AI and the Internet Need to Learn Color Rendering

Recently a colleague of mine from the IES Color Committee asked of one of the new AIs “Tell me about color rendering.”  If we assume that the answer was formulated based on information about color rendering that is available on the internet we have a problem because the answer the AI returned was almost completely wrong.  That strongly suggests that most individuals and manufacturers posting information about color rendering don’t fully understand color rendering and are perpetuating outdated or incorrect information.  Here’s what the AI had to say, and what’s wrong with it’s response.

Color rendering is a measure of how well a light source reveals the true colors of objects. (1) A light source with a high CRI (Color Rendering Index) will show objects in their natural colors (2), while a light source with a low CRI will make objects look dull or washed out (3).  The CRI is a scale from 0 to 100 (4), with 100 being the perfect score (5).  A light source with a CRI of 100 will show objects in their true colors (1) while a light source with a CRI if 0 will make all objects look the same color (6).

The CRI (7) is important for applications where it is important to see objects in their true colors (1), such as art galleries, hospitals, and retail stores (8).  It is also important for applications where it is important to see subtle variations in color such as fashion design and photography (9).

There are a few things that can affect the CRI of a light source (10).  The type of light source (incandescent, fluorescent, LED, etc.) will affect the CRI (11).  The color temperature of the light source will also affect the CRI (12).  Light sources with a higher color temperature will have a higher CRI (12).

The CRI (7) is a valuable tool for choosing the right light source for your application.  By understanding the CRI, you can choose a light source that will show your objects in their true colors (1) and help you achiever the desired look and feel for your space.


  1. Objects don’t have “true” colors, they have apparent colors.  The color of an object is determined by the spectrum of light striking it, it’s reflective properties, and the resulting light directed toward the observer.  An object’s reflective properties don’t change, but the spectrum of light changes every time the light source changes.  Changing the spectrum may, therefore, change the color appearance of the object, hence apparent color.
  2. While there is a correlation between fidelity and naturalness, they are not the same thing.  CRI measures fidelity, i.e. how well a given light source matches the color rendering of a defined reference light source.
  3. One of the problems with single measure metrics like CRI is that there’s a lot of information that isn’t conveyed.  As CRI values drop, the only thing conveyed is that the match to the reference light source is worse.  A worse match, however, doesn’t mean colors are made dull. It could be they are increased in saturation since both deviations from the reference are equally penalized.  That’s the advantage of TM-30.  As Rf decreases we can see why by looking at Rg and some of the other 147 measures.
  4. CRI can have negative values.  TM-30 Rf is calculated so that 0 is the lowest value.
  5. 100 is the highest value.  It’s dangerous to call it “perfect” though as that implies that high fidelity is the only color rendering goal, which it isn’t.  TM-30 provides information for the color rendering goals of preference and vividness, and may include more in the future.
  6. A CRI of 0 will certainly make nearly all colors look terrible and very similar, but not all the same.
  7. CRI isn’t a proper noun, and shouldn’t be preceded by “the”.
  8. There are strong arguments for emphasizing preference over fidelity in many applications, including retail.  Again, fidelity isn’t the only color rendering goal, although it is the only one CRI measures.
  9. Research shows that high fidelity isn’t necessarily the best spectrum for detecting color difference.  Additional research is needed, but the IES may eventually add a color difference metric to TM-30.
  10. Only one thing affects CRI value – the spectrum of the light source.
  11. This is true because different light producing technologies have similar quirks in their spectra.  Those similarities can lead us to blanket statements such as “all fluorescents are green” which are not true for all products.  Again, the individual light source’s spectrum determines everything.
  12. A common misconception, but not true at all.  Not in the slightest. CCT and CRI are separate metrics.

Updating the CCT Calculation

As I noted in Chapter 9 of the 2nd edition of Designing with Light, we calculate color temperature, correlated color temperature, and distance from the Plankian locus in a perverse way.  The calculations are performed in the CIE 1960 (u, v) chromaticity diagram (which is why distance from the Plankian locus is Duv).  However, since 1960 (u, v) is obsolete, we perform the calculation using CIE 1976 (u’, v’) chromaticity diagram, but then scale the v’ axis by .66 so that we’re using 1976 (u’, ⅔ v’) which is 1960 (u, v).

