Airplane Photography

Focussing.—The process of focussing aerial cameras was at first deemed a mystery, though undeservedly so. A belief was long current that “ground” focus and “air” focus differ. In other words, that a camera focussed upon a distant object on the ground would not be in focus for an object the same distance below the camera when in the plane. Belief in this mysterious difference went so far that certain instruction books describe in detail the process of focussing a camera by trial exposures from the air.

Careful laboratory tests performed for the U. S. Air Service showed that neither low temperature nor low pressure, such as would be met at high altitudes, alter the focus of any ordinary lens by a significant amount, and that the possible contraction of the camera body was of negligible effect on the focus (not more than 1
200 per cent. per degree centigrade with a metal camera). In complete harmony with these tests has been the experience that if the ground focussing is done carefully, by accurate means, then the air focus is correct. The whole matter thus becomes one of precision focussing.

The best method, applicable if the air is steady, is to focus by parallax. The ground glass focussing screen is marked in the center with a pencilled cross. Over this is mounted, with Canada balsam, a thin microscope cover-glass. The camera is directed on an object a mile or more away, and the image formed by the lens is examined by a magnifying glass through the virtual hole formed by the affixed cover-glass. With the pencil line in focus the head is moved from side to side. If the image and pencil mark coincide they will move together as the head is moved. If the image moves away from the pencil mark and in the same direction as the eye moves, the image is too near the lens. If the image moves away in the opposite direction to the motion of the eye, it is too far from the lens. In either case the focus is to be corrected accordingly.

In place of a distant object, which may waver with the motion of the air, we may use an image placed at infinity by optical means. The collimator, an instrument for doing this, consists of a test object (lines, circles, etc.) placed accurately at the focus of a telescope objective. The camera lens is placed against this and focussed by parallax, as with a distant object. Collimators are employed in camera factories, and should be part of the equipment of base laboratories where repairing and overhauling of cameras is done.

Lens Mounts.—All that is required for the mounting of an aerial camera lens is a rigid platform, with provision for enough motion of the lens to adjust its focus accurately. As already explained, the lens works at fixed, infinity, focus, and therefore needs no adjustment during use. It must be held far more rigidly than would be possible by the bellows, which is an almost invariable adjunct of focussing cameras. The use of ordinary types of hand cameras on a plane is rarely successful just because of the bellows, which is strained and rattled by the rush of wind.

The lens mountings thus far used have been simple affairs. In the French cameras the lens is merely screwed into a flange which in turn is fastened by screws to a platform in the camera body. Adjustment for focussing is not provided; instead, the flange is raised on thin metal rings or washers, cut of such thickness by trial as to bring the lens to focus, once and for all.

The U. S. Air Service method of mounting is to provide the lens barrel with a long thread, which screws into a flange that in turn is mounted on a platform in the camera cone, by means of thumb-screws. The lens is focussed by screwing in and out, and then clamped by a screw through the side, bearing on the thread. The whole mount may be quickly removed by loosening the thumb-screws, and once focussed in one cone, can be transferred to another similar, machine-made cone without change of focus. Fig. 18 shows a 20 inch lens mounted in this manner. The photograph shows as well the ring on the front of the lens by means of which circular color filters may be held in place. This ring screws down on the filter, and the catch is dropped into the nearest vertical groove to the tight position.

A somewhat different and better method of tightening the lens in the flange, when focussed, has been adopted in the English lens mount, which is in general similar to the American. The threaded part of the flange is split by a slot cut parallel to the flange base, and a screw is run into the flange from the front, through the split portion. By tightening this screw, which is always accessible, the split part of the flange is squeezed together, thus rigidly holding the lens barrel.

Permissible Exposure in Airplane Photography.—A definite limitation to the length of exposure in airplane cameras is set by the motion of the plane. If we represent the speed of the plane by S, the altitude of the plane by A, and the focal length of the lens by F, we obtain at once from the diagram (Fig. 19), that s, the rate of movement of the image on the plate, is given by the relation,

If we call the permissible movement d, then the permissible exposure time, t, is given by the relation—

As a representative numerical case, expressing all quantities in centimeters and in centimeters per second, let F = 50, S = 20,000,000
3600 (200 kilometers per hour), and A = 300,000, then

If we take for the permissible undetectable movement, .01 centimeter, which is, as has been shown, a reasonable figure for lens defining power, we have, then, that the longest permissible exposure is .011 second—in round numbers, one-hundredth.

