ΔE colour difference calculator

“ΔE” is not one number. The same two colours give different values under different formulas, and a tolerance written as “ΔE 2” means nothing until the formula is named. Enter a reference and a sample to see all four side by side.

Reference (the standard)

Sample (the thing being checked)

The same pair, four formulas
Formula	ΔE	What it is for
CIE76 (ΔE*ab)	4.13	Straight distance in CIELAB. Overstates differences between saturated colours.
CIE94, graphic arts	3.97	Down-weights chroma and hue differences as the reference gets more saturated.
CIE94, textiles	2.09	As above, with lightness differences counted at half weight.
CMC 2:1	2.27	Long-standing textile acceptability formula; tolerance ellipsoids vary with hue.
CIEDE2000 (ΔE00)	3.54	Current CIE recommendation; corrects blue-region and neutral-axis errors.

Where the difference is

Lightness (ΔL*): 3.90; Sample is lighter
Chroma (ΔC*): -0.60; Sample is duller
Hue (ΔH*): 1.23; Hue has shifted

Reference L*a*b* 38.5, 9.6, -46.3 · Sample 42.4, 8.3, -45.9 (D65, 2° observer, from sRGB).

How this works

Colours are converted from sRGB to CIELAB with a D65 white point and the 2° standard observer.
CIEDE2000 follows Sharma, Wu and Dalal's published implementation and reproduces their test data; CIE94 and CMC are the standard formulas with their usual application weights.
CIE94 and CMC are asymmetric: the reference's chroma and hue set the weighting, so swapping reference and sample can change the result.

Limitations

Values from sRGB screen colours are only as good as the HEX values: a measured sample needs a measured L*a*b* under stated conditions, not a colour picked from a photo.
No ΔE threshold is universal. What counts as acceptable depends on the industry, the material, the viewing conditions and the contract, and must name the formula.
ΔE formulas were fitted to judgements of small differences between uniform patches. They are unreliable for large differences, textured surfaces and very small or very large samples.

Where the difference is

Lightness (ΔL*): 3.90; Sample is lighter
Chroma (ΔC*): -0.60; Sample is duller
Hue (ΔH*): 1.23; Hue has shifted

Reference L*a*b* 38.5, 9.6, -46.3 · Sample 42.4, 8.3, -45.9 (D65, 2° observer, from sRGB).

How this works

Colours are converted from sRGB to CIELAB with a D65 white point and the 2° standard observer.
CIEDE2000 follows Sharma, Wu and Dalal's published implementation and reproduces their test data; CIE94 and CMC are the standard formulas with their usual application weights.
CIE94 and CMC are asymmetric: the reference's chroma and hue set the weighting, so swapping reference and sample can change the result.

Limitations

Values from sRGB screen colours are only as good as the HEX values: a measured sample needs a measured L*a*b* under stated conditions, not a colour picked from a photo.
No ΔE threshold is universal. What counts as acceptable depends on the industry, the material, the viewing conditions and the contract, and must name the formula.
ΔE formulas were fitted to judgements of small differences between uniform patches. They are unreliable for large differences, textured surfaces and very small or very large samples.

ΔE colour difference calculator

Where the difference is

How this works

Limitations

Read next

ΔE colour difference calculator

Where the difference is

How this works

Limitations

Read next