American cinematographer (Sept 1935)

September 1935 • American Cinematographer 377 Dimensional Motion Pictures Possible? by George A. Mitchell, Hon. A.S.C, Member of Research Committee, American Society of Cinematographers picture mode from o moving object. Not everyone, how- ever, knows that this trick, in addition to its legitimate uses, has many times been used to misguide investors, and has lured much money into the pockets of unscrupulous promoters. Ruling out both these tricks of perspective and the un-commercial eye-aids, let us consider, in the light of the principles already illustrated in Fig. 1, the average inventor's approach to stereoscopic m.otion pictures. Most of them start with the presumption that if two pictures are taken from the viewpoints I and I', either on two separate films or alternately on the some film, then projected ster- eoscopic results will follow. We have seen from Fig. 1 what will happen. If the axes of the lenses are shifted to converge on post A, then in each alternate projected picture post B will first be on one side of A and then on the other. Point A will be stereoscopic, but all other points in view will appear to oscillote with point A as the axis. The rapidity of the oscillation will be the same as the speed of projection, and the amplitude of oscillation will be the same as the distance between I and I'. To arrive at these results many thousands of dollars have been spent on special cameras and equipment. Also, at this point, most of the experiments cease—the in- ventor begins to realize the complexity of the problem, end abandons the idea. These results may easily be proven at practically no cost by setting up a motion picture camera on some suit- able still-life scene, centering carefully on some object, and marking the starting frame. Then with the stop- m.otion crank expose each alternate frame, capping the lens between exposures. Expose as many feet as desired, rewind the film to the starting point, shift the camera to one side about pupillary distance (approximately ZVi"). Again center on the same object as previously (by pan- ning the camera), and expose the other frames in the same manner as before. A Bell & Howell camera lends itself admirably to this experiment, as the camera may be shifted sideways easily in its mounting dovetail. Many inventors proceed from this point assuming the solution lies in some trick optical system—prisms seem to be the most common panacea—to correct the oscilla- tion. But no distortion or prism system can change the relative position of objects in a scene. Their relationship to each other is determined completely the moment we select the point from which they are to be photographed. When we secure an optical system which will correct this trouble, we will find that we have again a flat picture. Stereoscopy has vanished—and we are back at the starting point. Cne idea, out of the many tried to overcome this dif- ficulty, might be mentioned as being somewhat different from the general run of schemes. In this case it is as- sumed that all rays passing through a lens from a given object are brought to a focus at the same point. Then a lens of large diameter might have two apertures located off center at opposite sides of the axis. If now alternate exposures are made through these apertures, stereoscopic pictures would result. Unfortunately, this idea does not work. Referring to Fig. 1, let I and I' be the two apertures. The conditions in front of the lens are the same as if two separate lenses were used. But back of the lens another condition now exists, an explanation of which is of interest. A lens of large diameter, when fitted with two apertures, creates two images of a given object, side by side, and while with parallel focal planes set at an angle behind the lens, will bring the two images of this object together, since the condition in front of the lens has not been changed. We have the same old trouble that objects at different dis- tances in the scene bear a different relationship one to the other. A stereoscopic still photograph which may be compared with Fig. 1. The chauffeur and the lady may be likened to points A and B in the diagram.