Republication of Postings from the U.S. Department of Energy (DOE) Solid-State Lighting Program
by Jim Brodrick, U.S. Department of Energy
One of the panel discussions generating buzz at DOE’s tenth annual SSL R&D Workshop, held a few weeks ago in Long Beach, CA, was the one that focused on light spectrum and color perception. As you probably know, color issues are among the key obstacles to SSL adoption, which makes increasing our understanding of them critically important.
Characterizing the color of electric lighting is nothing new, but the emergence of SSL has created new controversies and raised new questions we’ve yet to resolve. One of the basic questions is how best to measure and describe the color of light and the way it renders colors. None of the existing or proposed metrics – including Color Rendering Index (CRI) and correlated color temperature (CCT) – is without limitations. And even supplementing them with additional metrics &ndash such as R9, which covers saturated red tones – doesn’t completely nail things down.
According to workshop panelist Jean Paul Freyssinier of Rensselaer Polytechnic Institute’s Lighting Research Center (LRC), no one metric can characterize color rendering with any degree of accuracy. The LRC advocates using CRI in combination with the gamut area index, which evaluates the relative area enclosed by sample colors on a chromaticity diagram when rendered by the source. Fellow panelist Cameron Miller of the National Institute of Standards and Technology (NIST) described an alternative metric, the Color Quality Scale, which was developed by NIST and addresses many of CRI’s limitations.
Miller’s main thrust, though, involved the potential advantages of discrete-spectrum light sources over continuous-spectrum sources. He explained that because of phosphor-conversion losses and the emission of light that’s not detectable by the human eye, broadband phosphor-converted white LEDs (pc-LEDs) are theoretically less energy-efficient than narrow-band LEDs, which have the potential to render colors equally well. Miller provided evidence that with three or four colors in discrete spectra that are properly tuned, high efficacy and excellent color rendering could both be achieved. But until all of the discrete emitters become efficient enough to make this practical, a hybrid approach that combines suitable pc-LEDs with discrete red emitters can serve as an interim solution.
Freyssinier questioned some of our most basic assumptions in rethinking how we define white light. The color-related standards for LEDs, fluorescent lamps, and many other types of light sources are based on the black-body curve (BBC), which he suggested is not the most suitable reference to use for white light. As an alternative, Freyssinier described a reference he and his colleagues at the LRC developed. Called the white-body curve, it’s slightly above the BBC at high color temperatures and below it at low color temperatures, and he cited studies showing that it corresponds to human perception of what is “white” light.
The color of light also affects us physiologically, which was the basis for the talk on spectrally enhanced lighting (SEL) given by panelist Brian Liebel of the Lighting Partnership. A design method for interior lighting applications, SEL is based on the principle that increasing the amount of short-wavelength energy (i.e., blue) within the light improves visual acuity by causing the pupils to narrow, which enhances one’s efficiency at performing detailed visual tasks. Liebel explained that this can save considerable energy by allowing for equivalent visual efficiency at lower light levels. Acknowledging the widespread industry belief that warm-white light is generally preferred over cool-white light, he cited studies suggesting that in a work environment – which is primarily where SEL is intended to be used – the majority of people have no real preference either way. A new technical memorandum on SEL, TM-24, is expected soon from the Illuminating Engineering Society.
Our understanding of the issues surrounding the color of light is far from complete, which may be limiting the adoption of energy-efficient SSL for some applications. Lively discussions like the one we had in Long Beach put a spotlight on the issues (pun intended), helping to pave the way for a broad-based understanding.