In flying with a slow plane, or in flying against the wind, the exposure can sometimes be increased to as much as double this length. Diminishing F would similarly extend the allowable exposure, but the ratio of F to A approximates to a constant in actual practice; in other words, a certain resolution and size of image have been found desirable. If flying is forced higher, a longer focus lens is used; if lower flying is possible, a lens of shorter focus. This relationship has, of course, been derived from war-time experience. Probably much of the prospective peace-time mapping work will impose substantially easier requirements as to definition and will thus allow longer exposures.

For low oblique views the longest exposure is much less. Taking 45 degrees as a representative angle for the foreground, and 500 meters as a representative height, the value of t becomes 1
600.

These figures will illustrate two important points: they show how severe is the limitation as to exposure, with the consequent heavy demand on lens and sensitive material speed; and they show how important it is to secure a shutter with the maximum light-giving power for a specified length of exposure. This leads to a study of the characteristics as to efficiency of the two common types of shutter, namely, shutters at or between the lens, and focal-plane shutters.

Characteristics of Shutters Located at the Lens.—Of the various shutters located at the lens the most common is the type that is clumsily but descriptively termed the “between-the-lens” shutter. This is composed of thin hard rubber or metal leaves or sectors which overlap and which are pulled open to make the exposure. It may require two operations, one for setting and one for exposing, or it may, as in some makes, set and expose by a single motion. Clock escapements, or some form of frictional resistance, are depended on to control the interval between opening and closing. This shutter is the one almost universally employed on small hand cameras and on all lenses up to about two inches diameter. It gives speeds sometimes marked as high as 1
300 second, although usually not over 1
100 on actual test.

Between-the-lens shutters have been used to some extent on the shorter focus (up to 25 centimeter) aerial cameras, notably in the Italian service. They suffer, however, from two limitations. In the first place we have not yet solved the mechanical problems met with in trying to make the shutter of large size (as for 50 centimeter F/6 lenses) at the same time to give high speeds. In the second place the efficiency of the type is low because a large part of the exposure time is occupied by the opening and closing of the sectors.

If we define the efficiency of a shutter as the ratio of the amount of light it transmits during the exposure to the amount of light it would transmit were it wide open during the whole period, then the efficiency of the ordinary between-the-lens shutter is of the order of 60 per cent. This means 1.6 times the motion of the image for the same photographic action that we should have with a perfect shutter. The accompanying photographic record (Fig. 20) of the opening and closing process of this type of shutter clearly illustrates its deficiencies.

Characteristics of the Focal-Plane Shutter.—Long before the days of aerial photography the problem of a high-efficiency high-speed shutter for photographing moving objects on the ground—railway trains or racing automobiles—had already led to the development of the focal-plane shutter. This is a type peculiarly adapted to the problems of the airplane camera. It consists essentially of a curtain, running at high speed close to the photographic plate, the exposure being given by a narrow rectangular slot.

If the focal-plane shutter is in virtual contact with the sensitive surface the efficiency, as defined above, is 100 per cent., since the whole cone of rays from the lens illuminates the plate during the whole time of exposure. But if the curtain is not carried close to the plate the efficiency falls off rapidly with distance, especially so for small apertures of the slot.

The efficiency of the focal-plane shutter may be calculated as follows: Let the focal length of the lens be F, its diameter be F
N, the width of the slot be a, and the distance from plate to curtain d (Fig. 21). Now if the curtain is moving at a uniform speed, the time taken for the slot to traverse the whole cone of rays, from the instant it enters till the instant it leaves, will be directly proportional to

If the curtain were in contact with the plate the time taken for the same amount of light to reach the sensitive surface would be proportional to a. Again defining shutter efficiency as the ratio of the light transmitted to what would have been transmitted were the shutter fully open for the total time of exposure, the efficiency, E, is given at once by the expression—

As an example let the lens aperture be F/6, so that N = 6; let d = 1, and a = 1, then E = 6
7. In the French deMaria cameras, where d = 4 centimeters, E = 60 per cent. for the aperture assumed, which is representative. Fig. 22 exhibits diagrammatically the chief characteristics of the focal plane shutter.

