American cinematographer (Jan 1933)

8 American Cinematographer • January 1933 KEYS70MIM6 LOSSES HORnaL mflZE /6 \ :: ft. - Plutesooi bis' TfipL Z-l- COMfCCTED IMAGE ViQTH-IM E£gT trble-Z - T HE establishment of the new standard camera and pro- jector apertures is fresh in everyone’s mind, so that the importance of the picture area on the film, in rela- tion to the projector aperture, is fully recognized. Theatre projection conditions demand the otherwise un- desirable compromise of projecting the picture at an angle above the horizontal—departing from the right angle rela- tionship with respect to the screen. The focal length of lens used together with the de- termined angle of projection become reciprocal functions for establishing the degree and character of the projected picture distortion. This Projection Keystoning is another condition which in- fluences the effectiveness of the cameraman’s composition on the screen. Not only does this affect the relative frame width at the top and bottom of the picture, but it also affects the 3:4 proportion of the composition. Fig. 1 illus- trates the effect graphically and shows how the shape of the aperture plate is altered to compensate for the distor- tion caused by projecting at an angle. As will readily be appreciated, the amount of distortion increases with increase in angle of Projection, the Projection distance, focal length of lens, and the size of the screen picture. The greater the distortion, the greater the amount of the keystoning of the aperture plate that is necessary and consequently, the greater the masking of the picture area when it is thus “corrected.” In this article, we propose to analyze some of these fac- tors and determine the extent to which they influence the cameraman’s “lining up” and composition of a scene. First of all, it is desirable to analyze Fig. 1 which illus- trates the cause of and the compromise for “keystoning.” Projection As soon as we project at an angle from horizontal, the top of the screen is nearer the projector lens than the bot- tom. Obviously the width of the picture on the screen is greater at the bottom than at the top, as shown by the dotted lines to the right of Fig. 1. The remedy adopted is to use an aperture plate which is narrower at the top than at the bottom. Figures 2 and 3, facilitate convenient determination of the various screen and aperture plate dimensions for dif- ferent conditions of projection angle, focal length of lens and throw. The average screen used in the majority of theatres is somewhere about 1 6 ft. x 22 ft. in size. In order to illus- trate the conditions involved in average practice, they are tabulated in Table No. 1. Table 1 is worth serious study and particular attention is called to the columns giving the percentage losses for key- stoning correction. These losses can also be read directly from Fig. 2 if desired. For example, taking the 3 inch lens at a 20 degree projection angle, read from the intersection of the 3 inch and 20 degree lines down to the figures at the bottom. This constant “C” is a multiplier to be used with the full plate width dimension to obtain the top width of the plate used in the keystone correction. In the instance given “C” reads .93 and indicates a 7 percent loss. It will be seen that the percentage keystoning loss varies from about 2 percent to 10 percent. The loss, in percentage, increases with increase of projection angle and also in- creases as the focal length of the lens decreases. On the average the loss will run about 6 percent. This means that about one-twentieth of the picture width at the bottom is sacrificed—it’s not much, but it is lost. What it really amounts to is that there is an average distortion of about this 6 percent. A doorway, for example, would be 6 percent wider at the bottom than at the top. The bottom of the door would be nearer the sides of the picture because the sides have Graph for determining projection angle and constant “A” for different conditions of height and length of throw. Sketch showing ex- tent of framing losses where correction has been made to the 3:4 proportion. POSSIBLE ERRORS tfi FRAMisc +T. plate T S41 PL4TE rOR LLKS-lS ' T~ .511 _L kr >j/ Li PcRTE EQR L’lEKS-LO* 3X4 icmttH NOTE - losses y>LL be rcouccd BY T**t Of T»C fig-5- L (FUJ) FIG-3 -