To complicate things, to present information graphically, most manufacturers transpose these calculations to the 1931 (x, y) chromaticity diagram, resulting in the industry using 2 ½  chromaticity diagrams for various calculations and illustrations.  Unfortunately, they also use 1931 (x, y) to illustrate the gamut of multi-colored luminaires even though it isn’t uniform, making the illustration of questionable value (they should be using CIE 1976 (u’, v’), which is perceptually uniform).

In a counter to this fragmented system, yesterday Leukos published a research article called Improved Method for Evaluating and Specifying the Chromaticity of Light Sources.  Among other proposed improvements to how we perform chromaticity related calculations, it introduces a new uniform chromaticity scale (UCS) diagram with coordinates (s, t), a measure of correlated color temperature (CCTst), and a measure of distance from the Planckian locus (Dst).  Importantly, it makes all chromaticity calculations in a single chromaticity diagram instead of the 2 ½ diagrams we use today.  It’s heavy on the science, but is an important step in fixing our current system.

Night Shift? Not Quite.

We’ve all heard about the effect blue light can have on our circadian rhythms.  It can suppress the release of melatonin, which can delay sleep and reduce sleep quality with possible long-term health consequences.  Circadian disruption has been associated with depression and increased risk of diseases such as diabetes, obesity and cancer.

Back in 2016 the AMA released a report recommending night-time outdoor lighting have a color temperature no higher than 3000 K to limit night-time exposure to blue light.  That report was quickly criticized by the Lighting Research Center and the IES, among others, as I noted here and here.  One of the key criticisms was that correlated color temperature is a poor measure of spectral content and says nothing about the amount of energy in the wavelength range that affects our circadian rhythms.  A better measurement is melanopic content, which isn’t discussed in the report.

Last year Apple unveiled a feature in their OS and iOS called Night Shift.  When enabled it causes the color of the display to become warmer in the evening.  The assumption, the same as the AMA’s, is that warmer light has less blue so it won’t impact melatonin production.

A recently published paper in Lighting Research & Technology looked at the effectiveness of Night Shift.  This preliminary study suggests that “changing the spectral composition of self-luminous displays without changing their brightness settings may be insufficient for preventing impacts on melatonin suppression.”  Even when Night Shift was used, the devices still suppressed melatonin production.  The authors noted that, “it is not known how this amount of suppression induces circadian disruption, delays sleep or affects health.  Larger, more comprehensive epidemiological studies should investigate how the long-term use of these self-luminous displays affects people, especially adolescents and children.”

While additional studies are clearly needed, it provides additional evidence that lower CCT alone is probably not enough.  With our display devices we should also lower the brightness.

Design Guide for Color and Illumination

As the co-chair of the IES Color Committee I am delighted (pun intended) to announce the publication of the Design Guide for Color and Illumination.  The guide is the result of over five years of work by more than a dozen researchers, engineers, manufacturers, and designers from across the globe.  Here’s part of the description on the IES site.

Color can be described using concrete values such as chromaticity coordinates, spectral power distribution, or others discussed later in this guide. However, one’s response to color can be much more personal and emotional—and therefore more difficult to quantify. This guide takes the reader from basic vision and color vocabulary, through methods of measuring and quantifying color, and culminates in the practical use of commercially available white light and colored lights. The definitions, metrics, and references discussed will assist in building a critical understanding of the use and application of color in lighting.

It is probably the best, most thorough discussion of light and color available today.  Everyone interested in color, color perception, color rendering, and their relationship to light should read it.  It will be available at the IES booth at Lightfair.

LRC Responds to AMA on LEDs

You may remember that in June of last year the American Medical Association (AMA) released a report called “Human and Environmental Effects of Light Emitting Diode (LED) Community Lighting.” The report made some noise in the general press because it supported the idea that blue light from blue-pump white LEDs contribute to disability glare and retinal damage.

In the lighting community there was a considerable amount of frustration and anger over the report for several reasons. First, there were quite a few references cited that were either hearsay, such as a New York Times article about Brooklyn residents who didn’t like their new LED street lights, or were irrelevant, such as several articles about the effect of skyglow on nesting turtles. The other reason was that there was not a single lighting designer or researcher on the panel. Overall, it was a poorly researched paper that didn’t deserve the attention it received.

