Short answer
Two surfaces match when they produce the same three cone signals, and that can happen with quite different reflectance spectra. Change the light and the spectra are reweighted differently, so the signals separate and the match breaks. This is illuminant metamerism; it is strongest when the two surfaces use different pigments or dyes and when the new light is spiky.
The eye reduces a whole spectrum to three numbers, one per cone type, and colorimetry mirrors this with X, Y and Z. Any spectral difference that happens to add nothing to all three integrals is invisible. Such a difference is called a metameric black: a wavy spectral curve, positive at some wavelengths and negative at others, whose products with the three colour-matching functions under a given light each sum to zero. Add one to a surface's reflectance and, under that light, the colour does not change at all. Every metameric pair is a surface plus a metameric black. The site's metamerism explorer builds pairs in exactly this way from CIE data.
A metameric black is zero only for the light it was built against. Under another light the weighting of each wavelength changes, the positive and negative lobes no longer cancel, and a real colour difference appears. The table shows it for a pair built on this site: a warm grey and the same grey with the largest physically valid metameric black added for daylight D65. Under D65 they are identical by construction. Under D50 the difference is negligible and under incandescent A and the broad halophosphate fluorescent it stays small; under the phosphor LED it becomes just visible, and under the spiky tri-phosphor fluorescent and RGB LED illuminants it grows to a clear difference side by side. The pattern generalises: lights close to the reference in spectrum preserve matches; lights with peaks and gaps expose them.
Real metameric pairs are rarely this extreme or this mild; the size depends entirely on how different the two recipes are.
The CIE's approach is a special metamerism index for a change in illuminant: take a pair that matches under a reference light, compute their colour difference under a test light, and report that difference. The CIE's method for grading daylight simulators turns this round, using sets of pairs that match under true daylight and scoring the simulator by the average difference it introduces, with a separate index for ultraviolet effects on fluorescent samples. Industry practice follows the same logic: a match is checked under more than one standard source — typically daylight, an incandescent-type source and a fluorescent or LED store light — and the worst result decides.
Metamerism is almost certain whenever the two things being matched are made differently: a paint mixed to match a fabric, a plastic part beside a painted one, a replacement car panel beside factory paint, or a dye lot from a different supplier. It is least likely when both come from the same recipe, which is why the site's lighting guide advises keeping materials from one range and batch where possible. The practical defence is simple and non-negotiable: approve matches under the lights where the objects will actually be seen, not just under the booth or showroom light.
| Illuminant | Colour difference (CIEDE2000) | Visibility |
|---|---|---|
| Daylight, CIE D65 | 0.0 | not visible |
| Daylight, CIE D50 (graphic arts viewing) | 0.1 | not visible |
| Incandescent tungsten, CIE A | 0.4 | not visible |
| Cool-white fluorescent, CIE FL2 | 1.1 | just visible side by side |
| Narrow-band tri-phosphor fluorescent, CIE FL11 | 2.3 | clearly visible |
| Phosphor-converted white LED, CIE LED-B3 | 1.1 | just visible side by side |
| Red-green-blue mixed LED, CIE LED-RGB1 | 2.5 | clearly visible |
Why: The retail mix used different pigments from the original; they are metameric under daylight only.
Fix: Repaint the whole wall section to a corner, or have the original measured spectrally rather than matched by eye.
Why: The booth's daylight simulator hides a difference the store's LEDs reveal.
Fix: Add the store's actual light source to the approval conditions.
Each statement is labelled by kind — established fact, a standard’s requirement, observed market data, a convention, or Colourwise’s own interpretation or analysis — with the strength of the evidence behind it.
FactStrong evidence
Two specimens with identical tristimulus values for a given illuminant and observer are metameric if their spectral reflectance differs within the visible range.
Source: CIE 080-1989 Special Metamerism Index: Change in Observer; CIE (International Commission on Illumination) publications
StandardStrong evidence
The CIE's method for assessing daylight simulators computes a visible-range metamerism index as the average colour difference of sample pairs that match under the reference daylight, with an ultraviolet-range index for fluorescent samples.
Source: CIE 051.2-1999 A Method for Assessing the Quality of Daylight Simulators for Colorimetry
Colourwise analysisModerate evidence
For a warm grey and a metameric partner built against D65, the colour difference is zero under D65, stays well below one CIEDE2000 unit under D50, and is largest under the spiky RGB LED and tri-phosphor fluorescent illuminants.
Based on: Colourwise calculation with the CIE 1931 observer and CIE illuminant tables at 10 nm, Bradford-adapted and compared with CIEDE2000; see the table.
Caveat: A synthetic pair; the ordering of illuminants is robust, the magnitudes depend on the pair.
Source: CIE datasets (colour-matching functions, illuminants)
Reviewed 29 September 2026. Colourwise summarises its sources in its own words and does not reproduce standards text or proprietary colour data. Spotted an error? Tell us.