International projectionist (Jan-Dec 1935)

DISTORTION (projection and viewing) IN MOTION PICTURES THE whole subject of warped perspective and distortion can conveniently be divided into three phases in accordance with three causes which adversely affect true rendition of form on the screen and in the eye of the observer. These are: (1) The discrepancy between camera point of view and audience point of view involving the relative values of camera and projector lens focal lengths. (2) The vertical elongation of figures and the keystone effect resulting from projection from a point above the screen. (3) The error in perspective caused by the off-center view seen by many members of the audience. The three types of distortion are, of course, additive in their effects, but for the sake of simplicity in this discussion it seems best to treat them singly. Projection at an angle and viewing at an angle are causes of distortion which can not be eliminated in practical presentation; but since both these are controllable to some extent in the design of theatres, it remains of practical interest to consider these matters and perhaps to specify limiting values for the guidance of architects and theatre owners. The question of projection angle has been frequently discussed before the S. M. P. E., the most thorough treatment of the problem being that presented by the Projection Committee in 1929. The Committee recommended, first, that the angle be kept as near zero as possible, and then proceeded to analyze the amount of distortion as a function of the angle of divergence of the projected beam and of the angle which the axis of projection makes with the horizontal. They expressed the distortion as the percentage increase in height which results when the picture is projected at an angle to a vertical screen. To quote from this report: "Now, the maximum permissible amount of distortion is a matter on which there seems to be considerable divergence of opinion. The Committee feels that in recommending 5 per cent as the maximum increase in picture height, it is erring on the side of laxity rather than on that of rigidity." For practical projection conditions a projection angle of 17 degrees results in an increase of image height of about 5 per cent. The Committee therefore fixed on this angle as the limiting value. The author has been supplied with data concerning the projection angles From I. P., October, 1933 existing in one chain of theatres. These data show that about 60 per cent of the theatres in this group have projection angles equal to or greater than that recommended by the Committee, and about 80 per cent have an angle greater than the more rigorous specification of 12 degrees recommended in the standards adopted by the Society as the limiting value. Since the theatres of large seating capacity have the larger projection angles, it follows that the vast majority of theatre patrons habitually see pictures which are projected at angles greater than 12 degrees. It would seem desirable either to bring the recommendation of the Society into line with practice or to exert pressure to bring future practice into line with our recommendation. It seems indeed that the 5 per cent increase in the ratio of height to width is a lax enough tolerance. Consider for a moment the effect on the human figure. The 17-degree projection angle should in its effect be roughly equivalent to the once highly advertised 18-day grapefruit diet. Greta Garbo, Ruby Keeler, Joan Crawford and others in the lightweight class apparently lose five or six pounds by the treatment. There is a strong tendency on the part of the general public to accept what it sees on the motion picture screen as the last word in fashion and beauty. Large projection angles may, therefore, be largely responsible for the vogue for slender figures. There is, of course, a second defect in an image projected at an angle which follows from the fact that the bottom of the screen image is magnified more than the top. What should be vertical lines in the image become convergent Illustrating binocular vision Object beyond point of fixation, seen double Left Eye Slfht Eye * By CLIFTON TUTTLE upward. A rectangular screen shape is maintained by shaping the mask, but nothing is. done to rectify the convergence of lines within the picture. It is well known that the eye is extremely sensitive to the lack of parallelism between two lines. Mitchell1 has recently thoroughly discussed this question for the edification of the cameraman, pointing out several matters which should be observed in scene composition. In some actual cases in theatre projection the convergence amounts to five or six degrees, which is very apparent at the edge of the screen. Fortunately, the attention of the audience is seldom concentrated on vertical lines in a picture. This is more true now than in the days of silent pictures with its numerous titles. It seems, therefore, that experimental demonstration of the practical effect of projection angle should be confined principally to such subjects as make up the greater bulk of motion picture presentation. In view of the data which show that many successful theatres have projection angles in excess of the arbitrary limit recommended by the Society, it seems of interest to proceed to experiment on the basis of determining the degree of distortion at which the illusion of naturalness breaks down. A number of still pictures of motion picture scenes were reproduced as lantern slides. These were projected on a screen at vertical angles which gave progressively 2.5, 5, 10, 15, and 20 per cent distortion. The screen picture was photographed at each position and lantern slides were made of the results. All the groups of slides thus obtained were thoroughly shuffled together so that during projection no one of a series of slides would follow another of the same subject. The slides were then shown to a group of persons, each of whom was asked to select all pictures which looked unnatural and to state the reason for the objection. Results of this test are summarized in Table I. In column one of this table,, the subjects have been classified in a general way. The terms "close-up," "semi-close-ups," and full-length figures" apply to human figures. Well-known inanimate subjects included pictures of houses, doorways, wagons (showing wheels) etc. The remaining columns headed by the value of distortion in per cent contain the record of the relative number of observers who objected. In lMitchell, R. F. : "Projection Keystoning from: the Cameraman's Viewpoint," Amer. Cinematographer, XIII (Jan., 1933), No. 9, p. 8. [lb]