# Color difference measurement

Comparing manufactured colors using specified color spaces, optical conditions, sampling methods and acceptance rules.

Reviewed: 2026-11-01

Tags: inspection, color, metrology, quality-control

Record the reference, formula and measurement conditions with each color difference. ISO/CIE 11664-6:2022, edition 2 (August 2022), defines CIEDE2000 from CIELAB coordinates for reflecting or transmitting objects, excluding stimuli that appear as primary lights or their specular reflections. [ISO/CIE 11664-6:2022](https://www.iso.org/standard/82662.html)

## Coordinates and difference formulas

**Original arithmetic example:** suppose an approved reference is (50, 10, 20) and a specimen is (51, 12, 18), both measured using the same method. Their signed differences are (+1, +2, −2), so ΔE*ab = √(1 + 4 + 4) = 3. This means lighter, redder and bluer relative to that reference. It is neither a three-percent difference nor a calculation of ΔE00. If a hypothetical contract specifies ΔE*ab ≤ 2.5, the specimen exceeds that particular numerical limit. A contract specifying ΔE00 requires a separate calculation and its agreed settings.

CIELAB uses L* for lightness, a* along the green–red axis and b* along the blue–yellow axis. With differences defined as **specimen minus reference**, positive ΔL* means lighter, positive Δa* means redder and positive Δb* means yellower. Keep the signed components: a total difference alone hides the direction of a production shift. CIE76, also written ΔE*ab, is the Euclidean distance between the two CIELAB coordinate sets. CIEDE2000, written ΔE00, applies perceptual corrections involving lightness, chroma and hue. The formulas produce different numerical results; a limit established with one must not silently be applied to another. [Konica Minolta: Delta E and formula choice](https://sensing.konicaminolta.eu/mi-en/colourblogtop/colour-blog/whatisdeltae)

For CIE76:

```text
ΔE*ab = sqrt((ΔL*)² + (Δa*)² + (Δb*)²)
```

## Define the measurement method before comparing results

HunterLab identifies color scale, illuminant, standard observer, instrument geometry, sample preparation and sample presentation as necessary communication parameters. For example, D65/10° identifies a computational illuminant and standard observer; it does not replace the instrument's illumination/viewing geometry. Directional 45°/0° and integrating-sphere d/8° instruments can return different results. Record aperture, UV configuration where relevant, number of readings and repositioning pattern. Use the same preparation and method across reference and specimens and across supplier sites. [HunterLab: measurement parameters and inter-instrument agreement](https://support.hunterlab.com/knowledgebase/article/KA-06563/en-us)

For a sphere instrument, **SCI** includes the specular component and **SCE** excludes it. SCI is useful for comparing material color while reducing the influence of gloss differences; SCE retains stronger sensitivity to surface appearance and can be useful when correlating with visual comparisons. They answer different inspection questions. Select the mode in the agreed procedure rather than switching modes to obtain a passing result. A material-color match does not by itself establish a match in gloss or texture. [Konica Minolta: color, gloss, SCI and SCE](https://www.konicaminolta.com/instruments/knowledge/color/part3/02.html)

Translucency makes backing part of the specimen configuration. HunterLab demonstrates that the same plastic plaque gives different lightness results over white and black backings. Specify the backing, specimen thickness or layer arrangement and contact configuration appropriate to the product. Maintain them for every comparison; neither white nor black is a universal choice. A change in backing can create an apparent color shift without a change in the material. [HunterLab: backings for translucent samples](https://support.hunterlab.com/knowledgebase/article/KA-05343/en-us)

## Instrument specifications are bounded tests

As a named example, Konica Minolta's CM-17d product table specifies inter-instrument agreement within ΔE*ab 0.12, **averaged over 12 BCRA Series II tiles**, in MAV SCI mode against a master instrument under the manufacturer's standard conditions. Its repeatability specification is standard deviation within ΔE*ab 0.02 when measuring a white calibration plate 30 times at five-second intervals after white calibration, again under manufacturer conditions. These are different tests. Neither establishes uncertainty on a textured production part or an acceptable customer tolerance. The same table identifies an 8 mm measurement area within an 11 mm illumination area for MAV: the visible feature must suit the optical footprint, not merely the displayed number. [Konica Minolta: CM-17d/CM-16d specifications](https://sensing.konicaminolta.asia/product/spectrophotometer-cm-17d-cm-16d/)

## Sampling, visual assessment and acceptance

Repositioning and sampling should represent the inspected surface. A repeated reading without moving the instrument tests something different from readings at several locations on a mottled part. HunterLab specifically calls for consistent measurement area, averaging and presentation patterns and recommends establishing repeatability for the actual application. Displayed decimal places do not demonstrate that production measurements resolve differences that small. [HunterLab: method repeatability](https://support.hunterlab.com/knowledgebase/article/KA-06563/en-us)

Define the calculation order as well. In an original simplified example, two locations differing from the reference only by +2 and −2 in L* have a mean-coordinate ΔE*ab of zero, while their mean individual ΔE*ab is 2. An average can therefore hide local mismatch. Specify whether acceptance applies to each location, a maximum, an average or another agreed statistic, and retain individual results when local uniformity matters.

Instrumental comparison also needs a visual context. Metameric specimens can match under one light source and mismatch under another. Konica Minolta recommends controlled lighting, consistent sample presentation and visual comparison under multiple relevant sources, with instrumental assessment of metamerism. A low difference calculated for D65 alone cannot establish a match under every installation light. Set product-specific tolerances with the customer and assess borderline results using the agreed measurement capability and decision rule; do not adopt a generic “invisible below 1” claim as an acceptance specification. [Konica Minolta: visual evaluation and metamerism](https://sensing.konicaminolta.us/us/blog/best-practices-for-visually-evaluating-the-color-of-samples/)

## Inspection record

A practical record should let another inspector reproduce the comparison:

- Identify the part, lot, reference specimen or reference-data revision and inspected locations.
- Record instrument identity, verification status, geometry, SCI/SCE mode, aperture and relevant optical settings.
- State color space, illuminant, observer, difference formula and any formula parameters.
- Describe preparation, backing, orientation, reading count and aggregation; retain signed differences and individual readings.
- State the contractual limit, decision rule, result and any required visual or alternate-light assessment.

These fields form a reporting template, not a new acceptance standard. The controlling drawing, customer specification or approved inspection procedure determines which observations constitute conformity.

## Source scope and related reference

Manufacturer web guidance and the CM-17d specification page cited here did not display document revision dates; they were reviewed on November 1, 2026. The dated standard cited above is the August 2022 edition. Product specifications and suggested practices remain subject to their stated conditions.

[Inspection and quality documents](../../sourcing/documents/index.md)