In view of the necessity for some distance between shutter and plate it is obviously important to keep a as large as possible, depending for the requisite shutter speed on the velocity of the curtain. Large aperture and high curtain speed are also found to be desirable when we consider the distortion produced by the focal-plane shutter.

Distortions Produced by the Focal-plane Shutter.—While the time of exposure of any point on the plate can, with the focal-plane shutter, easily be made 1
100 second or less, the whole period during which the shutter is moving is much greater than this. For instance, a 1 centimeter opening which gives 1
100 second exposure takes ⅒ second to move across a 10 centimeter plate, or nearly ⅕ second for an 18 centimeter plate. With a moving airplane this means that the point of view at the end of the exposure has moved forward compared to that at the beginning, by the amount of motion of the plane in the interval. If the shutter moves in the direction of motion of the plane the image will be magnified; if in the opposite direction, it will be compressed along the axis of motion. The amount of this distortion is calculated as follows:

Let the velocity of the plane be V, and that of the shutter be v. Let the focal length of the camera be F, and the altitude A. If the camera were stationary, a plate of length l would receive on its surface an image corresponding to a distance A
F × l on the ground. Due to the motion of the shutter the end of the exposure occurs at a time l
v after the start. In this time the plane has moved a distance V × l
v; hence the point photographed at the end of the shutter travel is Vl
v within or beyond the original space covered by the plate, depending on the direction of motion of the curtain. The distortion, D, is given by the ratio of this distance to the length corresponding to the normal stationary field of view:

When V = 200 kilometers per hour, v = 100 centimeters per second, F = 50 centimeters, A = 3000 meters, we have—

Or if the actual distance error on the ground is desired,

As a percentage error this one per cent. is small compared with other uncertainties, such as film shrinkage or the error of level of the camera. As an absolute error in surveying, thirty feet is, of course, excessive.

The distortion is diminished for any specified shutter speed by making the speed of travel of the curtain as large as possible and by correspondingly increasing the aperture. In connection with film cameras, another solution which has been suggested is to move the film continuously during the exposure in the direction of the plane's motion. The requisite speed of the film v' to eliminate distortion is given by the relation:

For the values of V, F, and A used above, v' = .92 centimeters per second. This speed is clearly that which holds the image stationary on the film—a fact which suggests another object for such movement, namely, to permit of longer exposures.

The effect of focal plane distortion may be averaged out in the making of strip maps, if the shutter is constructed so as to move in opposite directions on successive exposures. The first picture will be magnified, the second compressed, and so on, but a strip formed of accurately juxtaposed pictures will be substantially accurate in over-all length. Such a shutter is embodied in one of the German film cameras (Fig. 61).

Distortion of the kind above discussed is absent with between-the-lens shutters, which may conceivably be improved in efficiency and in feasible size. If so they would merit serious consideration for aerial mapping.

Methods and Apparatus for Testing Shutter Performance.—With a focal-plane shutter the desirable qualities in performance are three in number: (1) Adequate speed range, which may be taken as from 1
50 to 1
500 second for aerial work, (2) good efficiency, which has already been treated, and (3) uniformity of speed during its travel across the plate. Before the advent of aerial photography little attention was paid to speed uniformity, differences of 50 per cent. in initial and final speed being common in focal-plane shutters, and but little noticed in ordinary landscape work because of the natural variation of brightness from sky to ground. In the making of aerial mosaic maps the non-uniformity of density across the plate results in a most offensive series of abrupt changes of tone at the junction points of the successive prints (Fig. 140), an effect which must be minimized by manipulation of the printing light.

