When two colours look like a matching pair under one light source but noticeably different under another.
Metamerism occurs when two surfaces have different spectral reflectance curves that happen to produce the same perceived colour under one illuminant, but diverge under an illuminant with a different spectral power distribution.
Metamerism is not a flaw in the colours themselves — it's a property of how pigments and light interact, and can affect even carefully matched paint batches.
Two 'matching' paint samples from different product lines can look identical in the shop under fluorescent light and visibly different once home under daylight.
Metamerism is a genuine, well-documented reason two nominally matching colours can look wrong together in practice — always check a physical sample in the room it will actually be used in.
The CIE defines metamerism as the property of spectrally different stimuli that have the same tristimulus values. To test a pair, both reflectance curves are measured with a spectrophotometer and their colour difference is calculated under a reference illuminant and under one or more test illuminants. The difference under the test light, for a pair that matches under the reference, is a metamerism index. The quick version is a light booth: view the pair under daylight, a tungsten or warm LED source and a fluorescent source in turn.
| Light | Sample A looks | Sample B looks | ΔE00 between them |
|---|---|---|---|
| Daylight, CIE D65 | #383838 | #383838 | 0.0 |
| Daylight, CIE D50 (graphic arts viewing) | #383838 | #383838 | 0.3 |
| Incandescent tungsten, CIE A | #373938 | #353939 | 1.0 |
| Narrow-band tri-phosphor fluorescent, CIE FL11 | #393838 | #373137 | 5.2 |
| Red-green-blue mixed LED, CIE LED-RGB1 | #363838 | #2a3738 | 5.7 |
Under D65 the two samples are 0.0 ΔE00 apart: a match. Under tungsten light they drift to 1.0, still hard to see. Under the red-green-blue mixed led they are 5.7 apart, one grey turning #363838 and the other #2a3738. Lights built from a few narrow bands are the ones that pull a metameric pair apart, because they sample each reflectance curve at only a few points.
How this was calculated: Sample A is a smooth reflectance curve that reproduces Graphite #383838 under D65. Sample B adds the largest spectral ripple that leaves the D65 colour untouched. Both are modelled spectra, deliberately more different than most real pairs, each viewed by an observer adapted to the light.
A car panel is repainted, or a plastic trim part is moulded, to match an existing colour using different pigments.
The repair is approved under the workshop's lights and shows as a patch in daylight or under street lamps. Matching with the same pigments as the original, or checking the match under three light sources, is the remedy.
Black trousers and a black jacket are bought separately.
Different dyes on different fibres reach black by different spectra. Indoors they agree; in daylight one looks brownish or greenish against the other.