Color Temperature and Correlated Color Temperature CCT

People sometimes get confused about the meaning of correlated color temperature (CCT) and the relationship of this metric to color temperature and the D series of CIE standard illuminants (e.g., D50 and D65). I will offer some scientific insight here to help explain the differences.

  • Each color temperature (e.g., 5000 K) is a single point on the Planckian locus in a chromaticity diagram (e.g., CIE 1931 (x, y) chromaticity diagram).
  • Each CIE standard illuminant in the D series (e.g., D65) is a single point on the CIE daylight locus in a chromaticity diagram (e.g., CIE 1931 (x, y) chromaticity diagram).
  • Each correlated color temperature (e.g. 6504 K CCT) is not a single point in a chromaticity diagram. Many points in a chromaticity diagram can have the same correlated color temperature.

The color temperature of light is based on the concept of the black-body radiator, also known as a Planckian radiator, and the Planckian locus on a chromaticity diagram. The unit of measurement of color temperature is kelvin (e.g., 6500 kelvin or 6500 K). Each unit of color temperature has a corresponding set of chromaticity coordinates on a chromaticity diagram, and those chromaticity coordinates are on the Planckian locus.

The ability to associate the two-dimensional chromaticity coordinates with the one-dimensional scale of color temperature along the Planckian locus enables a simpler communication of the visual appearance of nearly-white light. A given color temperature (e.g., 6500 K) gives us an understanding of the light relative to other color temperatures (e.g., 5500 K or 7500 K). A higher color temperature is more blue in appearance. A lower color temperature is more red in appearance.

Since all color temperatures are restricted to the Planckian locus, we have a problem when we want to use the color temperature scale to communicate the visual appearance of nearly-white light that comes from a light source that produces a spectral power distribution that is different from a black-body radiator. Many light sources — particularly fluorescent lights — produce a spectral power distribution that is different from a black-body radiator.

Well, the color temperature scale was deemed too useful to be restricted to black-body radiators and the Planckian locus. The solution is a less restrictive correlated color temperature (CCT) scale. The correlated color temperature scale is based on the color temperature scale and isotemperature lines, which were proposed by D. B. Judd in 1936. All colors along a given isotemperature line have the same correlated color temperature. Here is Judd’s proposal for isotemperature lines:

“ The estimation of nearest color temperature has been facilitated by the preparation of a mixture diagram on which is shown a family of straight lines intersecting the Planckian locus; each straight line corresponds approximately to the locus of points representing stimuli of chromaticity more closely resembling that of the Planckian radiator at the intersection than that of any other Planckian radiator. ” (771)

The CIE adopted Judd’s proposal for isotemperature lines, but the concept was updated with the CIE 1960 uniform chromaticity scale (UCS) diagram, which was not available in 1936. The isotemperature lines are perpendicular to the Planckian locus in the CIE 1960 UCS diagram, but not in other color spaces (Note: The CIE 1960 UCS diagram has some advantages over other color spaces, and the ease of calculating isotemperature lines is one of them). The chromaticities of the CIE 1960 UCS diagram are denoted by u and v to distinguish the color space from the CIE 1931 (x, y) chromaticity diagram. The CIE provides equations that enable a conversion of chromaticity coordinates from one CIE color space to another. Therefore, we use the convenience of the perpendicular relationship in the CIE 1960 UCS diagram to determine the correlated color temperature of a light source by plotting the isotemperature line from a given set of u, v chromaticity coordinates for the light source to the Planckian locus. We can also translate the coordinates of any given isotemperature line in the CIE 1960 UCS diagram to any other color space that is useful.

The correlated color temperature scale has been, and continues to be, a useful means for companies to describe light sources and for users to specify lighting requirements. But there has been some confusion because the metric is not precise and is not comprehensive. For example, light sources with the same correlated color temperature can deliver different color rendering indices (CRI). If you are working in an environment where color rendering is important, then I recommend getting three metrics for a light source: 1) the correlated color temperature, 2) the white-point chromaticities in one of the CIE chromaticity diagrams, and 3) the color rendering index.

In summary, here is another way to describe the difference between color temperature and correlated color temperature:

  • Color temperature is a metric used to describe a color of light on the Planckian locus and produced from a Planckian radiator. This is a rather limited metric because it is only applicable to a color of light from a Planckian radiator. Each unit of color temperature has one set of chromaticity coordinates in a given color space, and that set of coordinates is on the Planckian locus.
  • Correlated color temperature is a metric used to describe a color of light located near the Planckian locus. This metric has broader utility because it is applicable to a variety of manufactured light sources, where each light source produces a spectral power distribution that is different from a Planckian radiator. However, it is less precise than the color temperature metric because many points in a chromaticity diagram along an isotemperature line will have the same correlated color temperature.

I will close with the description of correlated color temperature in CIE Publication 15.2:

“ The correlated color temperature of a given stimulus is the temperature of the Planckian radiator whose perceived colour most closely resembles that of the stimulus at the same brightness and under the same viewing conditions. ” (38)

Post written by Parker Plaisted

References:
D. B. Judd, “Estimation of Chromaticity Differences and Nearest Color Temperature on the Standard 1931 I.C.I. Colorimetric Coordinate System,” Journal of Research Nat. Bureau Standards, Vol. 17, 771-779 (1936).

Colorimetry, second edition. CIE Publication 15.2 (1986)

G. Wyszecki and W. S. Stiles, Color Science: Concepts and Methods, Quantitative Data and Formulae, John Wiley & Sons, New York, N.Y. (1986).

