CIE Method of Assessing Daylight Simulators
Daylight simulator is a “device that provides spectral irradiance approximating that of a CIE standard daylight illuminant or CIE daylight illuminant, for visual appraisal or measurement of colours” .
Development of the CIE Methods of Assessing Daylight Simulators
Although for decades after the acceptance of the CIE system (1931), Illuminant C was accepted and widely used in colorimetry, the practical implementation of Source C was limited to special laboratory use. In 1963, the Colorimetry Committee of the CIE decided to supplement the then existing CIE illuminants A, B, and C by new illuminants more adequately representing phases of natural daylight. These new illuminants (D55, D65, and D75) were defined by a new approach suggested by Judd et al.  based on Simonds’  method of reducing experimental data to characteristic vectors (eigenvectors) and calculating the relative spectral power distribution of daylight of any desired correlated color temperature.
As a consequence of this new approach, the new daylight illuminants have been defined theoretically (albeit based on experimental data), and it means that there are still no physically realizable light sources corresponding to the illuminants. It was clear to the CIE already in 1967 that daylight simulators were required that would serve as standard sources representing the daylight illuminants. As a first step, Wyszecki  published spectral irradiance distribution data on a number of daylight simulators and also suggested methods for evaluating how well these sources simulated the corresponding illuminants.
When asking the question “how close” a given source is to the illuminant of the same correlated color temperature, we must first decide how to measure this closeness.
The “fingerprints” of illuminants and light sources are their spectral power distributions, and for most of the colorimetric calculations, only the relative values are interesting, calculated from the spectral radiance or irradiance values and normalized to have the value of 100 at 560 nm or to have Y = 100.
The colorimetric properties of illuminants and light sources are generally described in terms of the x,y or u′,v′ chromaticity coordinates, the correlated color temperature, and very often the color rendering index. Although these properties may often give us sufficient information, for the more specific purpose of evaluating whether a given light source may or may not be considered an adequate realization (simulation) of the corresponding illuminant, more complex measures are needed. By and large, they can be divided into two groups: those comparing the spectral curves with or without applying weights to take the visual significance into consideration and those measuring the effect of the illumination on a selected group of object colors and then comparing either the change from one illuminant to the other (color rendering index or CRI type) or by calculating the color difference under the test source illuminant for pairs of samples which are perfect matches under the reference illuminant (metamerism index or MI type). In the early 1970s, a subcommittee of TC-1.3 on Standard Sources studied a number of proposals for different methods: based on MI-type indices for the visible range [5, 6, 7] and based on the effective excitation of three fluorescent samples for the UV range . At the Troy (1977) meeting of TC-1.3, a modification of the Ganz  proposal was adopted for the UV range evaluation: the method used three virtual metameric (isomeric) pairs for each illuminant, each consisting of a fluorescent and a nonfluorescent sample.
Chromaticity Limits of Daylight Simulators
As evaluated by the chromaticity limits both illuminants D55 and D75 are acceptable as simulators for Standard Illuminant D65, and D65 is acceptable as a simulator of both D55 and D75 (see also the category limits).
Visible Range Evaluation of Daylight Simulators
The method is based on the evaluation of the Special Metamerism Index: change in illuminants of five metameric pairs representative of practical samples in related industries. CIELAB coordinates of these sample pairs are calculated for the reference illuminant and the simulator source. The color difference between each standard and the respective comparison specimen is very near to zero for the reference illuminant; the average color difference for the five pairs under the simulator gives the quality grade for the daylight simulator.
Quality classification of daylight simulators 
M v or M u
Determine the relative spectral irradiance of the light source in the 380–780 nm range
Calculate the tristimulus values and the CIELAB coordinates for each of the five standards and the five comparison specimens under the reference illuminant and under the test light source
Calculate the CIELAB color differences between each standard and the respective comparison specimen under reference illuminant (should be near zero) and under the test light source
Calculate the average of the five color difference values
Determine the quality grade using Table 1
Ultraviolet Range Evaluation of Daylight Simulators
F(λ) is the ratio of the spectral distribution of radiance due to fluorescence to the sum of the tabulated values of this distribution, i.e., Σ λ F(λ) = 1.0. F(λ) is identical for the three fluorescent standard specimens and is independent of the SPD of the illumination.
βR(λ) is the ratio of the radiance due to the reflection of the medium in the given direction to the radiance of a perfect reflected diffuser identically irradiated .
As can be seen from the definition equations, the βF(λ) and thus the βT(λ) values depend on the SPD of the illumination, so we would have curves for the other CIE illuminants different from those in Fig. 7.
