There are various methods of measuring and specifying the color rendition capability of a light source; a commonly used metric was created by CIE in 1965, called the Color Rendering Index (CRI). The CRI measures a light source’s ability to reveal the intrinsic colors of the objects it illuminates, referred to as its accuracy, or color fidelity. CRI is based on the average fidelity of how a light source renders eight color samples.
CRI is measured on a scale of 100 with values at higher end of the scale indicating more accurate color rendering (as shown in the graphic below).
The Color Rendering Index (CRI) scale, where higher values indicate more accurate color appearance.
For example, the strawberry picture below illustrates the difference between color appearance at CRI of 95 (very accurate), compared to a CRI of 70, where colors appear washed out and not as “true to life.”
Comparison of color rendering with different CRI values. With poor color rendering (lower CRI), reds typically appear more “orange” and other colors appear generally washed out or dull. Light sources with good color rendering (high CRI) typically show vibrant colors and rich “cherry reds.”
To determine a light source’s CRI, it is tested using eight color chips (numbered R1 to R8), compared against a reference source with the same Correlated Color Temperature (CCT). Typically, test values are measured by a spectrometer and calculated using the light source’s spectral power distribution (SPD).
In certain lighting applications a light source with a low CRI value can be preferable. LEDs with high CRI values typically contain more red content in their SPD, providing more accurate color rendition. Red light requires more energy consumption per lumen due to low eye response in that wavelength region. Thus, high-CRI LEDs operate at a lower efficiency—they produce fewer lumens (a measure of brightness) per Watt (LPW). In environments where accurate color rendition is less important than energy efficiency and reducing cost, LEDs with a lower CRI value may be the best choice.
The CRI attempts to measure how accurately a light source can reveal color. The CRI gauges each lighting source based on the average of eight defined Test Color Samples (TCS), labeled in the image below. It compares the appearance of the samples under a real light source against how they would look under a reference illuminant.1
This color match is rated on a scale with a maximum of 100, with 100 meaning there is no difference in color rendition of any of the TCS samples between the measured light source and the reference illuminant.2 Light sources with a CRI score of 50+ are considered usable for most general applications. Top-grade LEDs, such as those from Luminus, boast CRIs with minimums of 80, 90, and 95 CRI.
Each result is expressed as an Rn value labeled R1 to R15. For example, R9 is a deep red—a critical reference color for many applications such as lighting in butcher shops, grocery stores, and medical settings. The CRI value of a light source is reported as the average value, Ra, of samples R1 – R8, which are the original eight TCS samples; the additional seven, R9 – R15 were added later.
While useful for some applications, there are many criticisms of using CRI for all color rendering determinations. Many have considered the eight CRI reference color samples to be limited, inaccurate, and too pastel to represent the vividness of the real world. Consequently, the seven additional reference colors were added, including the saturated red, R9. In addition, two skin tones and a vivid green make up the supplemental samples. However, only the original eight samples factor into CRI Ra calculations. Thus, with the additional reference colors, under sampling can still lead to inaccuracy.
With CRI, scoring also tends to be inconsistent between devices. Two light sources with the same CIE chromaticity coordinates (x,y) can have very different SPDs (referred to as metameric light sources). Similarly, two sources with the same CRI can render light differently because of the differences in their spectral power distribution (SPD). Gaps in the mathematical models and issues such as under sampling (with just eight CRI reference colors) mean that light sources with similar Ra values can nevertheless produce very different color rendering effects, as demonstrated in the image below.
Three LED bulbs with the same CCT (3500K) and an identical CRI (90) nevertheless render colors differently. The white shirt and the red scarf on the right and left appear visibly brighter; the shirt in the middle has a slightly yellow cast and the red scarf is a cooler tone (less orange, more blue).
This image demonstrates the differences between three LED sources. They have the same CCT value (3500K) and excellent (90+) CRI values, yet there is a visible difference to the human eye. All three are Luminus LED products, yet each is produced for a different use case.
Current research shows that the vibrancy of red tones is a determining factor in how humans perceive the quality of light sources. Thus, in many lighting applications such as retail environments, increasing the saturation of red wavelengths in a light is desirable. Yet the CRI, which is based on “outdated models of vision and color that underlie the calculation,”3 does not account for this perceptual preference. In fact, increasing the vibrancy of red in a light source is “penalized” in the CRI metric.
This issue—and the variation evidenced in the above image—both point to the need for more precise methods to quantify color rendition of light sources. In recent decades, new methods include the Gamut Area Index (GAI), Color Quality Scale (CQS), and TM-30.
To learn more about these various color rendering models refer to the Help Center articles:
NOTES & REFERENCES
- The reference illuminants, defined by standards committees, are full-spectrum SPDs that match the CCT of the artificial light source being measured. CCTs <5000K are compared to a Planckian black body and those >5000K are compared to the D-series, a mathematical representation of an ideal light source that resembles sunlight. In some color rendering models the black body SPD and the D-series SPD are blended in the CCT region between 4000K and 5000K. Reference illuminants are explained further in the Help Center article What is a Standard Illuminant (Reference Illuminant)?
- It is possible for light sources to have negative CRI values. For example, low-pressure sodium lights can have a CRI of -47. They render all colors in a muddy yellowish tone. Mathematically, the CRI (the Ra) is linearly scaled to indicate the amount of chromatic variation from the color of the object under the reference illuminant. The reference illuminance has a score of 100, and the distance from 100 indicates the accuracy of other light sources. Thus, a CRI value of 100 means no variation (perfect rendition); an Ra value of 60 means that there is four times the average color shift than is the case with an Ra value of 90.
- “Research Compares Measures of Average Color Fidelity.” Department of Energy, Office of Energy Efficiency & Renewable Energy, January 31, 2018. (Accessed October 25, 2021)
Related Articles About CRI-Centric Applications
Luminus Website https://www.luminus.com/
Luminus Product Information (datasheets): https://www.luminus.com/products
Luminus Design Support (ray files, calculators, ecosystem items: [power supplies, lenses, heatsinks]): https://www.luminus.com/resources
Luminus Product Information sorted by Applications: https://www.luminus.com/applications
Where to buy Samples of Luminus LEDs: https://www.luminus.com/contact/wheretobuy.
Technical Support Contact Information: email@example.com
Sales Support Contact Information: firstname.lastname@example.org
Customer Service Support Contact Information: email@example.com