A light defined by its spectrum on paper, so that colour calculations can say exactly which light they assume.
A table of relative spectral power by wavelength published by the CIE. A is incandescent tungsten, D65 and D50 represent daylight, and the F and LED series represent common lamp types. A real lamp only approximates one.
An illuminant is a specification; a light source is a physical lamp. 'D65 lighting' in a viewing booth is a source built to approximate the D65 illuminant.
The same CIELAB value is a different colour if it was computed under D50 instead of D65.
Every measured or calculated colour is only meaningful with its illuminant stated, which is why metamerism and print proofing depend on it.
An illuminant is a published table, not a reading: relative power at each wavelength, issued by the CIE. A is tungsten light at about 2,856 K, D65 and D50 are daylight at about 6,504 K and 5,003 K, the F series represents fluorescent lamps and the LED series, added in 2018, represents white LEDs. A real lamp is compared with the illuminant it is meant to simulate by measuring its spectrum with a spectroradiometer. For daylight simulators the CIE grades the result by how far metameric pairs that match under the true illuminant drift apart under the lamp.
| Illuminant | White x, y | Surface x, y | Surface Y |
|---|---|---|---|
| Daylight, CIE D65 | 0.313, 0.329 | 0.481, 0.355 | 29.0% |
| Daylight, CIE D50 (graphic arts viewing) | 0.346, 0.359 | 0.510, 0.369 | 30.6% |
| Incandescent tungsten, CIE A | 0.447, 0.407 | 0.571, 0.385 | 36.2% |
Under D65 the surface has chromaticity 0.481, 0.355; under illuminant A it is 0.571, 0.385, and its luminance factor rises from 29.0% to 36.2% because tungsten light is rich in the long wavelengths a terracotta reflects. A colour value quoted without its illuminant could be any of these rows.
How this was calculated: A modelled reflectance curve for Terracotta #e2725b is multiplied by each CIE illuminant table and integrated against the 1931 2° observer, with no chromatic adaptation applied.
A supplier measures with D65 and the 10° observer and the customer with D50 and the 2° observer.
Both report honest CIELAB values for one sample that disagree by several units, and the batch is rejected over a difference that exists only on paper. Textile work conventionally uses the first combination and print the second, so cross-industry jobs must state theirs.
Colours are assessed in a booth whose 'daylight' tubes are a poor match to D65 or D50.
Pairs that will match in real daylight look mismatched in the booth, or the reverse, and fluorescent whites are misjudged if the tubes lack ultraviolet. The booth's grade as a simulator matters as much as its nominal colour temperature.