Instruments for testing the speed and uniformity of action of focal-plane shutters are an essential part of any laboratory for developing or testing photographic apparatus and some simple device for setting and checking shutter speed should be available in the field. Every such speed tester must contain some form of time counting element—pendulum, tuning fork or clock-work. Elaborate shutter testers, suitable for determining all the characteristics of all types of shutter, have been developed and used in certain of the photographic research laboratories. For the study and setting of focal-plane shutters (whose efficiency need not be measured, as it can be simply calculated from linear dimensions), the following simple kinds of apparatus are adequate:

Clock dial type of shutter tester. This consists essentially of a black clock dial carrying a white pointer which makes its complete revolution in one second or less. If this dial is photographed by the camera under test, the width of the sector traced during the exposure by the moving pointer shows the time interval. If the dial is photographed at several points on the plate—beginning, middle and end of the shutter travel—the complete characteristics of the shutter can be determined.

Interrupted light type of shutter tester. For the study of uniformity of shutter action alone the apparatus shown in Fig. 23 may be employed. A is a high intensity light source, such as an arc or a gas filled tungsten lamp. L is a convex lens, focussing an image of the light source on a small aperture in the screen E. D is a sector disc which, driven by the motor M, interrupts the transmitted light with a frequency determined by the number of openings of the sector and by the speed of rotation, which must be measured by a tachometer. The light diverging from the aperture in E falls upon the shutter S, which for this test is reduced to a narrow slit of one millimeter or less. Passing through the shutter opening the light falls upon the photographic plate P. The principle is simple: If the light is uninterrupted, the plate P is exposed at all points; due to the interruptions, a series of parallel lines of photographic action result, and their distance apart gives a measure of the speed of the shutter at any chosen point in its travel. A performance curve of the French Klopcic shutter is shown in Fig. 24. The variation in speed lies over a range of two to one. So serious is this defect in these shutters that diaframs are sometimes inserted in the French cameras to cut off part of the light from the lens on the most exposed end of the plate. This expedient produces uniformity of photographic action, but does not overcome the movement of the image, which is one of the chief faults of excessive exposure.

A more complete apparatus, adapted both to absolute speed determinations and to the study of uniformity of action, is that worked out and used in the United States Air Service (Fig. 25). At A is a high intensity light source, an image of which is focussed by the lens L₁ upon a slit E, in front of which stands a tuning fork T, of period 1024 or 2048 per second. The light diverging from the slit is received by a second lens, L₂ which is arranged either to focus the slit image upon the shutter curtain or to render the rays parallel, so that an entire camera may be inserted. In the latter case the camera lens L₃ serves to focus the slit image on the curtain C. After passing through the curtain aperture the light is focussed by the lens L₄ on the rotatable drum D, which carries a strip of sensitive film.

The operation of testing a shutter consists in focussing the slit image on the portion of the shutter whose performance is required, striking the tuning fork to set it vibrating, rotating the drum rapidly and setting off the shutter. There is thus obtained on the sensitive film an exposed strip resembling in appearance the edge of a saw, the number of teeth showing the time interval in vibrations of the tuning fork. Three exposures usually give all the points necessary for a practical knowledge of the shutter's uniformity of action. A point of some importance, learned from numerous shutter tests, is that a focal-plane shutter should be tested in the position in which it is to be used. Aerial camera shutters should be tested in the horizontal position.

Types of Focal-plane Shutters.—A variety of means have been utilized for securing the necessary variation in speed in focal-plane shutters. Their success is to be measured by the actual speed range and by the uniformity of speed attained. In aerial cameras at present in use we find variable tension of the curtain spring, the aperture being fixed; variable opening with fixed tension; multiple curtain openings with fixed spring tension; and combinations of two or all of these methods of speed control. The problem of covering the aperture during the operation of winding up or setting the shutter has led to further elaborations of shutter mechanism. These take the form of lens or shutter flaps, auxiliary curtains, and shutters of the self-capping type. Shutters embodying all these features are briefly described below.

Representative Shutters.—The Folmer variable tension shutter is used on the United States Air Service hand-held and hand-operated plate camera and on some of the film cameras. It consists of a fixed aperture curtain wound on a curtain roller in which the spring can be set to various tensions, numbered 1 to 10. The range of speeds attainable is at best about three to one, or from 1
100 to 1
300 second, considerably shorter than the range indicated as desirable. Its uniformity of travel is variable with the tension, as shown by representative performance curves in Fig. 30. Lacking any self-capping feature the shutter is provided either with an auxiliary curtain, or in the hand-held camera with flaps in front of the lens, opened by the exposing lever before the curtain is released (Fig. 39). This shutter is made a removable unit in the 18 × 24 centimeter hand-operated camera, but is built into the hand-held and film cameras.

