Stiles & Burch (1959) 10-deg individual colour matching functions

Corrected data

Individual 10-deg colour matching functions of Stiles and Burch (1959). The raw data have been tranformed from the instrumental primaries to primaries at 15500 (645.16), 19000 (526.32) and 22500 (444.44) cm-1 (nm).

There are 53 sets of individual data for 49 unique individuals (four indiviuals ran twice, see below).

Two conditions, I and II, were used--as noted in row 5. Full details about the two conditions, also refered to as Part I and Part II of the study, can be found in Stiles & Burch (1959).

Names are provided in row 6, sex in row 3 and age in row 4. Those names in row 6 highlighted in blue (Stiles, Burch, Gray & Wilson) appear in more than one column, having participated in the experiments both in Part I and Part II of the study.


Uncorrected data and correction factors


Uncorrected data, correction factors and corrected data as an Excel file.

Click on the large icon for the Excel file. The file contains three pages. The first page contains the uncorrected data. The second page contains the corrected data; these are the same as the corrected data given here, but tabulated in order of wavenumber rather than wavelength. Other details for both pages are as in the previous section. The third page gives that correction factors that must be applied to the uncorrected data to account for slit-width changes, uniformity, polarization, chormatic aberration and absorbtion by an auxiliary lens. Separate factors for Part I (CFI) and Part II CFII are given for the red, green and blue CMFs in turn. Again, for further details, see Stiles & Burch (1959). Some notes by Stiles & Burch that accompanied the original data are given in the next section.


Additonal notes by Stiles & Burch, 1958

N.P.L. Investigation on Colour-Matching Complete sets of colour-matching functions for individual subjects

References:

(i) "The Average Colour-Matching Functions for a Large Matching Field” W. S. Stiles. N.P.L. Symposium No.8 (Visual Problems of Colour) Proceedings: Paper No.7, 1958.
(ii) "N.P.L. Colour-Matching Investigation: Final Report (1958)” W. S. Stiles and J. M. Burch. NPL colour-matching investigation: Final report. Optica Acta, 6, 1-26.

In the reports of this work cited above, only the mean colour-matching functions for the groups of subjects are given, together with some information on individual variations expressed as standard deviations at each wave-number. This information is incomplete as the variations at different wave-numbers arc not independent. Probably the only satisfactory way of making good this deficiency is to give the complete sets of observed colour-matching functions for all subjects. With these, the expected effect of individual variations on any metameric match for which the approximate energy distributions are known can be calculated, and hypotheses on the character of the variations among different 'subjects can be tested. Only a few specialists are likely to require the complete individual data and normal publication would not be justified. The accompanying photographic reproductions of the data sheets arc available, to those concerned with relevant problems, on application to the Secretary, The National Physical Laboratory, Teddington, Middlesex.

Explanation of the data sheets

With the exception of the sheets headed CF I, CF II and CF II/I each sheet I gives the colour-matching functions of a single subject. The specification of the subject and the primaries and other observing conditions used for his measurements is contained in the key symbols at the top of the sheet. These symbols represent:

Thus “54 M 22 II” means subject number 54, a male of age 22, measured under conditions II.
The colour-matching functions are computed with respect to primaries at 15500, 19000 and 22500 cm-1 respectively. While three or at most four significant figures would be a sufficient accuracy and, for some cases, two or even one significant figure would suffice, it saved labour to print off all the values to four figures without adjusting each one to the number of figures appropriate to its accuracy (for details of individual repeatability see Ref. (ii)).

The seven columns represent:
Column 1 The wave-number
Column 2   The colour-matching function at this wave-number for primary15500 cm-1, omitting a power of ten: the sign of the colour-matching function is positive unless the figure is followed by a minus sign, when it is negative.
Column 3 The power of ten by which the figure in the previous column has to be multiplied to give the actual value of the colour-matching function for primary 15500 cm-1. The power is always negative as shown by the minus sign following the number, or zero and in the latter case there is no entry in column 3.
Columns 4 and 5.  These give the colour-matching function for primary 19000 cm-1 in exactly the same way as do columns 2 and 3 for primary 15500 cm-1.
Columns 6 and 7  Those give the colour-matching function for primary 22250 cm-1 in exactly the same way as do columns 2 and 3 for primary 15500 cm-1.

The asterisk following the colour-matching functions at the primaries merely recalls that the colour-matching functions are normalised to exact values at these wave-numbers.

Correction factors
The computations of the individual colour-matching functions were made before the final corrections to the data for slit-width, etc. had been made. It is these uncorrected data that are given here. The correction factors by which the colour-matching functions must be multiplied for any subject observing under conditions I are given in the sheet headed CF I, and the similar factors for the results obtained under conditions II in sheet CF II. In using the data for the purposes mentioned earlier it will almost certainly be immaterial if these correction factors are ignored provided the I and II groups are handled separately.    The corrections to be applied to Group II data to make them immediately comparable with the uncorrected group I data are given in CF II/I. It will be seen that these differential corrections are small, generally less than 1.5 per cent, so that even if the uncorrected I and II data are lumped together - there will be very little error in most applications.

Acknowledgement
The work from which the present data have been taken was carried out as part of the research programme of the National Physical Laboratory, and the data are made available by permission of the Director of the Laboratory.

W. S. STILES
J. M. BURCH
August 1958


References

Stiles, W. S. (1958). The average colour matching functions for a large matching field. Visual Problems of Colour, Volume 1 (pp. 209-247). London: Her Majesty's Stationery Office.

Stiles, W. S., & Burch, J. M. (1959). NPL colour-matching investigation: Final report. Optica Acta, 6, 1-26.

Wyszecki, G., & Stiles, W. S. (1982). Color Science: concepts and methods, quantitative data and formulae. (2nd ed.). New York: Wiley.