Journal of the Society of Motion Picture Engineers (1930-1949)

absorbers and other filters be removed from the optics of the densitometer and that the S-4 phototube be replaced with a phototube having an S-l response. Wavelength Tolerances A standard, equal energy response curve of a phototube with an S-l response is shown in Fig. 2. The peak sensitivity of a standard tube occurs at 800 m/z. Such phototubes were not designed or ever intended for precise photometric measurements. As a consequence of their general purpose design, wavelength tolerance limits of ±100 m/x have been established for the peak sensitivity. These limits are entirely satisfactory for their general purpose service but can produce large differences in the measured density of silver sulfide sound tracks. For the spectrophotometric curve of silver sulfide shown and the maximum and minimum wavelength tolerances for peak sensitivity of the phototube, it has been computed that the measured density of the sample would be 0.68 and 1.3 respectively. The potential error due to wavelength tolerance alone for peak sensitivity of the phototube is ±30%. On a practical basis, this error seems to be about ±10%. Color Temperature Effect The output of the phototube is the product of the equal energy response and the relative energy of the light source at every wavelength. The curves of Fig. 3 show the product of the standard response and the energy of tungsten sources at 2500 K and 3000 K.1 Of course, some densitometers may have light sources with higher or lower color temperatures than these, but this variation can introduce a significant error in the densitometry of silver sulfide deposits. If a standard S-l phototube is placed in a densitometer and the color temperature of the light source is changed from 2500 K to 3000 K, the measured density of this sample will change from 0.90 to 1.02. This is a density discrepancy of ±6%. Variations in Spectral Sensitivity Another cause of differences in measurements is the effect of different spectral sensitivity for tubes having peak sensitivities at the same wavelength. From a small group of tubes having peak sensitivity at the same wavelength, it appears that this potential error is of the order of ±3 or 4%, possibly larger. These sources of error are important because they are quite large and yet may be overlooked in the estimate of reliability of densitometry. The tendency is to assume that because measurements on a silver deposit in the visual region indicate an accuracy of 1 or 2%, that measurements on the silver sulfide deposit are of the same order of reliability. Measurements With Densitometers Having Visual Response Measurements of silver sulfide sound track densities by densitometers with phototubes having an S-4 response or by visual instruments are much more affected by the variation of density with wavelength than are measurements with instruments having phototubes with S-l response. Density measurements, in the visual region, of the silver sulfide deposits are more sensitive in detecting density variations than are measurements at 800 mju, but they are considerably less precise. In addition, sound tracks on Kodachrome films have some unwanted magenta dye present which, although virtually unseen by the usual reproducer phototube, will contribute significantly to the visual density measurement. Visual density measurements may, therefore, show variations which in no way affect sound reproduction. The significance of visual measurements is further reduced by the possibility that as less silver sulfide is formed, more magenta dye will be formed, and although the visual density measure Robert C. Lovick: Densitometry of Sulfide Tracks 91