With many people interested in the science of investigation, I thought I would use some science to explain Color Rendering Index.
Many people have trouble grasping the CRI concept. Simply put, CRI explains how light affects the colors of objects we see and photograph. Take the image of a nice red apple. Certainly everyone knows what a red apple looks like, but exactly what color red is it? Many people will describe the color of the apple differently, and each person will undoubtedly disagree about its color.
Perhaps the light of a warm autumn sunset makes the apple look almost blood red, deep and rich in its color shade. A grey and overcast day may make the apple appear to be a lighter shade of red with a hint of blue shadow overtones in the crevices on the top and bottom. The warehouse worker pulling the apple out of his lunchbox under the mercury vapor warehouse lights may find the apple looks dull and somewhat colorless and unappetizing.
Much like the technical experiments of the TV show CSI: Crime Scene Investigation, we can use Color Rendering Index to properly ascertain exactly what color the apple appears to be under differing light sources.
CRI is derived from a series of visual experiments which aid us in determining how light affects perceived color. We can also refer to what is known as approximate color constancy, meaning that light reflected from a colored object varies considerably, depending on the spectral energy distribution of the light source. Casual observers are not ordinarily aware that there is quite a difference in the appearance of an object. In fact, many people think that most colors do not change at all, given our human tendency to “remember” colors rather than to closely analyze them, as the lighting changes.
The Color Rendering Index is an accurate color definition system which is derived from a series of experiments. It enables us to assess the impact of different light sources on perceived colors of objects or surfaces which reflect color.
Color is a very objective thing. There is no such thing as a “true” color. Color is a function of reflectance and absorption of the light waves under which the color is viewed. We each perceive colors differently. Research indicates that the translation of light waves into color perception is largely a function of our nerve connections; ie: how our brain is “wired”. This is why psychological factors come into play in our interpretation of color and color rendition.
To make things more interesting, color evaluation relies on comparison. Two red apples set apart and viewed separately may appear to be the same color red, but when set side by side and viewed together, subtle differences become apparent in hue, brightness and saturation. Not counting specific degrees of color blindness, normal human vision is then subject to various color interpretation.
So, what is the real color of the apple, you ask? This depends on the type of light illuminating it and our interpretation of those wavelengths of light being absorbed by and reflected off the surface!
In order to begin the investigation, we must first determine the color temperature of the illuminating light source. The color temperature of a light source is a number we give to measure the light’s color appearance. Generally we measure the color temperature of light on the Kelvin scale which is based on the apparent color changes of a no-color blackbody carbon radiator which changes color as it is heated.
Any object will emit light if it is heated to a high enough temperature. The emitted color of that light will then shift in a predictable manner as the temperature is increased. So, the blackbody radiator would shift gradually from dull cherry red to bright red to orange to yellow to yellowish white to white, and lastly, to bluish white; similar to the heating element on your stove, but far more intricate.
A light source’s color is the temperature which we will then measure in Kelvins, whereas the color of the blackbody radiator will match the color of the light source.
However, there are many modern light sources which are impossible to match exactly to the blackbody because they produce a discontinuous spectrum. These sources, such as mercury vapor street lamps, some HID lamps, and many fluorescent tubes are rated in what we call Correlated Color Temperature, close to a real measurable color but “no cigar” which furthers the difficulty of color rendition especially for the photographer.
With Kelvins for our light measurement in color temperature, we interpret the numbers simply as “warm” or “cool” light. This is about where most people prefer to leave the investigation up to the professionals! Many people find it difficult to understand that light sources with lower color temperatures, such as incandescent lights, are referred to as “warmer” sources of light, having more red content. Conversely, light sources with higher color temperature such as discharge lamps are “cooler” sources of light having more blue content. These descriptions of light source color temperature have very little to do with the temperature of the blackbody radiator. The descriptions actually refer to the way we perceive colors and color groupings. Herein lies what we call the psychological perception of colors.
It has been said that color, as perceived, has three attributes; hue, brightness, and saturation. All this of course then affects each of our perceptions of the true color of the red apple, psychologically.
The next step in the investigation requires another reference group, the actual color chip swatches used in the experiments. For this we use eight standard colors which have been carefully chosen by the international lighting community to measure CRI. These colors appear as pastel shades, which are quite susceptible to variations of reflectance, or color shifts, under different color temperatures of light. We can illuminate the eight color chip samples, first by the light source we wish to investigate, and then by a true blackbody matched to the same color temperature as our light source. If none of the samples appears to change in color, we can now give our tested light source a CRI rating of 100! If we see any changes of color in hue, chroma, intensity, saturation or purity on our samples, we will have to give our light source a lower rating.
Of course, the test requires very sophisticated and properly calibrated color testing and measurement equipment to accurately assign the proper CRI rating. Remember, if eight people viewed the eight color samples, we would most likely come up with eight different index numbers!
Our CRI index numbers will decrease as the average change in the color appearance of the color samples increases. A CRI rating of 80-100 is considered high and we can accurately assume that our source light has very good color properties, and an acceptable CRI.
So Now What?
We have determined that color temperature and color rendering index can provide us with some useful information, but what exactly can we do with it?
Well, CRI can help a photographer determine such things as what color type and amount of color correction gel to use. Lower CRI numbers between 60 and 80 tell us that color rendering is moderate and we may need to use some color correction. Ratings of 50 to 70 and below are considered very poor color rendering sources and have limited photographic capabilities, and are beyond correction. Such sources are however deemed acceptable lighting for applications like warehouses, parking garages, street lights, and other similar uses.
One can also use CRI to determine how set colors and costumes may appear under different light sources, and given the various CRI ratings, visual testing may be called for. The meat counter at the supermarket uses CRI to make sure the meat appears to exhibit the freshest, pinkest color possible!
Even though we have used very scientific technology to ascertain our CRI numbers, keep in mind that science is not perfect! Color temperature alone will not tell us how a given light source will actually render specific colors. Correlated color temperatures for light sources without a matching color temperature to a blackbody radiator produce results that are purely visual and open to interpretation. So for the photographer, correlated color temperatures can be very misleading, such as many fluorescent lamp sources with discontinuous spectra.
This is how it works…
Lets compare two similar “cool” light sources with similar color temperatures on the Kelvin scale. One of the light sources produces fairly uniform spectral energy distribution across the visible spectrum. The other light source, which looks the same to our eye, produces a similar visible spectrum of light but has a dip in the red wavelengths, and produces almost no visible red content.
The red apple appears the right color red and looks natural to the eye under the first light source with the even spectral distribution including red light. Under the second light source with the dip in the red frequency area, the apple now looks dull and colorless to the eye even though the two light sources have the same color temperature! Now that’s some weird science.
In conclusion, light sources with high CRI numbers will render colors well. However we must remember that CRI numbers are specific to light sources of any one given color temperature. Be careful to not compare similar light sources with different color temperatures even though they may have the same CRI numbers. A source which has a CRI number of 82 at 3200 degrees Kelvin, and another light source with a similar CRI number of 84 but a Kelvin number of 4200 will certainly give you very different color rendition of similar objects, even though the CRI numbers are very close. Keep in mind also that CRI is an average of the eight different color swatches.
A light source with a high CRI number will tend to render the eight colors all very well, but it is not a guarantee that any specific color will always appear natural under that source of light.
Now that the investigation is complete, we know the apple is the color red. However, it may not be the “correct” color you remember as the exact red you saw under your light source.
If we can use Kelvin and Color Rendering Index as useful tools and not exact science, chances are we will have more accurate and harmonious color reproduction in our day to day production scenarios.