Determine the relative spectral irradiance of the light source S(λ) in the 300–700 nm range
Take Q(λ′) for each of the three standards from the data illustrated in Fig. 4
Calculate N for each of the three standards from Eq. 1
Take the F(λ) values for each of the three standards from the data illustrated in Fig. 5
Calculate βF(λ) for each of the three standards from Eq. 2
Take βR(λ) for each of the three standards from the data illustrated in Fig. 6
Take the spectral (reflected) radiance factors for the three comparison specimens (like those illustrated in Fig. 8 for specimen 3)
Calculate the tristimulus values and the CIELAB coordinates for each of the three standards and the three comparison specimens under the reference illuminant and under the test light source
Calculate the CIELAB color differences between each standard and the respective comparison specimen under the reference illuminant (should be near zero) and under the test light source
Calculate the average of the three color difference values under the test light source
Determine the quality grade using Table 1
In the example illustrated in Fig. 9, M u = 0 for D65 (standard and comparison specimens are identical); M u = 0.79 (quality grade C).
Practical Application of the CIE Method
In the case of color measuring spectrophotometers measuring nonfluorescent samples, the SPD of the light source has no relevance only for their visual evaluation. For fluorescent samples, both the visible and the UV range evaluation is of importance. When classifying daylight simulators, generally both quality grades are given: first the visible range metamerism index Mv, then the UV range index M u . According to the CIE standard , “daylight simulators having [BC] grades have been found useful for many applications.” It is also interesting to note here that both illuminants D55 and D75 classify as grade CC simulators, while illuminant D50 is a grade DD simulator of the D65 standard illuminant.
Some national and international standards e.g.,  also consider quality grade BC as acceptable for critical match in visual evaluations. For the classification of instruments, the ASTM Standard Practice 991–11  states that the “requirement that the instrument simulation of CIE D65 shall have a rating not worse than BB (CIELAB) as determined by the method of CIE Publication 51 has often been referenced.” This standard comes with the caveat that “the method of CIE 51 is only suitable for ultraviolet excited specimens evaluated for the CIE 1964 (10°) observer. The methods described in CIE 51 were developed for UV activated fluorescent whites and have not been proven to be applicable to visible-activated fluorescent specimens.”
There are different technologies available for realizing daylight simulators for visual assessment: filtered tungsten lamps, dichroic lamps, filtered short-arc xenon lamps, fluorescent lamps, and LED-based lamps. In color measuring spectrophotometers, filtered pulsed xenon lamps are used nearly exclusively as daylight simulator sources. A detailed description and evaluation of the different implementations was described in CIE Publication no. 192  and some additional details in .
Many of the commercially available booths for visual assessment under D65 are quality grade BC to BE, i.e., they are acceptable daylight simulators in the visible range, but only one with fluorescent lamps and one with filtered tungsten and additional UV lamps were found to be acceptable in the UV range. For D50 and D75, the results were even worse; none complied with quality grade BC criterion. Well-calibrated color measuring spectrophotometers can have excellent (quality grade AB or BA) D65 simulators, but there are no reports of instruments equipped with D50 or D75 simulators [14, 15].
- 1.CIE Standard S 012/E: Standard method of assessing the spectral quality of daylight simulators for visual appraisal or measurement of colour (2004)Google Scholar
- 4.Wyszecki, G.: Development of new CIE standard sources for colorimetry. Die Farbe 19, 43–76 (1970)Google Scholar
- 7.Richter, K.: Gütebewertung der Strahldichteangleich and die Normlichtart D65. Lichttechnik 24, 370–373 (1972)Google Scholar
- 10.CIE: A method for assessing the quality of daylight simulators for colorimetry, publication no. 51. Central Bureau of the CIE, Vienna (1981)Google Scholar
- 11.CIE: A method for assessing the quality of daylight simulators for colorimetry, publication no. 51.2. Central Bureau of the CIE, Vienna (1999)Google Scholar
- 12.ASTM D1729-96: Standard Practice for Visual Appraisal of Colors and Color Differences of Diffusely-Illuminated Opaque Materials. ASTM International, West Conshohocken, PA (2009)Google Scholar
- 13.ASTM E991: Standard Practice for Color Measurement of Fluorescent Specimens Using the One-Monochromator Method. ASTM International, West Conshohocken, PA (2011)Google Scholar
- 14.CIE: Practical Daylight Sources for Colorimetry, Publication no. 192. Central Bureau of the CIE, Vienna (2010)Google Scholar