Volume 8 Issue 1 October 2004
##### How does the lighting industry measure color appearance?

Researchers and scientists, despite many attempts, have not been able to predict color appearance precisely, except under fairly restricted conditions (CIE 2004, 1986; Moroney et al. 2002; Fairchild 1998; Hunt, 1998; Nayatani 1997; Luo et al. 1996). However, the International Lighting Commission (Commission Internationale de l'Eclariage, referred to as CIE) established a colorimetry system for color matching that has, with minor changes, remained in use for nearly a century. Human color vision begins with absorption of light by the eyes' three cone photoreceptors. In the 19th century, scientists discovered that any light can be exactly matched in appearance with the proper combination of three different colored lights, known as primaries. They also discovered that color matching followed all the rules of linear algebra; addition, subtraction, multiplication, and division. The CIE system has been the foundation for all color calculations used by the lighting industry, in large part because color matching follows these algebraic rules. This three-primary principle is utilized today with color television and other electronic displays. By incorporating different amounts of just three highly saturated, red, green and blue primaries, a wide array of color perceptions can be created on the display.

Although lighting manufacturers publish spectral power distribution (SPD) data for their light sources, these data are cumbersome and more detailed than necessary for accurate, unambiguous color representation. Instead, the industry most commonly describes a light source's color appearance using chromaticity, which is derived from the SPD of the light source using the CIE system (CIE 1986). In that system the absolute amounts of the three primaries needed to match a given light are normalized so the sum of the amounts of the three primary lights equals one. In this way, any two of the normalized numbers give a complete description of a light source color. The two numbers used to describe a light source color mathematically are known as its chromaticity coordinates, or simply its chromaticity.

Light sources that have different SPDs but produce identical relative absorptions by the three cone types will have the same chromaticity. At the same luminance, these lights will also appear to be identical under the same viewing conditions. Light sources of this type are known as metamers; one metameric pair of light sources is shown in Figure 5.

 Figure 5. Spectral power distribution of two metameric light sources The SPD on the left is that of an incandescent lamp with a CCT of 2856 K. The SPD on the right is of a red, green and blue LED mixed spectrum that is metameric with the incandescent lamp.

Since it is known that the chromaticity of any light source can be determined by a linear combination of three primaries, it is possible to abandon the use of real primaries in favor of imaginary primaries that have some useful characteristics. The CIE 1931 system of colorimetry uses the photopic luminous efficiency function V(l) as one of the three imaginary primaries. In this way the CIE system of colorimetry was simultaneously integrated with the CIE system of photometry. Figure 6 compares the CIE 1931 two-dimensional chromaticity diagram with the CIE 1976 diagram, both of which utilize imaginary primaries. Also plotted in the diagrams is the blackbody locus, which represents the chromaticities of a blackbody radiator source, heated to incandescence.

 Figure 6. The CIE 1931 and CIE 1976 Chromaticity Diagrams

Appendix A demonstrates how the chromaticity of a light source can be calculated from its SPD and the three CIE 1931 color matching functions. Also included are the linear transformation equations for converting the color-matching functions in the CIE 1931 system into those for the CIE 1976 system. The main advantage of the CIE 1976 system of colorimetry is that distances within the chromaticity space are approximately "perceptually equal." That is, pairs of chromaticities separated by the same distance are presumed to be equally different in perceived color, no matter where on the CIE 1976 space they occur.

The true color appearance of a light source is too complex to be represented precisely by chromaticity for reasons previously discussed. However, the chromaticity of a light source is useful as an approximate representation of its color appearance. Lights with chromaticity coordinates at the bottom left of the diagram will generally appear blue, while those in the far right will appear red. Those near the blackbody locus will appear "white." Chromaticity diagrams like those in Figure 6 are often produced in color so that they resemble an artist's palette, but this approach is technically inappropriate despite its visual appeal.

Figure 7 shows the chromaticities of 67 commercially available "white" light sources (fluorescent, metal halide, mercury, and incandescent) plotted in a small portion of the CIE 1976 color space. Nearly all of the "white" light sources in Figure 7 fall close to the blackbody locus even though they are not incandescent sources. Given this close relationship between the blackbody locus and the chromaticities of "white" light emitted by these sources, the blackbody locus has become a useful reference line for describing the apparent colors of light emitted from electric light sources.

 Figure 7. Chromaticities of 67 commercial light sources plotted in the CIE 1976 color space