If a chart made of filters of pure, saturated colours is reproduced by a colour screen plate or film, it will be found that, despite the brilliancy of the copy, its colours are diluted with a certain amount of white light. This fault of saturation is obviously magnified in the course of copying the first transparency.' I f, in spite of the large amount of white light they transmit, the colours of the colour-screen plates and films appear to be saturated, this is due to the fact that the opacity of the mosaic screen only allows them to be seen in weak light. Now, it is known that a coloured paper always appears of more intense colour in the shade than in the sunshine.
This loss of saturation is, moreover, not the same for all colours. It is most marked with yellow. If it were the same for all colours, it could be remedied by intensification, but this treatment would darken unduly the colours which are sufficiently saturated.
This phenomenon is due, at least partly, to diffusion and refraction within the starch grains and to irradiation within the layer of emulsion. The image of a red object, which, during the first development, should be limited to the parts of the emulsion situated behind the red grains, impinges on the adjacent parts. After reversal, the grains partly uncovered near the red grains will therefore transmit some green and blue light. In the case of the image of a yellow object, this spreading effect is even more marked, for the blue grains will be partly uncovered both near the red grains and near the green grains. These same causes would appear to lighten in the same degree the greenish-blue and the pinks, but these colours in a pure condition are found less often than pure yellows.
It will therefore be seen that while a colour screen transparency with vigorous tones and free from pure yellow may usually yield satis factory copies, it is impossible to reproduce satisfactorily a colour transparency with soft tones.
893. The copying of colour screen trans parencies on plates or films with a regular or irregular tricolour mosaic is done either in a printing frame or in a triple-body reproduction camera. The light must generally be filtered.
Excessive sharpness must be avoided; it is desirable that the image of each element should spread over some elements of the screen of the sensitive emulsion. For copying by contact, a light source of considerable size will therefore be used, and possibly the surfaces of the two films may be separated by interposing a thin trans parent sheet of film.
To reduce the loss of saturation of the colours it has been proposed to effect copying by utilizing, simultaneously or successively, three narrow spectral bands, each passing only the elementary filters of the same colour (G. B. Harrison, 1933 ; G. A. Raguin, 1933).
894. Colour Separation Negatives from Colour Screen Transparencies. Autochrome photo graphy has considerably enlarged the field of application of three-colour work by enabling it to be used for various subjects such as land scapes and scenes in movement for which direct trichrome selection is practically impossible. It
is, however, necessary to avoid the intermediary of a colour screen plate or film in all cases where it is possible to make the colour-selection nega tives directly from the original. Still more is it necessary, for the reasons already stated (§ 906), to abstain from attempting a three-colour reproduction from a duplicate.
Three-colour separation from a colour screen transparency does not involve actual selection, but a sorting of the three selection images inter mingled on the original plate. The employment of the normal selection filters would give only fairly bad results by adding up the defects of two successive selections. In order to utilize just as it is the selection already made by the elementary filters of the screen plate or film filters are employed which completely cut out two of the three intermingled images, and trans mit the third. To ensure these conditions, the regions of transparency of the filters must be narrower than those of the elementary filters (§ 884, footnote) and must be included within the latter, or at least they must have a very marked maximum near the centre of transmis sion of the corresponding elementary filter.' For sorting out the three images it is therefore sufficient to employ filters of which the spectral limits of transmission are, for instance— below 4,80o ; from 5,ioo to 5,700 ; above 5,900.
It should be noted that such filters are not suitable for selection proper.
895. The copying may be done with advantage in a printing frame, using a lamp fitted with a ground glass and, successively, with the three special filters (L. P. Clerc, 1907). The use of diffused light on the one hand, and of a very ample exposure on the other hand, reduces to a certain extent the granularity of the images.
It has been repeatedly suggested (L. Didier, 1908 ; J. R. Fuller, 1924) to obtain, under the same conditions, selection positives from a colour screen negative (a colour screen plate or film fixed after development), such positives being usable direct for synthesis by hydrotype print ing, or forming the prints required for photo mechanical reproduction. But a drawback of this method is that it is impossible to examine the quality of the colour negative before copying it.
In cases where the copying is done in a triple body camera with the colour transparency illuminated by artificial light, precautions need to be taken to avoid, during the long exposures necessary, any heating which may injure the transparency. A stream of air may be directed on to the latter by means of an electric fan.
The structure of the image can be almost completely eliminated without seriously im pairing the sharpness of the image, by setting the camera very slightly out of focus.