Delta-E Calculator: CIE76, CIE94 & CIEDE2000

Pick two colors and see their difference under all three standard Delta-E formulas at once, with what each number means.

Short answer: enter two colors below and this calculator returns three numbers: ΔE76, ΔE94 and ΔE00 (CIEDE2000), each measuring how different the two colors are in the perceptually-oriented CIELAB space. They are not the same number for the same pair, because each formula corrects the space's unevenness differently. As a rule of thumb, a CIEDE2000 value of roughly 1 sits at the boundary most people can just notice; CIE76 tends to report a noticeably larger figure for the identical pair. This tool computes all three so you are not stuck guessing which one a source meant.

How to use this calculator

Set two colors with the pickers or by typing hex codes directly. The two swatches above the table update immediately so you can see the pair side by side, and all three Delta-E values recompute as you type. The line beneath the table gives a plain-language read of the CIEDE2000 result, since that is the formula most current guidance treats as the reference. If you already know roughly how different two colors need to be for your use case, jump to the worked examples below for concrete numbers to calibrate against.

Three formulas, one question

All three formulas answer the same question, "how far apart are these two colors," but they take increasingly careful routes to the answer. CIE76 is the simplest: the straight-line (Euclidean) distance between the two points in CIELAB space. It is fast and easy to implement, and it is also the least accurate, because CIELAB is not as perceptually even as its designers hoped; the same numeric distance can look like a big jump in one part of the space and be invisible in another.

CIE94 keeps the same lightness, chroma and hue decomposition but divides the chroma and hue terms by weighting factors that grow with the color's own chroma, which shrinks the reported difference for highly saturated colors relative to CIE76. It was a deliberate patch, not a rebuild, and it is still common in graphic arts and textile tolerancing workflows that adopted it in the 1990s and never fully moved on.

CIEDE2000 goes further still, adding a rotation term that specifically corrects the blue region where CIELAB misbehaves most, plus refined weighting for lightness, chroma and hue. It is the formula most current standards and color-management tools treat as the reference, and it is also the most complex to implement correctly, complex enough that its original authors published a dedicated paper of worked test cases purely so other implementers could verify their code against known-correct answers.

Which formula should you use?

If you are matching current industry practice or need the most perceptually accurate figure, use CIEDE2000; it is the current reference and what most modern color-management and print-QC tools report by default. Use CIE76 only when you need a fast estimate, are working with legacy tooling that already reports it, or are comparing against an old dataset that was measured in it, since converting old CIE76 tolerances to CIEDE2000 is not a simple scaling operation. CIE94 sits in between, and you will mostly meet it in graphic-arts and textile specifications written before CIEDE2000 became standard; treat a bare "Delta-E" figure from that world as CIE94 unless told otherwise, and always check which formula a spec actually names before you sign off on a tolerance.

Worked examples

Numbers make the disagreement concrete. Two near-identical reds, #ff0000 and #ff2a00, the calculator's own default pair, score ΔE76 = 5.37, ΔE94 = 2.37 and ΔE00 = 2.55. CIE76 alone would suggest a clearly different color; the other two agree the difference is real but modest, on the edge of "noticeable on close comparison." A pair of near-identical blues, #3366cc and #3366cd, differing by a single unit in the blue channel, comes out at ΔE76 = 0.60, ΔE94 = 0.20 and ΔE00 = 0.13, all comfortably under the just-noticeable range, which matches the fact that no one could actually tell these two apart on a screen. At the other extreme, primary red against primary green, #ff0000 versus #00ff00, produces ΔE76 = 170.58, ΔE94 = 73.43 and ΔE00 = 86.62: the formulas disagree sharply on the exact number but agree completely that the colors are nothing alike. Try all three pairs in the calculator above to see the swatches alongside the numbers.

Reading the just-noticeable boundary

The CIEDE2000 result is the one worth anchoring to a rule of thumb, since it is the current reference formula. Treat the bands below as guidance, not a hard wall; where a real threshold actually falls depends on the two colors, the size and shape of the patches, and the individual observer, which is the whole subject of our just-noticeable color difference guide.

For the deeper question of why these thresholds are averages rather than fixed constants, including how MacAdam's classic 1942 experiment first showed that color space is not evenly sensitive in every direction, see our Delta-E and the just-noticeable limit article, which this calculator complements rather than repeats.

Limitations

All three formulas here are computed from the sRGB values you enter, converted to CIELAB under a D65 white point with a standard gamma decode. That means the result assumes both colors are being viewed under comparable, well-lit conditions on a reasonably calibrated screen; real proofing and quality-control work in print or textiles uses a spectrophotometer under controlled illumination, not a monitor color picker, and will not exactly match what this page reports. CIE94 here uses the graphic-arts reference constants rather than the textile variant, which gives a slightly different number for saturated colors; if your specification names the textile constants, expect a small discrepancy. None of the three formulas account for color vision differences either: a Delta-E of 5 is "clearly different" for typical color vision but can still be a difficult call for some viewers, a separate question covered by our color blindness simulator.

Related reading

• Delta-E and the just-noticeable color limit: why CIE76 and CIEDE2000 disagree, in depth.
• The just-noticeable color difference: MacAdam's ellipses and what actually moves a threshold.
• How many colors can humans see?: the just-noticeable difference is what sets that count.

Sources: Sharma, G., Wu, W. and Dalal, E. N. (2005), The CIEDE2000 color-difference formula: implementation notes, supplementary test data, and mathematical observations, Color Research & Application 30(1), 21; International Commission on Illumination, CIE 116-1995, Industrial colour-difference evaluation (the basis of CIE94); Berns, Billmeyer and Saltzman's Principles of Color Technology, 4th ed., publisher page.