Republication of Postings from the U.S. Department of Energy (DOE) Solid-State Lighting Program
by Jim Brodrick, SSL Program Manager, U.S. Department of Energy
The emergence of solid-state lighting has challenged longstanding performance metrics used for assessing color quality, particularly the International Commission on Illumination (CIE) Test Color Method, commonly known as CRI. DOE technical support and leadership are assisting industry efforts to develop new methods for evaluating color rendition that address CRI’s widely acknowledged limitations. DOE played a key role in the development of IES TM-30-15, “Method for Evaluating Light Source Color Rendition” (the task group for which was chaired by a member of DOE’s SSL team) and is helping to facilitate its use by clarifying the complex technical issues involved. Webinars and presentations at major lighting conferences are augmented by online resources, including fact sheets and extensive FAQs.
Part of this DOE effort includes conducting foundational research to increase our understanding of color rendition, and reporting on it in peer-reviewed journals. The latest such article was published in the journal LEUKOS and is entitled “Chroma Shift and Gamut Shape: Going Beyond Average Color Fidelity and Gamut Area.”
CRI is a measure of average color fidelity, which means it’s intended to characterize the overall similarity of object color appearance between a test source and reference illuminant. And while more-accurate measures of average color fidelity are now available, average color fidelity is only one aspect of color rendition. When considered alone, average color fidelity has been shown to be poorly correlated with human perceptions of color naturalness, normalness, preference, and saturation. That is, regardless of the specifics of the color fidelity measure, a single-number quantification of average color difference is insufficient for characterizing human perceptions of object color appearance in architectural environments.
Gamut area, a measure of color saturation, has been considered an important complement to average color fidelity, and numerous iterations of gamut-area measures have been proposed. But it, too, is a limited characterization of color quality, providing no information about the rendition of objects having specific hues. Two light sources with equal average color fidelity and equal gamut area may be perceived differently because they distort different hues in different ways.
Though sometimes erroneously referred to as a two-measure system for evaluating color rendition, TM-30 includes key components that go beyond the two high-level average values, Fidelity Index (Rf) and Gamut Index (Rg). The journal article discusses two concepts in color rendition — gamut shape (defined as the average pattern of color shifts across different hues) and chroma shift — that have been recently formalized as part of TM-30. It discusses the calculation methods for these evaluation tools and provides context for the interpretation of the values. It illustrates why and how the TM-30 Color Vector Graphic and Local Chroma Shift values capture information about color rendition that’s impossible to describe with global average measures but is pertinent to more completely quantifying color rendition and understanding human evaluations of color quality in the built environment.
Both the Color Vector Graphic and Local Chroma Shift values are enabled by the greater number of samples in TM-30 (99, compared with eight in CRI), which allows the sample set to be divided into 16 hue-angle bins. The Color Vector Graphic allows rapid visual understanding of color rendition characteristics for different hues and enables high-level comparisons of light sources, but is not useful for establishing numerical specifications or identifying small differences between light sources. Local Chroma Shift values address this limitation via bidirectional, hue-specific measures of chroma change, which can be more informative than hue-specific color fidelity values, since the same color fidelity value may result from an increase, decrease, or no change in chroma. Local Chroma Shift values are denoted Rcs,hj, where j indicates the specific hue-angle bin. The values are percentages, typically between about -30% and 30% for sources with Rf ≥ 70, with negative values indicating a decrease in chroma and positive values indicating an increase in chroma.
The Color Vector Graphic and Local Chroma Shift values facilitate the prediction and understanding of how object colors rendered by a light source will be perceived in architectural environments, because certain hues are more influential during subjective evaluations. Neither average color fidelity, gamut area, nor a combination of the two can effectively identify how a light source will be perceived, because their computation requires averaging across all hues; and this limitation will carry forward with any future improvements or alternative measures that discard hue-specific information through the averaging process. The Color Vector Graphic and Local Chroma Shift values are more-granular measures that supplement blunt averages, although they also discard information about individual sample shifts in favor of localized averages that are more informative for predicting rendition of real object colors with a given hue.
DOE’s collaborative technical leadership has been instrumental in helping the lighting industry to look beyond CRI, in order to better characterize the color performance of LEDs. This will provide manufacturers and specifiers with more-complete information to effectively balance color quality and energy efficiency in a variety of lighting applications. The LEUKOS article adds to that body of information, and can be accessed online.