Monitor Calibration: D65 or 6500 K?

In my earlier post, I explained why the CIE standard illuminant D65 has a correlated color temperature (CCT) of 6504 K, not 6500 K. This is a trivial difference that is generally not perceptible, but it has led to some confusion when people use monitor calibration software.

Some of the software applications that are available for calibrating monitors allow the user to specify a target white point for the monitor calibration. The white point can be specified by selecting one of the CIE standard daylight illuminants (i.e., D50, D55, D65, etc.) or by selecting the desired color temperature on the Kelvin scale. Some of these software applications show the corresponding CIE chromaticity coordinates for the selected white point. People who are unfamiliar with the fact that D65 has a CCT of 6504 K think that D65 and 6500 K are interchangeable labels for the same CIE chromaticity coordinates, but they are not. And this is where that trivial difference causes confusion.

You may be asking why I am obsessed with this issue. I ran into this trivial difference when I was working on the development of OptiCal 2.5 in 1998, and it was my task to make sure that the OptiCal software had the right numbers for the CIE standard daylight illuminants, the CIE daylight locus, and the Planckian locus. I will also share with you the insight that monitor calibration software uses the CIE chromaticity coordinates to set the monitor white point, so it is very important for the software to have the right chromaticity coordinates associated with the CIE standard daylight illuminants and the color temperature scale.

If you read my earlier post about D65, then you know the CIE daylight locus and the Planckian locus are different, and the movement of the Planckian locus in 1968 changed the CCT of D65 from 6500 K to 6504 K. So now we have three ways to specify the white point for a monitor: 1) a CIE standard daylight illuminant, 2) the correlated color temperature on the CIE daylight locus, and 3) the color temperature on the Planckian locus. Each of these specifications of the white point has a different set of CIE chromaticity coordinates. The following chart shows three examples for each of the three methods of specifying the target white point.

CIE chromaticity coordinates

The CIE chromaticity coordinates shown in the chart provide a quick reference for you to use to evaluate monitor calibration software. If the software provides the ability to select the target white point by selecting a CIE standard daylight illuminant or a color temperature on the Kelvin scale and provides the CIE chromaticity coordinates for the white point you selected, then you can see if the software is using the right CIE chromaticity coordinates by comparing those numbers to the numbers in the chart shown in this blog post. To be clear, the CIE chromaticity coordinates in the chart above are based on the color-matching functions of the CIE 1931 standard observer (also known as the CIE 2 degree observer or the CIE 1931 Standard Colorimetric Observer).

Note: If the monitor calibration software allows you to specify the color temperature of the white point for the monitor, then the software must be using the data for the Planckian locus or the CIE daylight locus. To be technically accurate, the color temperatures for the CIE daylight locus are correlated color temperatures because the CIE daylight locus is not on the Planckian locus, but the monitor calibration software may not make this distinction between color temperature and correlated color temperature.

Post written by Parker Plaisted

References:
G. Wyszecki and W. S. Stiles, Color Science: Concepts and Methods, Quantitative Data and Formulae, John Wiley & Sons, New York, N.Y. (1986).

Colorimetry, second edition. CIE Publication 15.2 (1986)

The CIE chromaticity coordinates for the Planckian locus were calculated and posted online by Mitchell Charity at http://www.vendian.org/mncharity/dir3/blackbody/UnstableURLs/bbr_color.html

D65 has a correlated color temperature of 6504 K

The most common white points in digital color workflows are D50 and D65. Both D50 and D65 are D-Illuminants with xy chromaticity coordinates specified by the CIE (Commission Internationale de l’Eclairage). The “D” stands for daylight. The number is a loose reference to the correlated color temperature of the white point on the Kelvin scale. This is a loose reference because the current correlated color temperature of D65 is 6504 K, not 6500 K. To quote Wyszecki and Stiles, “CIE standard illuminant D65 represents a phase of natural daylight with a correlated color temperature of approximately 6504 K.”

Let me share with you the details on why the D65 white point has a correlated color temperature of 6504 K.

The xy chromaticity coordinates for D65, based on the relative spectral radiant power distribution for D65 and the CIE 1931 color-matching functions, are x=0.3127 and y=0.3290. These xy coordinates are on the CIE daylight locus in the CIE 1931 xy chromaticity space.

When the CIE defined the relative spectral radiant power distribution for D65, the correlated color temperature of the CIE xy coordinates of D65 on the Planckian locus for blackbody radiation was 6500 K.

In 1968 the scientific community changed the second radiation constant (c2), which is used in Planck’s radiation formula for blackbody radiation, from 0.014380 to 0.014388 (Reference: The International Practical Temperature Scale of 1968, IPTS-68). This change in the second radiation constant changed the location of the Planckian locus for blackbody radiation in the CIE 1931 xy chromaticity space. The shift of the Planckian locus away from the CIE daylight locus changed the correlated color temperatures of the CIE D-Illuminants. Thus the correlated color temperature of D65 shifted from 6500 K to 6504 K.

Post written by Parker Plaisted

References:
G. Wyszecki and W. S. Stiles, Color Science: Concepts and Methods, Quantitative Data and Formulae, John Wiley & Sons, New York, N.Y. 143-146 (1986).

Colorimetry, second edition. CIE Publication 15.2 (1986)

The International Practical Temperature Scale of 1968 (IPTS-68)