International projectionist (Jan-Dec 1935)

May 1935 INTERNATIONAL PROJECTIONIST 19 dividuals participating in the test were not informed of its object, and they probably were neither less nor more critical than members of the usual motion picture audience. The conclusion from these data is that for scenes in which actors or actresses form the principal interest, the image can be distorted 10 per cent before the illusion of naturalness is impaired. Stated in another way, this means that most persons unacquainted with what the real subject looks like are satisfied even though the picture is decidedly distorted. In the case of very familiar objects of definite shape as, for instance, a picture in which there was an axial "view of a wagon wheel, the tolerance is somewhat narrower — some persons objecting when the image is distorted as much as 10 per cent. Definition and Area Losses This discussion thus far has been limited to true image distortion. There are two other undesirable effects, however, which accompany projection at an angle. The first is the loss of image area which follows inevitably when the sides of the picture' aperture are cut to make the frame rectangular. The second is the effect upon top and bottom image definition because of the path difference. Though neither of these can properly be classified as distortion, an evaluation of both has been included in Table II. The third column shows approximately the area loss in per cent which results when a picture is projected with a 6inch lens to a rectangular vertical screen, and the fourth column shows the diameter of the circle of confusion, or, more accurately, the major axis of the ellipse of confusion at the bottom and top of the field for a perfect lens working at an aperture of f/2. The area loss, that is, the amount which must be masked from the lower corners of the picture, is not particularly serious at any commonly used combination of projection distance and focal length — provided that the elongated picture is not masked off to maintain the 3x4 picture. The lack of definition as judged from a viewing distance of twenty feet becomes noticeable if the image size exceeds about 1.5 mm. Since the image size resulting from the projection angle is superposed upon the effect of the lens aberrations, it is probably reasonable to state that with practical projection lenses of f/2 aperture, definition suffers notice ably at projection angles greater than 17 degrees. The second matter upon which some experimental data may prove of advantage concerns the viewing angle forced upon all the members of an audience except those located opposite the screen center. At first thought, this problem appears to be closely related to the former one. Both projection at an angle from the horizontal and viewing at an angle other than the normal, produce a similar kind of elongation of the screen image, and one might expect that a distortion tolerance set up for the one case might apply to the other. The conditions, however, are somewhat different: a person viewing the screen from an angle is conscious of his point of view, and instinctively makes a correction for some distortion of the image. He is not conscious of the projection angle, and therefore has no means of compensation which will aid him in rectifying his concept of the picture. It is common experience that motion pictures viewed from extreme front and side seats in some theatres appear badly distorted. At the same time, it is true that one's enjoyment is not adversely affected until the angle becomes fairly large. The author has attempted to determine the limiting angle experimentally by projecting before a group of persons motion pictures of a screen image photographed at different angles. There seems to be quite good agreement among the persons before whom these and other pictures of the same kind were shown that any angle less than 30 degrees is not objectionable. Forty degrees seems to be passable, but the opinion was unanimous that the illusion is spoiled at angles greater than 40 degrees. One can not say that these demonstrations adhere closely enough to theatre conditions to justify any general conclusions. One is quite justified in asking: How much does motion in the picture affect the feeling that the illusion has failed? And how much does the angular field of view change one's judgment? A few trials were made in which a large black border was shown around the rotated picture. This, is was thought, would supply a comparison reference as to the amount of foreshortening to be expected in the picture. The judgment of distortion did not seem to be much changed. Pictures filling the screen appear better to represent the view which a member of the audience has in a seat close to the screen — the only location in which the viewing angle problem is serious. On the basis of a number of observations in theatres during the projection of pictures, it appears that motion in the picture does not affect the result to any great extent. A seat which forces one to see any part of the picture at an angle greater than 40 degrees is undesirable. For fulllength figures the judgment is more critical. The larger the screen, of course, the worse is the distortion at the farthest edge at a given viewing distance. In recommending practice for the guidance of designers and architects, the specification for the position of the extreme seat should be based upon the angle at which this edge can be seen. Assuming that the distance from the first row of seats to the screen is equal to 1.5 times the screen width, the first row of seats should be not longer than 1.5 times the screen width if the extreme viewing angle for the edge is not to exceed 40 degrees. Data supplied to the author give the average maximum viewing angle of a number of theatres at 34 degrees at the screen center — an angle which makes the edge viewing angle somewhat in excess of the distortion limit. Distortion Remedies Many attempts have been made in the past to cure the evil of viewing angle distortion — usually by the use of curved screens. Anamorphotic lens systems also have been suggested. The fallacy of such suggestions has been pointed out so frequently that apparently no proposal of this kind has been made for several years. There is, of course, no remedy except proper design of the seating space. Any correction of the screen image for one position can be made only at the expense of the perspective from other positions. Correction of projection angle distortion is not so impossible theoretically. Partial compensation can be effected by tilting the screen. This means probably presents some mechanical difficulties, since it is not commonly used. Other partial remedies are at least theoretically possible. Putting the results of the experimental data in the classic form attributed to Barnum, most of the people most of the time are totally fooled up to a projection angle distortion of 10 per cent. All of the people all of the time are dissatisfied with a seat which forces them to see part of the screen at an angle greater than 40 degrees. TABLE I Summary of Data Showing Susceptibility of an Average Audience lo Different Degrees of Distortion TABLE II Effect of Projection Angle upon Screen Definition and Loss of Area with Rectangular Masking Close-ups Semi-close-ups Full-length figures Well-known inanimate subjects Relative number of observers who objected per cent of total number of observers 15 20 0 0 0 0 0 20% 0 0 0 0 5% 25% 0 0 0 0 15% 50% 0 0 0 10% 40% 50% Projection Angle (degrees) Per Cent Distortion Area Loss, Per Cent Image Size Bottom and Top of Screen Mm 0 0 0 12.5 2.5 2 0 9 17 5 5 0 3 1 3 21.5 7.5 4 1 7 24 5 10.0 5 2.0 27.0 12.5 5 2 2 29 5 15 0 G 2.6