Projection print resolution as a function of illumination cone angle
OA Version
Citation
Abstract
Military aerial photography is, at the present, beginning a stage of transition. The camera systems in use until this time usually have produced photography resolving approximately ten to twenty lines per millimeter on the print, in practice. For such photography to be useful in reasonably detailed photointertation, it has been necessary that it be at relatively large scales, normally in the range of 1:5,000 to 1:15,000. Since high flight altitude is desirable to provide a reasonable probability of successfully completing a mission in the presence of opposition, the scale requirement has necessitated the use of long focal length cameras. However, in the newer aircraft, the space and carrying capacity available is decreasing drastically. It appears certain that to provide further increases in aerodynamic performance in the future, this trend will continue. In addition, technical intelligence requirements are becoming much more severe, requiring that aerial photography resolve much smaller dimensions on the ground than has been the case previously. Fortunately, a new generation of aerial cameras is being developed to keep pace, the so called "high-acuity" cameras. Resolution with these may be expected to exceed considerably that obtained with previous cameras. These high-acuity systems offer the potential for keeping pace with the increasingly severe intelligence, aerodynamic and photographic interpretation requirements.
The human eye may be considered to have a maximum resolving power of approximately ten lines per millimeter. Assuming that photography from high acuity systems will resolve at least fifty lines per millimeter, and possibly exceed one-hundred, a magnification of from five to more than ten will be required to fully benefit from this resolution. Such magnification may occur either in printing the negatives, in the viewing device used by the interpreter, or in some combination in the two stages. From several standpoints, however, it appears certain that an appreciable part will be in the printing process.
In photography, it is axiomatic that resolution is lost in virtually every process. In projection printing, this loss is attributable to not only the projection: lens and printing material, but also to other characteristics of the projection printer. One of these of potential importance is the cone angle of the illumination furnished by the printer at a given point on the negative. Projection printers fall into two classes: those employing diffuse illumination, in which the illumination cone angle may approach 180 degrees, and those employing illumination directed by a condensing-lens system, in which type the cone angle may be in the vicinity of twenty degrees. Most authorities agree that, other conditions being equivalent, a print made with a diffuse-type printer will have less contrast than one made with a condenser-type printer. From theory, any loss of contrast should result in a loss of resolution. To fully benefit from the increased capability of the new high-acuity camera systems, it is important that losses of resolution which may arise in the projection printing process be minimized. Accordingly, this investigation was made to determine the dependence of projection print resolution on the cone angle of the illumination incident at a point on the negative.
It can be shown that, so far as factors affecting print resolution are concerned, the difference in illumination cone angle represents the only practical difference between the two classes of projection printers. Therefore, the procedure used in the investigation consisted in varying the illumination cone angle in a projection printer in eight steps between the limits of seven and 180 degrees. Print resolution was evaluated, and the maximum at each step was plotted against the illumination cone angle of that step.
The illumination source used was a four inch diameter disk of flashed opal glass, evenly illuminated from the reverse side by six 15-watt daylight-type flourescent tubes. The various illumination cone angles at the negative were obtained by varying the distance between the negative and the source disk. Two negatives were used in the course of the experiment. Both were 1:22.1 reductions of an original Buckbee-Meers U.S. Air Force resolution target, were made in a resolving-power test camera, and were high-contrast targets. One was on an Eastman Kodak 548 High Resolution Plate and had a limiting resolution of 446 lines per millimeter; the other was on Eastman Kodak special emulsion S.O. 1213 on film, and had a limiting resolution of 125 lines per millimeter. The latter was the one intended to approximate photography obtained with a high-acuity camera system. The projection lens was a Schneider Componon 80 mm focal length f/4 enlarging lens. The lens was used at f/5.6 throughout the investigation; at this aperture and within the limited angular field used, it gave essentially diffraction-limited performance. The printing material employed was Kodabromide projection printing paper, series were made on both Contrast Two and Contrast Four papers.
For each cone angle (with each negative and printing paper contrast grade), exposures were made at five lens-to-easel distances differing by a small increment; this insured a run through best focus. At each distance, five exposures were made differing in exposure by a one-half stop increment in terms of time; this insured a run through optimum exposure. All prints were developed for 90 seconds in Kodak Dektol 2:1 at 68°F in a tray with continuous agitation, and fixed for seven minutes in one bath of Kodak Rapid Fix. Print resolution was evaluated by an experienced analyst; readings were made using a binocular microscope, reflected light, and either 13X or 30X magnification depending on the resolution level of the particular series. For a given level to be considered resolved, it was necessary that both the tangential and radial lines be resolved. Each print was evaluated twice, with the two readings being separated by approximately ten days.
Analysis of the data resulting from the experiment revealed that with the H.R. 548 and S.O. 1213 emulsions used, there is no significant dependence of projection print resolution on the illumination cohe angle at the negative. Some differences were found, but these were both small and inconsistent. They are felt to have been the result of random inconsistencies, principally in the reading of resolution values at the justresolved level.
The major factor which fubelieved to have been responsible for this lack of effect can be seen from a geometrical consideration of the passage of light through a negative. Consider light to be directed toward the negative at such an angle that, undeviated, it would reach the projection lens. For a clear area, the light will pass through largely undiminished and undeviated. Assuming no losses in the lens, this light will reach the image of the area. However, for an area having density, three effects will occur. A portion of the light will be absorbed in the negative and thus will be lost, photographically. Another portion will pass through undeviated. A third will be scattered in all directions; only so much of this as is within the angle subtended by the lens aperture will reach the image of the denser area. Presumably the latter two portions when combined in the image will render the dense area in the proper tonal relation. Now consider additional light to be made incident on the two areas of the negative, with the direction of this light being such that, undeviated, it would not reach the projection lens. Obviously a portion of this light will also be scattered by the dense area, and a portion of this will reach the image of that area. This, then, will alter the tonal relation of the print; it will, in fact, reduce the contrast of the two areas. This accounts for the frequently mentioned difference in contrast rendition between condenser-type and diffuse-type projection printers. The latter supply a large portion of their light at angles which are larger than that subtended by the projection lens from a given point on the negative.
The reason that the foregoing did not affect the results of the present investigation lies in the fine grain nature of the negative materials used. While a coarse grained material of high density may scatter nearly all of the light transmitted, the same is not true of fine grained materials. The materials designed for use with high-acuity camera systems are fine-grained; and future materials will probably be even more so as emulsion technology advances.
Two other factors probably contributed to the observed lack of dependence of resolution on cone angle. One is the fact that for a negative illuminated by a flat Lambertian source, the intensity reaching a point on the negative at a plane angle "a" to the nomal is proportional to cos^3a. Further, the portion which is lost by reflection rapidly increases for larger values of angle a. The combined effect is that once a moderate angle is exceeded, further increases should have relatively little photographic effect. This, of course, applies regardless of the nature of the negative emulsion.
As a result of this investigation, two major conclusions have been reached:
a. With the newer films having very fine grain and a high resolution capability, any dependence of projection print resolution on negative-illumination cone angle is negligible.
b. Projection printers intended for use with high-acuity camera systems need not be limited by a requirement for a particular illumination cone angle, but rather may be designed to best meet other pertinent considerations.
Description
Thesis (M.A.)--Boston University
License
Based on investigation of the BU Libraries' staff, this work is free of known copyright restrictions