American cinematographer (Jan 1930)

January, 1930 American Cinamatographer Seven "Photographic (Control in Fariable "Density "/^cording A Paper Included in the Technical Digest of the Academy of Motion Picture Arts and Sciences By Wesley C. Miller Chief Yransmission Engineer, Metro-Goldwyn-Mayec Studios S OUND reproduction presents three basic problems. First, the original sound must be picked up by some device—a microphone. Later, sound must be re-created by a loud speaker or its equivalent in such a manner as to reach a listener as a good substitute for the original. In between are all of the steps of amplification and recording, involving various kinds of energy conversions. The entire philosophy of the process is based upon the attempt to keep the results of each step strictly proportional to the original. In cases where this is not en- tirely practicable, correction may be made in one step for some unavoidable distor- tion introduced by another. For the purpose of a photographic analysis of variable density recording and reproducing, suppose we modify this fundamental division and •use that shown in Fig. 1. Up to the point of exposing the sound negative we can maintain good proportionality with the original sound. Also, from the time the sound positive pro- duces electrical variations which later are converted to sound, we can maintain similarly good proportionality. But if the same condition fails to exist in the photographic part of the pro- cess, our overall result suffers. Let us then state the photographic problem thus—the transmission of a sound positive must be strictly proportional to the original negative exposure which it represents. Any technical process has certain limitations. Recognition of them is an incentive to remove them or to decrease their effect. In any case, to determine them and to work within them is good engineering. The author’s intimate contact with heavy produc- tion using the light valve variable density method since the beginning of commercial work with it has led to certain con- clusions in this respect. It is the desire in this article to describe some features of the method and to present a means of analysis which has been found to be of great value in controlling results. The latter is not original in principle. It has. however, the dis- tinct merit of affording a way of stating the problem simply and of readily determining how well practical success has been attained. The usual means of arriving at the con- ditions for proportionality in the photo- graphic processes is by reference to the familiar Huerter U Driffield curves for the emulsion used, such as those shown in Fig. 2 plotted between density D, and the logarithm of the exposure E. These curves exhibit the characteristic of having a curved toe in the region of under ex- posure, a curved shoulder at over ex- posure, and in between a section which approaches a straight line. The slope of the straight portion determines gamma (y), the contrast factor, which up to a certain point increases with time of de- velopment. The intersection of the straight portion extended, with the log E axis, determines the inertia i which is not of immediate interest. The straight part of the curve may be represented by the equation 1 D = log — — y (log E — log i) T where T, the transmission, is the ratio of transmitted light to incident light—through the film. This equation may be written — y T = k E This relation holds for both negative and positive, that is — 9n Tn = kn En — yp Tp = kp Ep where the subscripts n and p apply to negative and positive, respectively. The various k factors are constants. In printing, the positive exposure is proportional to the negative transmission, hence Ep = k' T„ From these relations we may determine the overall relation (Continued on Page 18) co&aoitmm Of t*»o*ooe Fig. 2 Fig. 3 Fig. 4 Fig. I