Fluorescents and the Kelvin Myth: Changing the Calibration for Color
By Joseph N. Tawil
Measuring color temperature in Kelvin (once degrees Kelvin) is a system of designating a light source's spectral distribution. The basis of this measuring system is the "Theoretical Black Body Radiator" as shown in the "Black Body Locus" on the CIE Chromaticity Chart (see illustration A).
To understand use of the Kelvin scale, one must first define what a "Black Body Radiator" does. At absolute zero - minus 273.3 ° Centigrade - this theoretical object radiates no energy. As the temperature of "Theoretical Black Body" is raised, it begins to emit energy. At about 700° C, a faint red glow becomes visible to the naked eye. As the TBB's temperature is raised to about 1517° C, it glows with a light similar to candle or firelight, 1800° K. (We started at minus 273° C and raised the temperature of the Theoretical Black Body to 1517° C - the result 1800° K). Continuing to raise the temperature will cause the Kelvin factor to rise: first to 2850K, a typical household incandescent lamp; then to 3200K, a typical studio incandescent lamp; and finally on to 5600K, nominal daylight. The Kelvin scale continues to rise to a blue sky of 28,000K (see illustration B).
The spectral distribution of the TBB is continuous as it varies from 1800K (candlelight) to 28,000K (north sky). That is the spectral energy is not "interrupted" in the visible range from 300 to 700nm (see illustration C). However, the balance or ratio of red to blue is shifting - the higher the Kelvin, the "bluer" the light.
When describing a light source as having a Kelvin rating, it refers to a light source that emits energy across the entire visible range from 300 to 700nm. Sunlight and incandescent lamps are very good "Black Body Simulators"- they behave in a manner very close to the predictable spectral distribution of the "Theoretical Black Body Radiator." With these light sources, one need only measure the energy at two places in the spectrum - red and blue - to determine the Kelvin. However, when the light source has an "interrupted" spectrum - as do all fluorescent and discharge lamps - it is not correct to describe that light source as having a Kelvin temperature.
Let's examine the spectral distribution curve of a typical fluorescent lamp (see illustration D). Note the high peaks of energy at certain wavelengths and the areas of "interrupted spectrum." These published charts can be misleading because it is acceptable to shorten the high peaks of energy and widen the bar to represent the total energy (it saves paper). In reality, these bars represent high peaks of energy at certain points in the spectrum, and the smoothed out continuous line is artistic license with areas of little or no energy. The Kelvin scale should not be used to describe light sources with an "interrupted spectrum." Sadly, the term is often misused, and by those who should know better: the very lamp manufacturers being depended on for this important information.
Once upon a time, the term "Apparent Color Temperature" was used to describe light sources that "looked like" but did not behave exactly like a certain Kelvin. The use of the term "Apparent Color Temperature" was a warning that this was not an accurate description but only a guide. Unfortunately this useful description has been dropped and often replaced with "Correlated Color Temperature" (CCT) - read "Apparent Color Temperature." My objection to the use of this language is that it no longer suggests the warning. Even worse, since Correlated Color Temperature is a mouthful, lamp manufacturers will just describe a light source with an interrupted spectrum (all fluorescent and gas discharge lamps) as having its color temperature in a factor of Kelvin. This is a misuse of the Kelvin temperature scale, and can lead to many problems in film and television production or any application that requires accurate color reproduction.
Misuse of the Kelvin scale has grown with the use of light sources that do not behave in a manner similar to the "Theoretical Black Body Radiator." I can appreciate the need to use the Kelvin scale to describe the proximity of these newer light sources' color temperature to the well established incandescent and day light equivalents. However, there is no excuse for abandoning the caution "Apparent Color Temperature. "Correlated Color Temperature" does not imply as clear a caution as "Apparent Color Temperature." Worse yet, it suggests an accurate measurement, which it is not.
When one encounters a light source with an 'interrupted spectrum"(all fluorescent and gas discharge lamps), problems with color reproduction will arise. Just look at the full spectral distribution chart for that lamp. The peaks, often in yellow green range and elsewhere will cause problems. Color filters can be used to eliminate spikes. However, if the color is not in its spectrum, it cannot be added with a filter. Colors that fall in this interrupted area of the spectrum will be dulled or lost. If using these light sources is a must, it is best to augment lighting with a full spectrum source.