Short answer
Because many whites fluoresce. An optical brightener absorbs ultraviolet and re-emits it as blue light, so the instrument sees reflected light plus emitted light, and the emitted part rises and falls with the ultraviolet in the instrument's lamp. A reading is only meaningful with a statement of how that ultraviolet was handled: left as it comes, filtered out, or adjusted against a fluorescent reference to imitate a chosen daylight. Non-fluorescent samples are unaffected by any of this.
For an ordinary surface, reflectance at a wavelength is a property of the surface alone: double the lamp's power there and the reflected light doubles, so the ratio is unchanged. A fluorescent surface adds a second component. It absorbs energy at shorter wavelengths, often in the near ultraviolet, and emits it at longer ones. ASTM E991 describes the quantity an instrument records as the sum of an ordinary reflected part and a fluoresced part, and notes that the fluoresced part varies directly with the lamp's power in the exciting range. The curve the instrument draws is therefore partly a description of its own lamp. Around 450 nm a brightened paper or textile can show apparent reflectance above 100%.
Instruments deal with this in three ways. Some simply use the lamp as it is; two such instruments, or one instrument before and after a lamp change, will disagree on brightened whites. Some insert a filter that removes the ultraviolet altogether, which measures the substrate as if the brightener were not there. The third approach adjusts the amount of ultraviolet. Datacolor's technical note describes the usual arrangement in a sphere instrument: a pulsed xenon lamp emits proportionally more ultraviolet than D65, and a cut-off filter is moved partly into the beam until a fluorescent reference specimen reads its assigned whiteness. Because the sphere coating absorbs more ultraviolet as it ages and the lamp changes with use, that setting drifts and has to be renewed, weekly by that maker's advice, as a procedure separate from the daily white calibration.
A whiteness index is computed from colorimetric values, so it inherits the ultraviolet question. The paper industry runs two conventions side by side: CIE whiteness for D65 with the 10° observer, meant to represent paper seen outdoors, and CIE whiteness for illuminant C with the 2° observer, meant for indoor viewing with less ultraviolet. Each needs its own filter setting, established against its own reference papers. Textile practice often uses a different index and different reference fabrics again. The same note is blunt about the consequences: results from different calibration schemes should not be compared, a calibration made at one aperture or specular setting holds only for that aperture and setting, and even two units of one model can differ by a couple of whiteness units.
All of the above are attempts to make one lamp imitate one illuminant. They cannot answer how the sample would look under a different light, because the fluoresced part would change. The complete description is bispectral: illuminate with one narrow band at a time and record the whole emitted spectrum for each. ASTM E2153 calls this two-monochromator method the definitive one, since it yields a quantity independent of any instrument's lamp from which colour can be calculated for any illuminant and observer. It is slow and the instruments are rare, so it is used for reference specimens and research. It is also the proper approach for colours excited by visible light, such as fluorescent oranges and yellow-greens, for which E991 notes that methods developed for ultraviolet-excited whites are unproven.
| Approach | What the lamp does | What the reading describes | Limit |
|---|---|---|---|
| UV included, uncontrolled | Whatever ultraviolet the lamp happens to emit reaches the sample | Reflection plus an arbitrary amount of fluorescence | Differs between instruments and drifts as the lamp ages |
| UV excluded | A filter removes ultraviolet before the sample | The surface with ultraviolet-excited fluorescence switched off | Not how the sample looks in daylight; does nothing for dyes excited by visible light |
| UV calibrated | A filter is adjusted against a fluorescent reference to imitate a stated daylight | An estimate of appearance under that one illuminant | Valid only for that illuminant, reference type, aperture and specular setting |
| Bispectral | The sample is excited one narrow band at a time | A lamp-independent description usable for any illuminant | Slow, specialised instruments; reference work rather than production |
Why: The two lamps deliver different amounts of ultraviolet, so the brightener emits different amounts of blue.
Fix: Calibrate both instruments' ultraviolet against the same type of fluorescent reference, or agree to measure with ultraviolet excluded and judge the brightener separately.
Why: The sphere coating and lamp have aged and less ultraviolet reaches the sample than when the filter was last set.
Fix: Repeat the ultraviolet calibration against the fluorescent reference, and replace reference specimens on the schedule their supplier gives.
Why: One is working to the outdoor D65/10° convention and the other to the indoor C/2° convention, or their apertures differ.
Fix: State the illuminant, observer, calibration reference, aperture and specular mode with every whiteness value.
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.
StandardStrong evidence
For a fluorescent specimen the total spectral radiance factor is the sum of a reflected component, which is independent of the illumination's spectral distribution, and a fluoresced component, which varies directly with the illumination's power in the excitation range.
StandardStrong evidence
The accuracy with which a one-monochromator colour spectrometer measures a fluorescent specimen depends directly on how well its illumination simulates CIE D65, and the CIE has recommended no standard source corresponding to D65.
FactStrong evidence
In UV-adjustable sphere spectrophotometers a variable-position ultraviolet cut-off filter in front of a xenon source is set against fluorescent reference specimens to simulate D65, and the setting must be renewed periodically because the sphere coating's ultraviolet absorption and the lamp's ultraviolet output change with age.
Caveat: One maker's description of its own instruments; the weekly interval is that maker's recommendation, not a standard's.
ConventionModerate evidence
For paper and board, CIE whiteness for D65 with the 10° observer is associated with outdoor illumination and the method of ISO 11475, and CIE whiteness for illuminant C with the 2° observer with indoor illumination and ISO 11476; each uses its own ultraviolet calibration.
Caveat: As summarised in an instrument maker's note. The ISO catalogue pages for the two standards could not be read when this was written, so the standards themselves are not cited.
StandardStrong evidence
The bispectral (two-monochromator) method yields an instrument- and illuminant-independent characterisation of a fluorescent specimen, from which its colour can be calculated for any illuminant and observer.
StandardStrong evidence
ISO 13655 distinguishes measurement conditions by the ultraviolet content of the instrument's illumination, including one matched to D50 (M1) and one with ultraviolet excluded (M2), so that readings of brightened papers can be reproduced.
Source: ISO 13655:2017 Graphic technology — Spectral measurement and colorimetric computation for graphic arts images; ISO 13655 measurement modes M0–M3
Reviewed 6 October 2026. Colourwise summarises its sources in its own words and does not reproduce standards text or proprietary colour data. Spotted an error? Tell us.