Shortly after it was issued, the Lighting Research Center at RPI issued a response paper. On March 15 the authors of that paper held a webinar to further address the AMA report. A video of that webinar is now available. If you’ve got an hour, take a look.

The key takeaways regarding the hazard of blue light from LEDs and the report are:

  1. The criteria of blue light hazard for retinal damage is much more than just color temperature, and includes the source size, intensity per unit area on the retina, and SPD of the light source.
  2. Disability glare is not a function of light source SPD, as the AMA paper suggests, although discomfort glare is. Short wavelengths increase discomfort glare.
  3. Color temperature is the wrong measurement to determine whether or not a light source will affect the circadian system and melatonin production because color temperature does not provide complete SPD information.       For example, some 3,000 K LEDs can have a greater impact than 4,000 K LEDs.
  4. The criteria of blue light hazard for circadian disruption from a light source include – the intensity, duration of exposure, timing of exposure, and SPD.

Who Needs A Lighting Designer? Museums and Galleries!

A few weeks ago I gave a three-hour seminar on lighting museums and galleries to the graduate students in an art curating program at a university here in New York. Condensing everything I’d like to say into less than three hours was tough. The two big questions were what to include and what to leave out. I started with a quick overview of how to think about light and lighting before moving on to basic vocabulary and some common lighting techniques. Then, since LEDs are clearly the future, even when lighting art, I moved on to an overview of both color temperature and color rendering. I talked about reference materials such as the IES Lighting Handbook, intensity and brightness ratios, and other considerations before we moved into their gallery space to use their track light system for some demonstrations.

After the whole affair a faculty member, who sat in on most of the seminar, said he had hoped I would have spent much more time talking about how to use track lights and less time on unimportant issues like design, color temperature, and color rendering (!). I was respectful, but stunned. Focusing track lights is so complex that it requires extensive demonstrations? Understanding that with LEDs the color qualities of the light vary widely, and can only be properly selected when they are understood is unimportant information? Uhh…NO. Or, as my 20 month old niece says, “no no no no.”

Yes, five or ten years ago the default light source in museums was an incandescent or halogen lamp. The color temperature difference was minor and the color rendering of both was excellent. That’s not true today. Look at the cut sheet for any museum grade track light and you’ll see that you have a choice of several color temperatures and CRI values. If ANYONE needs to understand the qualities of light that must be selected when using LED fixtures, if anyone needs to understand the affect that color temperature and CRI have on how colors are perceived, it’s certainly people involved in displaying and lighting art. To me, that means the curators of exhibits and the lighting designers they hire.

As I’ve discussed earlier, changing the color temperature of the light changes the color appearance of objects, as shown below.

Illuminated with Warm White Fluorescent Lamp
Illuminated with 3000 K light
Illuminated with Cool White Fluorescent Lamp
Illuminated with 4000 K light

The phenomenon of color consistency means that the shift in color appearance isn’t as great as one might expect or as these photos suggest, but the shifts are real. If you’ve ever bought a black garment only to discover later that it was actually dark blue you’ve experienced this shift. A similar thing happens when we compare a high CRI light source and a low CRI light source. If your work involves color perception this is basic and critical information.

Curators can be forgiven for not knowing much about this, but if they know nothing how can they collaborate with their lighting designer to show the art as they intend? Administrators and curators of museums and galleries – educate yourselves, then hire a lighting designer!

Circadian Lighting

A recent article in The Wall Street Journal discussed the possibilities and benefits of lighting systems that shift color to mimic the changes in daylight. It’s a complicated subject so it’s not surprising that some of what’s reported is inaccurate, so let me clarify a few things.

First, our current understanding of how light affects our circadian rhythms is that a light activates a type of cell in the retina called intrinsically photosensitive retinal ganglion cells (ipRGC). These cells, unlike rods and cones, are unrelated to vision. The signal they send to the brain is received by the suprachiasmatic nucleus (SCN), which is the body’s hormone regulator. As the graph below illustrates, the ipRGCs are most sensitive to short wavelength (blue) light and, unlike rods and cones, are unaffected by long wavelength (red) light. (A more detailed explanation can be found in Chapter 16 of Designing With Light, and in this article from the Journal of Circadian Rhythms.)

