Researchers KONSTANTINOS PAPAMICHAEL, MICHAEL SIMINOVITCH, JENNIFER A. VEITCH, and LORNE WHITEHEAD argue that CRI is the proper metric for use today as a complement to efficacy in SSL regulatory policy.
The four undersigned scientists write with deep concern regarding the November 2016 article by our respected colleagues, Mark Rea and Jean Paul Freyssinier, entitled "CRI should never be used in efficacy regulations but a new lumen definition should." We share their interest in having a suitable regulatory system for new lighting products, including LED-based products, to ensure that energy efficiency and human value in lighting are suitably balanced. We also believe that lighting regulations should rest on a firm scientific foundation and will explain our concerns here regarding the article by Rea and Freyssinier.
We work at three different research organizations and we come from different research backgrounds. Our long experience in illumination and vision research qualifies us to provide an expert opinion and we feel a professional obligation to do so. Our concerns are (1) that the Rea/Freyssinier article is misleading with regard to the adoption of metrics in regulation, and (2) that its key conclusions about these metrics are scientifically incorrect.
The authors of the aforementioned article present their case as though regulators have the responsibility or authority to define fundamental metrics. Regulations ought to be based on consensus-based metrics that have undergone a thorough examination and analysis by international experts, and that have been accepted by the relevant international body. In the case of the lumen, that is the Bureau Internationale de Poids et Mesures (bipm.org). There may come a day when the universal lumen has achieved that status, but until then it would be premature for regulators to base their requirements on it. We believe that if the opinions offered by Rea and Freyssinier were to be presented for scrutiny in a science consensus process they would be rejected, but the relevant point here is that the place for this debate is first, in the peer-reviewed literature, and second, in the committees where standardization decisions are undertaken.
As concerns the scientific argument put forward by the authors, we are puzzled by inconsistencies and find the argument illogical. Rea and Freyssinier seem to argue both that color fidelity is important to light source quality (this is the premise with which the article opens) and is not important (they argue that by using a suitable universal lumen criterion, any light source, even high-pressure sodium [HPS], could be accepted regardless of its CRI [Ra] value).
The basis for the seeming contradiction seems to be the authors' rejection of the CRI as an effective or useful metric. Most scientists today agree that the accuracy of the CRI as a measure of color fidelity is in need of modest improvement, and, separately, that there is a need for one or more additional metrics to fully characterize the color rendering characteristics of light sources. This is why both the Illuminating Engineering Society (IES) and the International Commission on Illumination (CIE) are working on a more accurate metric for color fidelity and also on important perception issues beyond color fidelity.
Regulators are well aware that available metrics often improve over time. Nonetheless, at any given time, regulations must be based on what is available. Therefore, when a regulator today decides to address color fidelity, it uses the CRI metric because it is the currently accepted means to quantify this important dimension of light source performance. Until there is international agreement on improvements, the CRI metric is the appropriate way for regulators to specify color fidelity requirements.
We also find it puzzling that the authors have tried to support their argument by citing observations of participants' preference judgments about object color appearances under various light sources. These preference judgments are not appropriate data for light source regulations, in part because the preferences are not a function of the light sources alone. What is reported in these experiments is a preference for object color appearance based on the interaction between the light source and the illuminated objects. This is seen clearly in Fig. 1 of the article, where the light sources rank differently for general and skin tone preferences.
Other scientific literature also shows that these preferences differ for different contexts, and that there are wide individual differences among people. This topic is a popular one among lighting and vision researchers today. As our understanding of these preferences improves, we can expect that context-specific lighting recommendations might evolve to support achieving preferred color appearances using specialized light source spectra. Nonetheless, this information is likely to remain unsuitable as a basis for general light source product-performance regulations.
The authors also argue that regulators seeking to ensure that efficiency regulations also provide minimum light quality ought to use a new definition of the lumen based on Rea's universal lumen concept, arguing that it would avoid the need for a separate indicator of color fidelity. This position also puzzles us, because it obviously does not ensure that all products will deliver a suitable level of color fidelity. That the universal lumen is a poor indicator of color fidelity is evident from the table in the article, in which the light sources with both the best and the worst CRI value have the same value of U(λ). If color fidelity is an important quality consideration for light sources, then the metric used in light source regulations ought to differentiate between light sources in terms of color fidelity.
We see no alternative to using multiple metrics to describe light source performance, because different desirable characteristics compete against one other - that is, they cannot be simultaneously maximized. Instead, for any specific use, we can at best seek optimization of value per watt based upon the thoughtful relative prioritization of factors including (but not limited to) luminous efficacy, chromaticity, and color fidelity. We look forward to working with the authors and other international experts in light and lighting to further the development of suitable consensus-based, scientifically grounded metrics upon which regulators can base impartial rules to the betterment of society, the environment, and trade.
Commentary from the editors of LEDs Magazine
This is one of several recent articles that we have published on the topic of regulatory policy that guides the adoption of LED-based lighting and industry standards for metrics that are referenced in regulations. This contribution was prompted by an article in our November/December issue. Readers may also find some of the other recent content compelling to review and maybe it is needed to fully appreciate the views expressed in this article.
Indeed, three of the current authors previously contributed an article supporting the regulatory policy developed by the California Energy Commission (CEC) that essentially requires 90-CRI performance in consumer lighting products. The CEC action has been controversial of late and we at LEDs Magazine have questioned the theory behind the regulations. We first detailed the new regulatory requirements back in February 2016.
Surely the debate will continue. There are qualified scientists on both sides of the issue. But we'll maintain our stance that consumer choice should take the place of a requisite color fidelity regulation.
KONSTANTINOS PAPAMICHAEL and MICHAEL SIMINOVITCH represent the University of California at Davis and the California Lighting Technology Center; JENNIFER A. VEITCH, the National Research Council of Canada; and LORNE WHITEHEAD, the University of British Columbia.