TO explain the action of tri-color filters it is necessary to recall the statement that all colors can be formed from three fundamental colors. But while we have seen that these three fundamental colors give rise to the three color sensations, it must not be thought that there is any closer connection between the sensation curves, shown in Fig. 3, and color photography. The color sensation curves are quite useless for color photography, although this is actually founded on their physiological action, except when we use the additive process, when the illumination of the three positives should be made by the physiologically correct or fundamental colors. But for subtractive work, and for the making of the negatives for the additive processes, the color sensation curves are not of the slightest practical value.
The first essential is to split up the spectrum or any colored subject into three regions; the one representing all the red, another all the green, and the third all the blue. It has been found that the most satisfactory results are obtained when the spectrum is divided into three zones, shown in Fig. 6, in which R represents the region transmitted by the red filter, G that of the green filter and B that of the blue-violet filter. A comparison of the curves with the color sensation curves in Fig. 3 will at once show that there is but little similarity; in the former the curves are gradual and all three overlap one another at many points, whereas with the filters they cannot be legitimately called curves, but are abruptly ending regions with limited overlap at only two points, in the golden-yellow, at the D lines, from 5900 to 5800, and in the blue-green from 5000 to 4800.
If we separate these three curves we shall obtain a fairly clear idea of the action of the filters. In Fig. 7, R shows the action of the red and the blank portion is that which corresponds to the shadows of the negative. In G the action of the green filter is shown, again with the blanks representing the shadows, and B represents the blue with the blank shadows. Now as it is the shadows of the negatives that print, the bare parts of the above diagram represent the parts that would print, therefore, positives or prints from the above would be represented by Fig. 8, in which the black parts show the parts that must not show in the additive process, that is when we project by colored lights; while in sub tractive processes, that is in prints, they represent the colors that must be present. We may, therefore, tabulate these results as follows:
Additive process Projecting color
Subtractive process Printing color
It will thus be seen that one may assume that the sub-tractive printing colors are the complementaries of the filter colors, and for this reason the corresponding negatives are frequently referred to as the "minus color" plates, as shown in the fourth column.
As the limits of the transmission of the filters are set by the above statement, we have merely to make filters that will show these cuts. The most satisfactory method is by the use of aniline dyes, dissolved in gelatine and cemented between glasses to protect the stained film from damp and mechanical injury, for every photographer knows that gelatine readily absorbs moisture, and that it is easily damaged by dirty or damp fingers.
In the early days of color photography it was usual to employ glass-sided cells filled with colored solutions, and they are still employed sometimes in commercial establishments, such as photo-mechanical studios. But for the dilettante they are the most unsatisfactory form of filter that can be used, and also when of sufficiently good quality to allow large aperture lenses to be used, are very costly, so that we shall not take them into consideration at all. The preparation of the correct filters does not present insurmountable difficulties for the average worker, and it should be looked upon as part of the preliminary training for every worker to make a set of filters. Yet it must not be overlooked that it is actually much more expensive to make them and the results are rarely as satisfactory as the commercial filters. However, we will proceed to describe the preparation of a set.
The apparatus required will be a sheet of plate glass, not less than 10 x 12 inches in size and about half an inch thick; three leveling screws and a level; some sheets of plate glass about one-fourth inch thick and sufficiently large to allow a margin of at least one-fourth inch all round the largest filter that is required; a beaker of 500 ccm capacity, another of 100 ccm, and an accurate 10 ccm pipette. The materials required are some soft photographic gelatine, of which about 100 g will be ample, and small quantities of the following dyes; rose Bengal, tartrazin, patent blue, acid rhodamin, tolui-din blue and naphthol green. About 10 g of each will be plenty. We also need some squares of patent white plate glass, about one-sixteenth inch thick and of the necessary size to cover the lens completely without cutting off any of the light from the plate, and some Canada balsam, dissolved in xylol, as used by micro-scopists, which can be obtained commercially; it should be noted that a chloroform solution of the balsam must not be used.
The large sheet of glass is merely used as a support for the gelatine-coated glass, so that the film shall set of an even thickness all over, and therefore, its actual size is not of real moment. It is frequently possible to pick up a fairly large piece from a glass dealer at a very reasonable price, because scratches and flaws will not matter. The smaller pieces of glass, one-quarter inch thick, are used to coat the dyed gelatine on, while the one-sixteenth inch pieces are used for the final filter. Both should have perfect surfaces. The reader may possibly wonder why the dyed gelatine cannot be coated directly on its final support; if this were done, the chances are that the filter would not be flat, for the gelatine contracts in drying and distorts the glass, that is bends it into a shallow curve, and the result would be that the filter would slightly alter the focus of the lens; as this filter curvature might not be equal in all three filters, the foci for the three colored images would differ and the images would be of unequal size, and possibly not equally sharp, and as we have to accurately superimpose the constituent positives, the outlines of the subject would not coincide and a confused jumble would be the result, or we would get color fringes.