Since the 1960s, there has been an internationally agreed-upon way to describe the color rendering of a light source using a simple numeric value. It's called the Color Rendering Index, or CRI. It's widely used and you will normally see it listed in lamp catalogs, specification sheets and on lamp cartons. Typical values of CRI for CFL and LED bulbs range from 80 to 90 and, more recently, CRI ratings over 90 have been appearing for LED products. Because the CRI attempts to rate the appearance of colors lighted by a specific light source, it must take into account the variety of colors used in interiors, from walls to furnishings and objects to complexion colors of people. The idea is to show, numerically, how people and things lighted with a high-CRI light source should appear more natural and normal than the same people and things lighted with a lower-CRI light source. And that means, for residential spaces, the higher the CRI, the better. That's because the reference or "ideal" light source for visually warm residential spaces has been the incandescent lamp, which has a CRI of 100 by definition.
A CRI rating of 100 is as good as it gets, this is because the CRI value represents a mathematical comparison of the measured color characteristics of a light source illuminating a defined set of color samples using a reference or "ideal" light source compared to the same measurements for the light source being evaluated. In other words, if all of the sample colors look the same under both light sources, that's a 100% match and a CRI of 100. But, there's an important caveat - the light sources being compared have to have the same or similar chromaticities for the CRI value to be meaningful. The chromaticity of a light source is a metric that describes the tint or tone of that source. For lamps typically sold to consumers and used in residential lighting, chromaticity is expressed in Kelvins or "K". Incandescent lamps, for example, have chromaticities in the range of 2700 - 3000 K. On the DOE "lighting factsCM " labels, which now appear on bulb cartons, that range is described as "Warm White". Some incandescent lamps, such as the halogen-incandescent MR16 reflector types appear "whiter" because their chromaticities are designed to be at or slightly over 3000 K.
However, there are problems with applying CRI values to LEDs. As LEDs began to be used for general lighting starting in about 2005, people noticed that what they saw with their eyes did not correspond with the CRI ratings. In some cases, colors illuminated with LED lighting looked better than the CRI values indicated; in other situations, the colors looked worse.
The International Commission on Illumination (CIE), which wrote the CRI standard, has been aware of the problems and has now been working for several years to find an agreed-upon fix, with no success. In the interim, several new color rendering mathematical models have been proposed, including a metric called the Color Quality Scale or CQS, which was developed by the National Institute of Standards and Technology (NIST) in 2006. CQS addresses many of the problems with CRI, and several LED lighting manufacturers now report CQS ratings for their products, but CQS has not been adopted by the CIE, thus, the controversy continues.
Last year, the Illuminating Engineering Society (IES) formed a Color Metric Test Group to develop an improved or complimentary color rendering index appropriate for LEDs. The IES reports that this work is in process; however, this past August, the IES said that the CRI "should not be used in energy regulations to characterize color attributes for solid state lighting until there is industry consensus on the issue." The full IES statement can be viewed here.
While the DOE and EPA/Energy Star have not yet commented on the IES statement, the California Energy Commission (CEC) heard testimony on September 29 proposing that the CEC ignore the IES statement for new Title 20 rules involving the regulation of small-diameter reflector lamps such as the LED versions of MR11 and MR16 products.
The DOE, however, already recognized the problems early in 2008 via a technical bulletin "Color Rendering Index and LEDs," which you can find here. It is brief and contains some helpful background information.
So, considering that resolving the CRI issues and implementing a new metric have now already taken several years, and will likely take several more, should we continue to use the CRI just as we have been doing or is it time to make some changes?
I recommend continuing to use the CRI, but also taking into account its limits, especially when dealing with LED lighting products. Here are a few practical suggestions that I hope you will find useful:
-Consider the CRI an indicator rather than a measurement. At its best, CRI is an out-of-date, limited-use metric that tries to turn a complicated visual experience into a simple number. It's useful, but not as precise as the CRI rating numbers suggest.
-Differences in CRI values of 3-5 cannot usually be seen by normal eyes.
-Currently, CRI values are calculated based upon a set of 8 or 15 color samples. See those color samples here: http://alturl.com/e6w64 (top row). Some lighting specifications call for the so-called "R9" value to be reported separately. That's the color sample shown as TSC9 and sometimes described as "strong red."
-The color characteristics of LEDs are such that they can trick the CRI into reporting values that are misleading.
-The color rendering values of an LED lighting product reported as CQS, rather than CRI, are more reliable. Both use the same 0-100 scale where higher is better. Note the color samples used for CQS (bottom row) are more saturated.
-Expect to see some new and better metrics for evaluating the color rendering of light sources. One is called Gamut Area Index (GAI), but I'll leave that subject for another column.
-Human color vision is subjective, so the most reliable color rendering information is going to arrive via your eye. When other people are involved, say in a lighting showroom where products are being compared, nothing works better than an actual side-by-side color comparison in a full-scale demonstration.
Check LED lamp and product manufacturers literature for interesting and clever ways to explain and describe the color characteristics of lighting products. There are some useful ideas out there.
I recently gave a presentation on this subject at an LED lighting conference. That presentation can be downloaded from the ALA web site here.
Your comments and questions are always welcome.
