Metamer Mismatching as a Measure of the Color Rendering of Lights

We propose a new method for evaluating the colour rendering properties of lights. The proposed method uses the degree of metamer mismatching for flat grey (constant reflectance of 0.5) quantified in terms of the metamer mismatch volume index proposed by Logvinenko et al. . The standard CIE color rendering index (CRI) [1] has been widely criticized, especially when evaluating LEDs. One of the key problems with it is that it is based on measuring the colour differences that arise across a small sample of 8 (sometimes 14) coloured papers. Not only may such a sample not represent what the colour differences for all other possible surface reflectances may be, it also gives manufacturers the opportunity to tune the spectra of their lights to perform well on the standard sample. The background for the proposed measure of colour rendering is the concept of metamer mismatching. Consider a colour signal XYZ (in CIE standard coordinates) observed under a first light. Metamer mismatching refers to the fact that the possible XYZ that might be observed under a second light are only constrained to lie within a convex volume of possible XYZ values. The size of the volume depends on the XYZ and the lights involved; however, the volume for flat grey is the largest. The metamer mismatch volume represents the range of possible XYZ that can arise under the second light and so provides a measure of how different the XYZ under the second light can be. As such, it provides a measure of the colour rendering properties of the second light relative to the first. Since the metamer mismatch volume scales with the intensity of the illuminant, it is helpful to normalize the mismatch volume by the volume of the object colour solid for that illuminant, since it scales correspondingly with intensity. The result is the metamer mismatch volume index (MMVI) . In terms of colour rendering, the larger the MMVI, the poorer the colour rendering is likely to be. Since the MMVI is volume based, it is more intuitive to consider MMVI^(1/3). The Metamer Mismatching Colour Rendering Index (MMCRI) is then defined as MMCRI = 1 – MMVI^(1/3), where the MMVI is that of flat grey. We have computed the MMCRI and CIE CRI for numerous light spectra and compared them. In particular, we measured the spectra of several commercially available LED lights of various correlated colour temperatures and also used the spectra of CIE standard illuminants. When plotted (unavailable in this text-only abstract), we see a good correlation between the two indices--an indication that the MMCRI behaves reasonably--but with notable differences for some illuminants. It is exactly such differences that the proposed new method is intended to reveal. A complete analysis is presented in the full paper. References [1] Method of measuring and specifying colour rendering properties of light sources. CIE Publication No. 13.3, 1995. [2] Logvinenko AD, Funt B, Godau C (2014) Metamer mismatching. IEEE Trans on Image Processing, 23: 34 43. Metamer Mismatching as a Measure of the Color Render g of Lights Hamidreza Mirzaei and Brian Funt School of Computing Sci nce, Simon Frase Univ rsity