Technidyne Header Image

Monday, November 9, 2015

Color Measurement: What is Observer?

When making color measurements it is important to understand the terms Illuminant, Observer and UV-Level (or source). There will be three separate blogs addressing each of these points.

Standard Observer

The color data obtained from spectrophotometric measurements has no relationship to human visual assessments of color.  The spectrophotometer measures the reflectance of different wavelengths of light which is a physical phenomena.  It is of no value to measure color unless the measurements relate to what we see.  In order to make use of spectrophotometric data in the setting of product specifications, acceptance tolerances, etc., it is necessary to first convert the spectrophotometric data to a color scale which relates to visually observed color.

Figure A
In 1931, the CIE (Commission Internationale de l'Eclairage) established the "Standard Observer" utilizing and experiment depicted in Figure A.  Monochromatic light of individual wavelengths is produced by the prism system shown and is projected onto a white screen. Three controllable colored light sources, red, green and blue are also projected onto the white screen superimposed upon each other. To begin the experiment, a single wavelength of 400 nanometers is projected onto the screen. The observer then adjusts the three colored light sources until the resultant color matches the individual 400-nanometer wavelength. The observer then records the relative output of the three colored light sources. To continue, a single wavelength light source of 410 nanometers is then projected onto the screen and the observer produces a matching color by again adjusting the three colored light sources. This procedure was repeated over and over until individual wavelengths throughout the visible spectrum of 400 to 700 nanometers were reproduced by the three primary light sources, red, green and blue. After this experiment was performed by many observers, average observer responses were plotted and are shown in Figure B.
Figure B

            Thus, the CIE Tristimulus functions X (red), Y (green) and Z (blue) were developed. These results obtained in 1931 were based on a cone angle of 2º viewing. Color measuring instruments which faithfully reproduce the standard average human eye response normally produce excellent agreement with human color assessments. Figure 3.2 shows the response curves of Technidyne colorimeter (dashed curves) superimposed on the average human eye response (solid curves). Effective wavelengths for the CIE functions based on illuminant C, 2º observer are: X (595nm), Y (557nm), and Z (455nm).

In 1964, CIE developed a new set of tristimulus functions for a 10º observer to correlate with the viewing of larger samples. The relationship between the 2º and 10º standard observers is illustrated in Figure B. At approximately arms length the 2º observer would view an area the size of a U.S. quarter and the 10º observer would view an area the size of a softball.

Remember that the so-called “Standard Observer” is actually an Average Observer. Every person has a different spectral response from every other person and that response is continually changing. The CIE standard observer has quantified a single, average, spectral response, which can be built into every color-measuring instrument. The instruments will, therefore, all agree with each other but an instrument will only agree with a human observer if the human has average spectral response.


1931 Standard Observer (2º Standard Observer) - how the average person saw color as a result of the 1931 CIE experiment where samples were observed via an area of approximately 2º at arms length e.g. an area the size of a U.S. quarter

1964 Standard Observer (10º Standard Observer) - how the average person saw color as a result of the 1962 CIE experiment where samples were observed via an area of approximately 10º at arms length e.g. an area the size of a softball

Since a change in observer is just a calculation in a spectrophotometer, after the sample is measured, color values for different observers can be determined with just a click of a button. It is important to understand, though, that color values obtained under one observer will be different than those obtained under a different observer, This is because the size of the area being viewed does have an effect on how people see color. 

In general, if illuminant C is used the observer is normally 2º.  Likewise, if standard illuminant D65 is used the observer is normally 10º.