Colorimetry: A brief history

Editor's note: This is one in a historical series of documents published in our Lab Notes series around the turn of the century. The series is a snapshot of current understanding at that time, we hope that you find it interesting.

Starting in the early 1800s investigations were made into the perception of colour and how it relates to the physical properties of light. Over time, a number of metrics to describe the relationship have been formulated.
 The Commission Internationale De L’Eclairage, otherwise known as the CIE was established in 1913 and is the internationally recognised authority on photometry and colorimetry. It was integral in guiding the development of colour related metrics.

Trichromatic colour measurement

Figure 1: The VLambda curve

The idea that a colour could be matched by a combination of red, green and blue stimuli has been accepted since 1853, when Grassman first postulated this theory.
A later milestone in the development of the CIE Colour System was the acceptance in 1924, of the standardised spectral response curve of the human eye which is usually known as the VLambda curve.
In 1955 the CIE formally adopted the so called Colour Matching Equation. This was based on the premise that a certain quantity of red light (R), plus a certain quantity of blue light (B), plus a certain quantity of green light (G), could match to give a given quantity of most colours that can be perceived by the eye (C).

Written as a shorthand algebraic equation this can be stated as below, in this equation the quantity of C represents 1 trichromatic unit.
 
C  = 1   [is matched by]      r (R) + b (B) + g (G) 

Therefore :
r =         R              b =      B           g =        G     
         R+G+B               R+G+B               R+G+B

Hence : r + b + g = 1.

The quantity C may be found on a two dimensional CIE chromaticity chart. Note that the chromaticity does not encode any information about brightness.

Figure 2: CIE chromaticity chart

CIE 1931

Fig 3. RGB factors

By the 1930s colour measurement had moved from observer comparisons to a Colorimeter. This move to an instrument based approach refined colour measurement accuracy and reproducibility. In 1931 J. Guild published in the results of his experiments which showed the spectral co-ordinates for the three primary colours, Red, Green, and Blue, matched against a standard white light (Figure 3). At each wavelength the R,G,B co-ordinates added to 1.0 however one or more of the co-ordinates were always negative.

In 1930 D.B. Judd had put forward a proposition that if three new imaginary stimuli to be known as X, Y and Z could be used, the disadvantages of the R,G,B, system could be overcome. When the colour triangle was redrawn as in Figure 4, most of the colours fell within this new triangle.

In addition, the standard colour “White”, which later became known as the “equal energy white”, is matched as follows :
 
W [is matched by] 0.333 (X) + 0.333 (Y) + 0.333 (Z) 
Further refinement of the system found that a spectral locus could be drawn within the X,Y,Z triangle as depicted in Figure 5. The primary advantage of this method is that all of the real physical stimuli lie within this X,Y,Z triangle, and the line X,Y lies along the spectral locus at the red end of the spectrum where the locus is a straight line. In this way the blue co-ordinates are made to equate to zero, so simplifying the calculation procedures. Similarly the blue/red line is also a straight line.

Fig 5. XYZ representation

The 1931 CIE Chromaticity Chart is plotted in terms of the reference stimuli of X, Y and Z with units based on an equal energy White (Figure 6). To calculate the colour co-ordinates for a specific light source requires its spectral power distribution, and the colour matching functions xbar, ybar and zbar.

The colour matching functions were determined empirically in the late 1920s through experiments conducted independently by Wright and Guild. The human observers matched pure monochromatic colors using a combination of red, green, and blue light primaries to establish a 2° Standard Observer.

Further research undertaken by Stiles and Speranskaya in the 1950s indicated that a 10 degree field of view was more appropriate for observers. The colour matching functions for a 10° observer were adopted in 1964.