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

As is the case with any subjective experience, it’s a challenge to quantify color preference, which varies from person to person and is affected by age, culture, and application as well as by many other factors. At the same time, it’s important for specifiers and consumers to choose sources that render the color of objects in accordance with human preferences, and to be able to make intelligent tradeoffs between lighting color quality and other performance attributes.

The recent development by the Illuminating Engineering Society (IES) of TM-30-15 offers an improved tool for characterizing average color fidelity (the overall similarity of colors illuminated by a test and reference source) and average change in color saturation (i.e., vividness), using TM-30’s Fidelity Index (Rf) and Gamut Index (Rg). There’s also added information that extends beyond average fidelity and average gamut area, providing insight into how specific hues will be rendered. But how color-rendering measures relate to the preferences of end users in real-world applications remains an open topic, with little practical guidance available to specifiers.

A new study published in the journal Lighting Research and Technology and coauthored by several members of DOE’s solid-state lighting team provides initial evidence to support the effectiveness of the measures defined in TM-30 for characterizing subjective visual perceptions. In turn, this opens doors for engineering and specifying light-source spectra with different color-rendering characteristics than are commonly available today.

The researchers conducted an experiment to determine how subjective impressions of color quality depend upon the shifts the light source causes in the color appearance of objects. Each of 28 participants of varied age and gender evaluated 26 different lighting conditions in a room filled with objects that were selected to cover a range of hue, saturation, and lightness — with illuminance (20 fc) and chromaticity (3500 K, on the blackbody locus) kept constant for all lighting conditions.

By itself, average fidelity — especially CRI — was found to be a weak predictor of human perception. Nine of the 12 most-liked products had a CRI ≤ 73, indicating that the common practice of using CRI ≥ 80 as a cutoff point may exclude many preferred light sources. The findings also suggest that gamut shape (the specific changes in color appearance across different hues) is more important than average gamut area (average saturation level) for modeling human preference. In particular, a distinct preference for the increased saturation of reds was observed, which is the opposite of what occurs with most energy-efficient lighting on the market today. What’s more, CRI disproportionately penalizes increases in red saturation, due to underlying fundamentals of the calculation, which may have been a deterrent to the development of more-preferred sources in the past.

It’s important to note that TM-30 is only a tool and is not a solution in and of itself. To be of maximum help when choosing a light source, it should be used in combination with other information — e.g., chromaticity, intensity, distribution, efficacy, and cost. But as this new research shows, better tools can aid in the development of new light sources that effectively balance all aspects of lighting quality and better fit the needs of a given application.

For more information, download the journal article.