Page:JOSA-Vol 06-06.djvu/62

From Wikisource
Jump to navigation Jump to search
This page has been proofread, but needs to be validated.
584
L. T. Troland
[J.O.S.A. & R.S.I, VI

ences may now be reconverted into color valence terms, as directed under “(d)” of the preceding paragraph.

At the present time, owing to the unreliability of the magnitudes assigned to the luminosity valences for the three excitations, the conversion of data obtained by monochromatic analysis into excitation values, and thence into terms of other methods of color specification, cannot be accomplished with as great an accuracy as could be wished for. Present indications, moreover, are that these luminosity valences vary considerably among individual observers, without parallel variations in the color valences. Such variations evidently accompany deviations in the form of the observer’s visibility curve from normal, and demand that special care be taken in the selection of observers for use of the monochromatic method. In general, this method would appear to be more sensitive to the personal equation than the trichromatic and certain other methods.

The reverse conversion, of color excitation values into monochromatic specifications, is theoretically possible without ambiguity for all sets of values represented in the color triangle by points lying within the area determined by the spectral locus (vide supra). The easiest means for accomplishing this conversion consists in the use of a color triangle having represented upon it not only the spectral locus, but also the loci of the spectral colors, and purples, mixed with various proportions of white. In the absence of such a diagram, which is provided by Fig 9, a cumbersome “trial and error’ method is necessitated. To determine the monochromatic equivalents of any color excitation specification, its position should be plotted on the color triangle of Fig 9, and a straight line drawn through this point and the point representing the white (the white of the given monochromatic system). The intersection of this line with the spectral locus will indicate the dominant hue or wave-length,—direct or complementary, as the case may be—, and the relation of the color point on this line to its intersections with the loci for various percentages of admixed white will serve to determine the “per cent. white.”