In the last column, I talked about LED lighting and color temperature. I recounted the experience a small studio had when evaluating LED lighting to install in a new stage. They were somewhat surprised at the variance of color temperature and, more importantly, how colors look on camera. While their tests weren’t scientific, they pointed out the limitations of using old methods to measure new technology. In this case, I’m talking about measurement of CRI—or the color rendering index—of a light source.
CRI attempts to evaluate how well a light source will illuminate an object to present accurate color. Using a scale of 1 to 100, an illuminator with a score of 100 will provide good color rendering. The sun would have a score of 100.
This sounds simple and pretty black and white (for color, that is). But, there are a few issues with the system. First, CRI was developed to evaluate light sources on objects as seen by the human eye, not image sensors. While image sensors are good, they don’t work like the human eye. CRI also doesn’t take into account that there’s an entire process for creating images, from the image sensor to the display device. Second, CRI measures across eight color patches. These measurements are then averaged. Averaging means that if one patch is poorly represented, the overall measurement might hide the problem. Also, the color patches chosen aren’t that helpful to evaluate how light will affect some skin tones.
Since new light technology like LEDs can have gaps or spikes in their wavelength spectrum, using CRI doesn’t show where you might run into problems. You could have an LED that has a high CRI on the spec sheet, but when you use it for an interview, the flesh tone of the interview might be off. Manufacturers are aware of the issue of trying to create LED lighting that illuminates a broad color spectrum. One method is to test LEDs during manufacture and group them according to the color they produce. This technique, called "binning," allows manufacturers to create LED fixtures that minimize the spikes and gaps in the spectrum.
While manufacturers are striving to produce LED lighting that will yield better color fidelity through the camera and onto the monitor, it would help to have a better way of measuring their performance. A possible "better way" is through the Television Lighting Consistency Index (TLCI). The BBC first started work on a new method of measuring light sources in the 1970s. The European Broadcasting Union is moving forward with that work and has now developed an EBU Recommendation for TLCI. This recommendation doesn’t just look at the light source, but instead looks at the whole production chain, from camera to display.
To learn more about the EBU Recommendation, visit tech.ebu.ch/tlci-2012. The Science & Technology Council of the Academy of Motion Picture Arts and Sciences has a Solid State Lighting Project website that has information about their work, too, at www.oscars.org/science-technology/council/projects/ssl/index.html.