The Ica shutter used on the standard German aerial cameras is a good example of the multiple slit curtain (Fig. 26). Four fixed aperture slits are provided, with a single tension, the openings roughly in the ratio 1, ½, ¼, ⅛, which when the spring tension is properly adjusted give exposures of 1
90, 1
180, 1
375, 1
750 second. To pass from one exposure time to another the setting milled head is wound up to successively higher steps or else exposed one or more times without resetting, depending on the direction it is desired to go. Capping during setting, or during exposure, in order to change the opening, is provided for by a pair of flaps on the shutter unit, which open into the camera body. The mechanical work on these shutters is of excellent quality, the curtain running with exceptional smoothness. Provision is made for adjusting the tension until the marked speeds are attained; this is presumably done in a repair laboratory to which the shutter only need be sent, as it is a removable unit. Tests made on one of these shutters wound to its highest tension are shown in Fig. 30. The marked speeds are not attained, and there is considerable lack of uniformity from start to finish of the travel.

L camera variable-aperture shutter. The shutter of the L type camera (Fig. 27) is representative of one of the most primitive methods of varying aperture. The two jaws of the slit are held together by a long cord passing completely around the aperture, fastened permanently at one end and attached at its other end by a sliding clasp or saddle. As this saddle is forced in one direction the slit is closed, in the opposite direction the cord becomes slack, and after the shutter is released once or twice the slit assumes a wider opening. A chronic trouble is the breaking of the cords. Its opening can be changed only after the plate magazine is removed.

U. S. Air Service variable-aperture shutter. This shutter is incorporated in the American deRam and in other late American cameras (Fig. 28). Its characteristic feature is the introduction of an idler, whose distance from the main curtain roller can be varied. Tapes whereby the following curtain is attached to the spring roller pass over this idler, and by changing its position the aperture or distance between the two curtain elements is altered over a large range. Tests of this shutter are shown in Fig. 30. A speed of 1
50 second is provided for by a slit width of five centimeters, and the highest speed is fixed only by the practical limit of approach of the jaws. Experiment shows great uniformity of rate of travel to be attainable by combining careful choice of spring length and tension with good workmanship in the mechanical features. Variable-aperture fixed-tension shutters have a definite advantage over the variable-tension type in that they can utilize for all speeds that tension which gives uniform action. The capping feature of this shutter is provided in the American deRam by flaps, in the automatic film camera by an auxiliary curtain. The shutter is removable in the deRam, but built into the other camera.

The Klopcic variable-tension, variable-aperture, self-capping shutter is an example of an attempt to meet all shutter requirements with an entirely self-contained mechanism. It is shown diagrammatically in Fig. 29. Tapes G₁, G₂ are used to connect the following curtain B directly to the spring roller T, at a fixed distance, while the leading curtain, A, may be slid along the tapes by small friction buckles, C₁, C₂, auxiliary springs R₁, R₂ serving to keep it taut in any position. When the shutter is being set the buckles are arrested against stops while the winding-up continues for what is to be the following half of the curtain in exposing. When released the curtain moves across with an aperture fixed by the point of setting of the buckle stops. At the end of the travel the buckles are arrested by other stops, while the following portion of the curtain continues its travel to the end. On re-winding, therefore, the aperture is closed. Variable tension as well as variable aperture is provided, although little used. In the French cameras a lens flap is also inserted behind the lens, but this is not needed if the self-capping feature functions properly. On the hand cameras this flap is said to be necessary in order to prevent a curious kind of accident: if the camera is held on the knee, pointing upward, an image of the sun may be formed on the curtain and burn a hole through it.

The performance of the French shutter in respect to uniformity has already been shown in Fig. 24. It leaves very much to be desired. Besides non-uniformity of action during its travel it exhibits another common defect of variable-tension shutters, namely, the curtain must be released several times after a change of tension before the new speed is established (Fig. 30, tensions 5 and 5´).

