THE PHYSICS OF TUNABLE WHITE

Color Space

This image is what’s known as a color space. It is a representation of all possible visible colors of light. This particular space is the x,y space. Other color spaces serve different purposes. Each point within the color space has a set of coordinates describing its location within the space, known as its chromaticity coordinates. These coordinates identify the exact color of light and result from a spectrophotometer, among other tools.

Image courtesy of Lutron Electronics Co., Inc.

Color space represents all possible colors of light.

The boundary of color space is defined by monochromatic light, a single wavelength of light, and everything inside from the edge is a mix of wavelengths. The axis, x, and y, are unitless and represent relative amounts of light: x being the relative amount of red light, and y being the relative amount of green light.

As discussed earlier, the curve highlighted in the middle is the black body curve or black body locus. It represents all of the possible white lights along the range of color temperatures. Along this curve are the commonly used CCTs.

MacAdam Ellipse

Many light source spec sheets refer to color difference or SDCM (Standard Deviations of Color Matching) or MacAdam ellipses. The SDCM indicates the likelihood that someone would notice a difference between a light source with a given set of x,y values and a comparison light source with another set of x,y values.

This color-matching deviation is an important concept when specifying a project’s lighting. Sometimes, a manufacturer may say their LEDs are within a two-step MacAdam ellipse. This means that 95% (two standard deviations) would notice a difference between the center of the ellipse and the boundary of the ellipse between those two points. In this drawing, the ellipses are scaled up to be able to see them. They are 10 times the size they would normally be for a 1-step MacAdam ellipse.

Image courtesy of Lutron Electronics Co., Inc.

The MacAdam ellipse indicates deviations in light color.

Color Rendering Index (CRI)

CRI represents how well the test source renders color compared to a reference source of the same color temperature. The reference used is the black body curve, which means that the baseline source is an incandescent lamp up to 5000K and daylight above that value. As such, most incandescent sources score nearly 100 on a scale up to 100. CRI was created in the mid-1900s by the International Commission on Illumination, abbreviated CIE due to its French name, the “Commission Internationale de l’Eclairage.”

Image courtesy of Lutron Electronics Co., Inc.

The CRI compares a test source to a reference source of the same color temperature.

The CRI value arises from comparing the test source to the reference source for eight pre-defined color samples. The small sample stems from the manual calculation used at the time. One of the issues with CRI is that these eight color samples, R1-R8, do not correctly represent commonly found colors. The eight samples used are pastel colors and do not do a great job representing the color red, which is very widely seen in the world and critical for rendering skin tones. In addition, incandescent sources are very good at rendering the color red, so it is expected that light sources will be able to do this accurately even though it isn’t reflected in their CRI value. The standard defines other color samples, R9-R14, but they are not used in calculating the CRI value. Because of these challenges, CRI does not guarantee overall color rendering quality but how closely they match the colors compared to the reference source.

These specific color samples have enabled some LED companies to focus on the wavelengths that will give them high CRI values and ignore other wavelengths, leading to sources that produce a good CRI but do not render other colors well.

To find these values, no physical test with color samples is ever performed. A Spectral Power Distribution, or SPD, for the source is created using lab equipment. The R-value for each sample is calculated using the values from the test source’s SPD and the known values of the reference source. Each R-value represents the variance between each color sample’s test and reference sources. The values for R1-R8 are then averaged together to determine the “CRI value.”

Color Rendering: TM-30-18

TM-30 is a color metric that the Illuminating Engineering Society (IES) developed. It addresses some of the shortcomings of CRI and looks to expand how color rendering is discussed to relate better to what a source is doing. It uses 99 color samples for its comparison that better represent commonly occurring colors to strengthen the reliability that the metric truly represents the source’s rendering ability within the built environment. This metric also conveys other information, such as saturation (or how vivid a color it can render). Most importantly, red is a primary factor in this calculation.

Image courtesy of Lutron Electronics Co., Inc.

TM 30-18 color metric graphic; source IES.

Color Rendering: TM-30-18, Annex E

TM-30 has many components to it in the calculation. A wide variety of metrics represent how a light source performs. Annex E is designed to simplify the metric and provides recommendations for target levels of preference, fidelity, and vividness. Each of those characteristics has three levels. The metrics output from TM-30 determines which level a light source falls into. The visual is attempting to represent that some targets are mutually exclusive. For instance, you cannot have a priority 1 level for vividness while maintaining fidelity.

Image courtesy of Lutron Electronics Co., Inc.; Source: IES

Annex E helps simplify lighting recommendations for target levels of preference, fidelity, and vividness.

Color Shift/Stability

While color consistency refers to sources being close to each other after manufacturing, color shift describes variations that occur to the LEDs over time. The phosphors used in LEDs can deteriorate over time, changing the LEDs’ visible color. This deterioration is usually due to heat, which means that fixture design can help mitigate these effects by helping to minimize that thermal impact. LED fixtures shifting colors over time can make an installation look poor and limit the ability to replace a few fixtures in a space or add fixtures later, as there is the potential that they won’t match the other, older fixtures already installed. Some LEDs have fixed this problem by adding calibration of the source over time.

Low End

The low-end performance of a fixture has always been a vital spec to determine its suitability for different applications. That becomes even more important with LEDs since they often output more lumens than the sources they are replacing, and they are particularly prone to issues at the low end. These issues include pop-on, pop-off, flicker, strobing, and other undesirable effects, often leading to their relatively high low-end capability.

For the sake of this conversation, low end refers to the lowest, stable light output that a source can generate. Low end is measured in a percentage, representing the percent of the high-end light output. Low end is critical because of how eyes perceive light. As the measured amount of light decreases in a space, our pupils dilate and absorb more light, which leads to plateauing of the perceived amount of light in a space.

It is not uncommon for LED sources to have a low end of 20 percent measured light output, which means that a light meter on the table would show an 80 percent decrease in light between high end and low end. Unfortunately, this only comes out to a 45 percent perceived light output, meaning that the lights would appear to still be at half of their original output. This perception makes it critical to find good-performing LEDs that can get to low ends and allow the person in the space to perceive the lights dimming to a reasonable level.