 

ipRGC (black), rod (blue), and cone (red) sensitivity curves
ipRGC (black), rod (blue), and cone (red) sensitivity curves

 

The ipRGCs signal to the SCN, which keeps our circadian rhythms are kept in sync with the day/night cycle, seems to be affected by three factors: color balance; intensity; timing. Cool (bluish) light should be delivered at a relatively high brightness in the morning hours. Delivering cool light in the evening hours can disrupt sleep and other processes controlled by the SCN.

Next, there’s nothing special about LED lighting that makes it uniquely appropriate for this application. Fluorescent, HID, and incandescent light (adjusted with color filters) can all be used to create a system that delivers cool light in the morning and warm light in the evening. In all cases, this involves the use of light sources of several tints (warm, cool, and possibly neutral) that are individually controlled. The WSJ is completely wrong when it says “unlike incandescent or fluorescent lights, LED lights’ materials and electronic components allow for finer adjustments of color, brightness and intensity.” A single LED creates light of a single color. To shift colors a second LED is required. Brightness and intensity are the same thing. With LEDs set to surpass sales of all other light sources within the next several years, research involving sources other than LED is nearly non-existent. Naturally, researchers are using LEDs to test theories and to develop demonstrations and products.

Finally, it would be helpful if the WSJ had defined two terms – color temperature and CRI (Color Rendering Index). Both are explained in Chapter 8. Color rendering and CRI are briefly explained in this post. CRI and color temperature are both addressed in this post.

Basking in a New Glow

The New York times has an “I Heart LEDs” article in today’s paper that leaves out some important information about evaluating them.  Here are some additional thoughts.

The government hasn’t done a very good job of publicizing or explaining that the Energy Independence and Security Act of 2007 (EISA) set minimum efficiency requirements for general use light bulbs (the act excluded decorative and colored products).  The incandescent lamp that’s been around for over 100 years doesn’t meet the energy efficiency standard.  Rather than re-engineer incandescent lamps, the lamp manufacturers have focused on expanding and emphasizing compact fluorescent (CFL) and light emitting diode (LED) technologies.  Again, you can still purchase 40 – 100 watt decorative incandescent lamps but not A-lamps, the most common shape in use.

The easiest substitution, one that requires no thinking about rewiring, dimming, etc., is the halogen lamp.  Halogen lamps are an improvement on standard incandescent lamps, and many of them meet the EISA energy efficiency requirements.

If you’re looking for higher energy efficiency, and are willing to pay a higher price up front to get it, CFL and LED lamps are available in a wide range of wattages and shapes.  However, they  need to be approached with caution.  Both technologies can be difficult to dim, especially with older dimmers that were designed with incandescent lamps in mind, so your existing dimmers may need to be replaced.  They can also produce unsatisfactory tints of white light.  LEDs are especially notorious for not matching the information provided on the packaging, as demonstrated through the Department of Energy’s CALiPER program.

Here’s what to look for.  Every light bulb package should have a Lighting Facts Label that looks like this.

Lighting Facts Label
Lighting Facts Label

The orange/yellow/white/blue color bar is where you’ll find information about the warmth or coolness of the light, both with an arrow on the color bar and with a number.  The number is called the Color Temperature (actually the correlated color temperature) and measures the warmth or coolness in Kelvin.  The important thing to know is that a lower number (2700 to 3000 K) is roughly equal to an incandescent light bulb.  As the number gets higher the light gets cooler.

Warmth/coolness isn’t the only measurement of the quality of light.  Another consideration is how well the light source allows us to see the colors of objects.  This is called Color Rendering (Color Accuracy on the Lighting Facts Label) and is indicated by a Color Rendering Index number.  Higher numbers (with a maximum of 100) indicate better color rendering, so a light with a Color Accuracy of 95 should be visibly better than one of 80.

The Color Rendering Index is not very specific, however, and is known to misrepresent LEDs.  Therefore you are the best, final test of whether or not a given light bulb is appropriate.  I recommend purchasing only one or two and trying them out for a few days before committing to changing over your entire house.

My other recommendation is to stick with the major manufacturers (GE, Philips, Sylvania) for most lamps that you test.  These companies have a track record of product consistency and quality that many of the newer manufacturers don’t.  I can almost guarantee that with an off-brand 5-pack of lamps for $10 you’ll get what you pay for and hate the results.  It’s not the technology that you’ll hate, but the manufacturer’s poor execution of the technology.

I hope this helps.