What does Delta E measure? It expresses the distance between two colours as a single number, calculated from their positions in a colour space rather than from ink percentages. A Delta E of zero means the two measurements are identical. Everything above zero is a difference, and the size of the number is meant to track how visible that difference is to a person looking at the two samples side by side.
Which formula should be used? The original 1976 calculation treats all directions in colour space equally, which overstates differences in saturated colours. The 1994 version weights lightness, chroma and hue differently, and the 2000 version adds terms for hue rotation and for neutral greys. Delta E 2000 tracks perceived difference closely, and it is the one to name in a specification, because the same pair of samples scores differently under each formula.
What tolerance is realistic? Under Delta E 2000, a value below about 1 is invisible to an untrained eye, 1 to 2 is visible to a trained eye in a direct comparison, and above 3 is visible to anyone. Brand guidelines often set 2 for spot colours and 3 for process builds, with tighter limits on large solid areas where any variation shows. A tolerance of 1 across a whole print run is expensive to hold and rarely necessary outside luxury work.
Why do two passing samples still look different? Because a number averages. A Delta E of 2 driven by a lightness shift looks different from a Delta E of 2 driven by a hue shift, even though both sit inside tolerance. Measurement also depends on the instrument geometry, the illuminant, the backing and the spot chosen, so two labs measuring the same sheet can report different numbers without either being wrong. The specification has to fix those variables, or the argument is about method rather than colour.
How is the measurement taken? A spectrophotometer reads the spectral reflectance of the printed surface, usually at 45 degrees illumination and 0 degrees viewing, or with an integrating sphere, and converts it to Lab values against a standard illuminant such as D50. The sample is measured over a defined backing, often white, at agreed spot locations. Reading a small printed logo with a large aperture, or measuring over a different backing, moves the number more than the print did.
Can Delta E catch metamerism? Not on its own. Two samples can match under D50 and diverge under tungsten, because metamerism is a difference in spectral curves that happens to cancel under one light source. Catching it needs spectral data and a check under two or three sources, which is why a brand colour approval should include a visual check under controlled lighting rather than a number alone, and cross-batch colour matching keeps a reference piece for each batch for the same reason.
What should a buyer put in the specification? The formula, the tolerance per colour and per area, the instrument type, the illuminant, the backing, the spot locations, and who measures , supplier, buyer or a third lab. Add what happens when a sample fails: reprint, sort, or accept with a concession. Those lines turn a colour argument into a decision, and they make a colour control routine possible across batches and across suppliers.
Chengtop measures colour on printed cartons and labels in Wuhan and reports Delta E against the approved target with the formula and illuminant stated. Send the brand colour reference and the tolerance your quality team works to, and we will confirm how the job will be measured and what happens on a miss.


