Abstract
The naturalness and diversity of colors produced by a light source are important aspects of color rendering which are often difficult to reconcile with conventional illumination. With the advent of modern solid-state light sources almost any lighting spectrum is possible and a new wide range of possibilities for color rendering is available. What spectral profiles optimize these two aspects? We addressed this issue by studying computationally chromatic effects of a variety of light sources with arbitrary spectral distributions. A large number of chromaticity points were chosen around the Planckian locus ranging from 2,000 K – 20,000 K, and for each point a large set of metamers with variable degree of smoothness was generated using the Schmitt's elements method. The general color rendering index (CRI) and the volume of the Munsell set spanned in the CIELAB color space were calculated for each metamer. Consistently with previous findings obtained with smaller samples, the metamers with the maximum CRI at each chromaticity had smooth spectra and the metamers producing the maximum volume of the Munsell set had more structured spectrum. The product of CRI and volume of the Munsell set showed a regular variation in color space and was maximized for chromaticities in the yellow-green region and for spectra medium structured. Similar tendency was obtained with the product of CRI by the number of discernible colors in high-resolution hyperspectral images of 50 natural scenes illuminated by these metamers. In these optimal conditions CRI above 90 could be obtained. The present study shows that it is possible to reconcile chromatic naturalness and chromatic diversity in color rendering by optimizing the spectra and the chromaticities of lighting.
Fundação para a Ciência e a Tecnologia (PTDC/EEA-EEL/098572/2008).