The French shutter as made for the deMaria cameras is a removable unit. The small size (13 × 18 cm.) sets by the straight pull of a projecting pin, the larger (18 × 24 cm.) by winding up a milled head. The former is the more convenient motion for an aerial camera. Care must be taken with either type that the motion of setting is not stopped when the first resistance is encountered; this occurs when the tape buckles strike their stop and the slit begins to open.

In the earlier days of airplane photography the ordinary plate-holder or double dark slide was used to some extent, but it is ill-suited to the purpose because of the considerable time and attention required for its operation. It has nevertheless the merit of adding little to the length of the camera, and it works in any position. For these reasons it has remained in occasional use for the taking of oblique views with long focus cameras in a cramped fuselage.

Next in order of progress rank the simple box magazines, for holding a dozen, eighteen or twenty-four plates, as used in the English C, E, and L type cameras. These are little more than boxes with sliding lids which when open permit the introduction or removal of the plates. Figs. 45 and 46 illustrate the magazine of this type as made for the English C and E cameras. It is constructed of wood, grooved to fit tracks on the camera, and is furnished with a sliding door or lid hinged in the middle to fold down out of the way when open. The eighteen plates are carried in metal sheaths, both to provide opaque screens between them, and to protect them from injury in the mechanism of the camera. Fig. 27 shows the all-metal magazine made for the American model L camera. This differs from the English in material of construction, plate capacity (24 instead of 18) and manner of operating the slide, which is built up of three thicknesses of phosphor bronze and draws out through metal guides bent into semicircular form. A snap catch holds this slide at either end of its travel. The leather strap introduced in the American model for carrying and handling is a distinct improvement. These magazines contain no springs or other mechanism, as the cameras with which they are used depend upon the action of gravity for emptying the upper (feeding) magazine, and filling the lower (receiving) one.

Next in order of complexity may be ranked the bag magazine (Figs. 31 and 44). In this the exposed plate is pulled out of the magazine proper by a metal slide or rod into a leather bag. The rod is then pushed back, the plate in its metal sheath is grasped through the leather bag, lifted to the back of the magazine, and forced in behind the other plates. The number of plates exposed is indicated either by numbers on the backs of the sheaths, visible through a red glazed opening in the back, or else by a counter actuated by the metal slide rod. Usually twelve are carried in a magazine. For aerial work the common design of this magazine as used for ground work must be modified by providing extra large easily grasped hooks both on the draw rod and on the dark slide, which must be drawn before making the first exposure and replaced after the last. The small rings and grips of the standard commercial magazine are almost impossible to handle through heavy gloves.

The next type of magazine is represented by three designs, the Gaumont and deMaria, used very generally by the French during the war, and the Ernemann, used almost universally in the German air service (Figs. 32, 40 and 42). In all of these the operation of plate changing is the same: the end of the magazine is pulled out and thrust back, a more simple operation than the bag manipulation just described. The internal workings are different according to size. In the smaller French magazines (13 × 18 cm.) the camera is first pointed upward, all the plates are drawn out except the one to be changed, and this, with the aid of springs, drops to the bottom, after which the other plates push back over it. The plates pull out in the direction of their long dimension. In the larger French magazine (18 × 24 cm.) only the exposed plate pulls out. The pull is in the direction of the shorter dimension of the plate, which is lifted up by heavy springs and slides back over the top of the pile. In the Ernemann magazine only six plates are carried, which there is good reason to believe represent the maximum feasible number, judging by the reports of jambs and breakages in the twelve-plate French magazines. In all of these magazines laminated wood slides pull out and in at each operation, and while satisfactory if made and operated in one climate, experience indicates that if made in America and sent abroad swelling of the wood may be expected to prevent their successful operation.

Alternative forms of magazine, somewhat more practical from the standpoint of manufacture and export, are several designs embodying two compartments (Fig. 32). In the most simple of these the plates are moved, immediately before or after exposing, from the unexposed to the exposed side. Illustrative of this type are the Folmer designs, in which the to-and-fro motion is imparted by a rack geared to a pinion actuated either by a lever, in the hand camera, or by the power drive, in the automatic design (Figs. 33 and 53). Another illustration is afforded by the Piserini and Mondini magazine, in which the operation of changing is performed by a back-and-forth motion of a hand-grip, which also sets the camera shutter (Fig. 47).