much closer to what a human observer would actually judge. For example, two blue shades with the same CIE76 distance as two yellow shades may appear very different to your eye — CIEDE2000 accounts for this, while CIE76 would misleadingly report identical differences. In short: CIEDE2000 matches human perception far better.

The calculation involves multiple steps:

1. HEX/RGB → Linear RGB: Remove gamma correction (inverse sRGB transfer function)
2. Linear RGB → XYZ: Apply the sRGB-to-XYZ transformation matrix (D65 reference white)
3. XYZ → CIELAB: Convert to the perceptually uniform LAB color space using the CIE standard formulas with D65 illuminant reference values (Xₙ=95.047, Yₙ=100.000, Zₙ=108.883)
4. LAB → ΔE: Apply the chosen Delta E formula (CIE76, CIE94, or CIEDE2000) to the two LAB values

This tool performs all these conversions automatically in real-time, so you only need to input HEX or pick colors visually.

In commercial offset printing, a ΔE tolerance of 3–5 is common for general work. For brand-color matching (like Coca-Cola red or Tiffany blue), tolerances tighten to ΔE < 2–3. Proof-to-print matching typically targets ΔE < 3. ISO 12647-2 specifies tolerances for process colors: ΔE < 5 for CMY solids and ΔE < 4 for black. Pantone spot colors often have target tolerances of ΔE < 1.5–2.5 depending on the substrate and ink system.

Absolutely. Display calibration software (like DisplayCAL, CalMAN, or i1Profiler) extensively uses ΔE to measure how accurately a monitor reproduces colors. A factory-calibrated professional monitor typically achieves an average ΔE < 2 across the gamut. After hardware calibration, average ΔE values of 0.5–1.5 are achievable on high-end displays. Consumer-grade monitors often have average ΔE values of 3–6 out of the box. For color-critical work (photo editing, video grading), aim for average ΔE < 2 and maximum ΔE < 4.

CIELAB was designed to be perceptually uniform — meaning equal numerical distances in the space correspond roughly to equal perceived differences. While not perfect (CIEDE2000 exists precisely to correct its remaining non-uniformities), LAB is vastly more perceptually relevant than RGB or XYZ. The three axes map to human vision: L* represents lightness (0=black, 100=white), a* represents the green–red opponent channel (negative=green, positive=red), and b* represents the blue–yellow opponent channel (negative=blue, positive=yellow). This structure mirrors the opponent-process theory of human color vision.