THE possibility of using two instead of three colors was pointed out by Ducos du Hauron in 1895, and he suggested that red and blue were quite sufficient to produce a color result, provided that too bright a light was not used for examination, or the light was yellowish, or the support was tinted yellow. This idea was further utilized by J. Gurtner, who, in 1902, proposed to use orange and green as the printing colors, thus ignoring the deep reds. Some five years later G. A. Smith adopted the red and green dyad for the mixing of colored lights for cinematographic projection. That the results obtainable by a two-color process can never be theoretically correct is unquestionable; but such pictures, when examined by artificial light, are so satisfactory as to delude even experts. The only colors that are really defective are the violets and deep blues, assuming red and green as the printing colors.

Gurtner's method was to place two sensitive plates in contact. The film of the front one, which was a slow chloro-bromide emulsion, was stained orange and turned with its film towards the second plate, thus acting as a filter for the latter. Ducos du Hauron suggested a tri-pack; that is three sensitive plates or films made up into a block with filters in between, so as to limit the action of the spectral rays. Obviously, in both cases, but one lens was required and any ordinary camera could be used with slight alteration of the plate-holders. Some such process would be invaluable if practical, but there is a fly in the ointment.

It is impossible to obtain critical sharpness by any such method. In the first place, critical sharpness exists in one plane, and one plane only; but, with a comparatively small diaphragm in the lens, there is an appreciable distance along the optical axis, through which the sensitive plate may be moved without apparent loss of sharpness. But this is not sufficient to allow of three plates or films being used and critical sharpness obtained. The most serious obstacle to sharpness lies in the films themselves. If such a plate-block be outlined, the subject will be more readily grasped. The front element should preferably be a glass plate with its glass towards the lens, for if films are used it is difficult to obtain the necessary flatness of the surface. Behind this front element is placed the second sensitive surface, and a filter may or may not be interposed; this second element may be a film and here the sensitive surface may face or be turned from the lens. The third element is again preferably a plate, as this helps to keep the surfaces flat; a filter may or may not be interposed between the second and third films. The light has to pass through two sensitive films containing silver salts and these are normally by no means transparent. It is true that the front element of the pack may be one of the transparent Lipp-mann plates, as was suggested by Du Hauron, as this is acted upon by the blue rays, and can, therefore, be the slowest and consequently transparent. But when we come to the second sensitive surface, which is usually devoted to the record of the green rays, we can no longer allow it to be transparent, for it would be too slow. Therefore, it has to be an emulsion of the normal type, faster than the front but slower than the rear plate. Hence the plate is more or less translucent and the particles of silver salt diffuse the light and make it impossible to obtain in any case, no matter how thin this coating may be, critical sharpness on the rear element. In the case of the two-color process we can use a transparent front plate and a fast panchromatic plate in the rear, and obtain better results. There is also another factor that has been ignored in the consideration of these processes, and that is that the rear element, being the one on which the red is recorded, that is, the minus-blue plate, is printed in blue. Now this blue impression is the one that gives us what the artists call the "drawing" of the picture. The yellow print may be hopelessly out of focus, the red less so, and the results will not be objectionable. But the moment loss of sharpness is shown in the blue it is instantly detected.

Were it possible to overcome this defect, the tri-pack or bi-pack system would be one of the simplest to use for color photography, because the only alteration needed in the camera would be the alteration of the plate-holder so that it would take two or three plates.

For two-color work, however, the bi-pack offers considerable advantages. We can use a slow, more or less transparent chloro-bromide or chloride plate, and place behind this an orange filter and a panchromatic plate. If we decide to use commercial filters, such as Wratten & Wainwright, then we can safely say that the distance between the front and the rear elements will be only 0.1 mm, or approximately one two hundred and fiftieth of an inch, for this is the mean thickness of their gelatine filters.

Two Color Processes Bi Packs And Tri Packs PracticalColorPhotography 30

Fig. 26.

The best results are obtained in two-color work by using for the two negatives the regions shown in Fig. 26; R represents those rays which should act on the panchromatic plate and G those for the front plate; but unfortunately the sensitiveness of the normal chloro-bro-mide plate is shown in P by the continuous curve; while that of a chloride emulsion is shown as the dotted curve. There are no commercial plates of either kind which are color-sensitized, so we should have to sensitize them. The most satisfactory dye for this purpose would be the new German dye pinaflavol, but whether this is yet obtainable commercially is not known. This dye sensitizes far more satisfactorily than all others for just the particular region that we want to record on our front element. But it is also necessary to stain this front plate to cut out the action of the violets and deep blues, while this new dye is a basic dye and we do not know whether it would stand admixture with a staining dye, which must be an acid dye. The sensitizing and screening dyes may be combined in one bath and probably the following would be a suitable combination: