Our conclusion, briefly
Lighting has one universal light quality metric: the Color Rendering Index, or CRI. Pronounced "C R I" (occasionally pronounced "cree", although that causes confusion with Cree, the lighting company), this light quality measure inadequately distinguishes between good and bad lighting, especially where LED products are concerned. Traditional lighting manufacturers appreciate its predictability (or less charitably, its inaccuracy), so it has remained virtually unchanged for 50 years. LEDs are forcing an update, however, and a new standard currently called TM-30-15 is being considered (TM stands for "testing method;" 30-15 is just a filing metric with no particular meaning related to the test). It broadens the palette of test colors and adds a dimension to quality measurement. We think it's better.
How CRI works
Currently there’s only way for the lighting industry to communicate how “good” light looks: a measure called the Color Rendering Index (CRI). The CRI has been the reigning light quality standard since 1965. Every lighting product or bulb undergoes a CRI test and is given a score between 1 and 100, with 100 being the light quality of a reference incandescent light bulb.
The test light and reference light are each shined at eight specific colors, and a score of 1-100 is given on each color. The average of the scores delivers the final CRI.
Other than slight tweaking to measurement methods in 1974, CRI has been unchanged since inception. This past summer, however, lighting’s governing body (the CIE - there isn't much to justify a visit to their website, but just the same: www.cie.co.at) announced progress on a new method for determining how nice light looks. The challenge in the past has been finding common ground between scientists and the lighting industry to move to a new standard, but the growth in LED lighting has driven a need for change.
The new method, currently referred to as “TM-30-15” does indeed represent change. First a look at where CRI comes from, and then on to why this new standard is interesting.
The origins of CRI, and how the original intent undermines its usefulness today
CRI is a reference measure, so where most measures are absolute (weight, distance, brightness, color temperature), CRI is a comparison of one light to a reference light. The reference light has always been incandescent, meaning CRI was created in response to the advent of non-incandescent light sources, specifically fluorescent and high-intensity-discharge lighting. These "new" light sources were, and are, agreed to be worse than incandescent lighting in quality terms, so the CRI existed to prevent lighting products from falling below an acceptability threshold. CRI scores are out of 100, where 100 is the rendering delivered by incandescent lighting - no fluorescent light achieves a score above 90.
In 1965 it made sense to devise a test to set boundaries around what is considered minimally viable lighting, but when looked at through the lens of today's lighting product set, the CRI leaves room for serious questions. Why only 8 color samples? Why are they all pastel? Where’s red? Why has this test not been changed? Manufacturers have typically stood in the way of color quality overhaul because the pioneering technologies of the 60s and 70s - fluorescent, metal halide, and mercury vapor lighting - had seen vast adoption in lighting applications around the world. Overhaul of the CRI stalled because a true quality metric would have damaged the perception of these gas-based lighting technologies.
With LEDs, we now have TM-30-15 (maybe)
LEDs changed everything in the quality measurement game. With superior light quality now available in an energy efficient product, lighting manufacturers have come around to what lighting scientists have wanted for 50 years - a metric that measures for the best possible light.
Standards change slowly in large industries, so TM-30-15 is just an iteration, but it has been agreed upon by various market participants - scientists, specifiers, designers, and manufacturers. The new standard proposes two major changes:
Raise the number of color samples from 8 to 99. The leap there acknowledges how the current system is not anywhere close to adequate.
Adds “gamut” to the testing criteria. Before, there was only “fidelity” which refers to similarity between the color’s appearance under the test light versus the reference light. Gamut, now, is in plain terms the richness or intensity of the color.
So, where CRI has for seventy years existed only to stack-rank different fluorescent and HID bulbs, the new TM will actually go some distance to measuring light quality.
It’s good news for LED, no news for incandescent, and bad news for fluorescent
The shortcomings of CRI have been highlighted in recent years when improvements in LED light quality have not correlated with higher CRI scores. The two below images are both 80 CRI scores, for example, but the fluorescent on the left fails to render much of the color in the scene that would be visible under an LED.
Where currently both good light and bad light are able to score well in CRI, bad light will start to actually score badly. This will not affect incandescent lighting, which remains the reference light and faces myriad regulatory challenges of its own, but will strongly skew the marketplace toward LED lighting over fluorescents. Whatever the relative costs and benefits of LEDs, the idea that they are equivalent to fluorescent bulbs from a quality perspective will disappear.
New standards, new ratings, and the eye of the beholder
In our work specifying, implementing, and managing lighting, we do not focus on CRI. As discussed here, we find that on its own CRI is unable to indicate light quality. Unless the client has a pre-existing CRI minimum requirement (and some, especially in the hospitality industry, do), we do not make it part of our process. TM-30-15 offers a much clearer future for lighting product rating and specification, and if it is passed we'll take it seriously.
And yet, people to know what they like. Trite as that may sound, we believe users untrained in the science of lighting measurement are capable of walking into a room and determining for themselves whether the light looks nice. Rating systems are useful for doing a high-level comparison across many products (extremely helpful in the diverse LED business), but we will continue to rely on implementing lighting systems the look beautiful to the people who use them.