Challenges of Stereoscopic Motion Picture Photography
Hoch, Winton C, American Cinematographer
A great deal has been written about stereoscopic photography. This most fascinating field seems to be periodically rediscovered. The question is why? Since stereo photography is the subject, it is inevitably tied into the technology of-and interest in-photography and stereoscopy by individuals and/or groups, generally for either pleasure, science or business, or any combination of these. Money, technology and interest are essential ingredients.
Currently we are presented with a breakthrough in 3-D "photography" in the form of holograms. This development is, in many opinions (and also in mine), one of the more fascinating current scientific developments. Holograms are only a very small part of a laser technology that is developing with great strides. If we accept their limitations in "photography" or recording of the images, and the limitations of viewing, they have no equal at present. This is a superb technique!
The limitations require close examination. Holography uses very expensive, very sophisticated equipment and imposes severe restrictions on scope, subject matter, lighting techniques, etc. For instance lasers of any appreciable power necessary for even medium-size areas are hazardous.
Many intriguing demonstrations are available and many applications are of great value. However, the many inherent restrictions make it essentially unattainable and highly impractical for general motion picture applications.
In brief, holography appears not at all feasible for the type of motion picture applications that form more than 99% of our current operational pattern.
Accepting this conclusion, let us consider more flexible and much less restrictive depth-portraying techniques.
Depth cues are conveyed to the mind in several ways. A "flat" (non-stereo) two-dimensional picture can present a feeling of depth by several well-known techniques, such as: perspective, scaling, use of haze, etc. Motion picture photography has the added inherent advantage of a moving viewpoint when the camera is moved as in "dolly" shots and "running" shots. If there is no stationary foreground, these exceptional shots can, on occasion, approach stereoscopic viewing. The primary principle can be easily verified. Sit in a room with stationary objects at various distancesthe objects preferably overlapping their outlines and at least one close to you. Close one eye. The scene is "flat". All sense of depth is lost except that in your memory. Keep one eye closed and move your head side-ways. Immediately you have evaluated the depth of the scene and the distance of the various objects! A motion picture "running" shot gives similar information, such as a side-angle running shot or a forwardlooking shot (with a wide-angle lens) and projected onto a curugd, theatre screen. The roller coaster ride in "THIS IS CINERAMA" is an outstanding example.
While these exceptional flat" scenes can be effective in presenting depth on occasion, they cannot truly compete with bona fide stereo presentations. The only time I have ever seen an audience dodge a missile coming out of the screen was in a stereo presentation. Stereoscopy does have an overwhelming noncompetitive audience participation capability!
Therefore, accepting this stereo approach and using the time-tested tools of photography (or any other practical image recording technology), we are faced with the basic problem of taking the picture to the audience or of taking the audience to the picture. This is the classical premise.
If we take the picture to the audience we must photograph, record or otherwise "capture" a multiplicity of viewpoints with the effect that each member of the audience sees the picture oriented to his viewing position in the audience. This infers not only laterally displaced viewpoints, but also vertically displaced viewpoints. Let us assume a simplified version of just laterally displaced viewpoints. Even this simplified approach requires a multiplicity of images for both recording and reproducing (or photography and projection).
The possible techniques have been pretty well explored. One technique involves photography through the film base which is embossed with extremely small cylindrical lenses that allow the camera film to capture a multiplicity of viewpoints by either swinging the camera about a predetermined pivot point or by laterally moving the camera a predetermined amount. Another timehonored approach is to simply take a multiplicity of pictures, each picture on a separate film or film area and each picture photographed from a different but predetermined place.
Presenting these pictures to a viewing audience involves a multiplicity of problems and expenses. Multiplicity is the name of the game and would appear highly impractical for motion picture presentation. The added motion picture requirement of simultaneous exposure of stereo images at a rate of 24 pictures per second (more or less) appears to put the multiplicity of images approach in the indefinite future.
The second approach requires only two images. A separate viewpoint for each eye. This approach utilizes a much less complicated and a more economical technology for both recording and viewing.
A great deal has been written about, and much effort spent toward developing, satisfactory two-image systems. Regrettably, they have not been truly compatible with motion picture production and exhibition hardware, techniques and practices.
Probably the biggest problem with past 3-D motion picture systems has been poor basic design of the system and the violation of proper techniques in photography and/or exhibition. The most serious abuses appeared to be in exhibition. In the failure to properly present the images for viewing. The resultant discomfort was erroneously blamed on the viewing "glasses".
Discipline is essential in both photography and exhibition.
An inherent challenge existed to devise a single-camera, single-film, singleprojector system compatible with current production and exhibition hardware and techniques. A system that is versatile, essentially foolproof, easy to use and free from eye strain.
Toward this goal some efforts have been directed to a single-lens, doubleimage system. Although a single camera lens apparently contradicts the premise of two viewpoints this is not the case. For example, any lens of appreciable aperture diameter gathers light from viewpoints within that aperture. Therefore, if one half of the aperture is covered with a green filter and the other half is covered with the complementary magenta filter (the dividing line being vertical), green and magenta images can be said to be simultaneously recorded and superimposed (on color film, of course). When the positive picture is viewed with the complementary green and magenta filters a stereoscopic picture is observed. The color picture, when viewed without filters, looks like a normal color photo except that objects in front of and behind the plane of focus show color fringing due to the green and magenta filters on the camera lens. Any two properly balanced complementary color filters may be used. However, the same filters should, of course, be used for viewing-filters of the same colors and proper densities. This system of using the color filters has a handicap in the .unnatural viewing, due to each eye being filtered with a separate color. This system and approach has many serious restrictions, but for some limited applications it can be of interest. This paragraph does not pretend to be at all exhaustive regarding this approach.
Other "systems" have been presented claiming that 3-D viewing can be obtained from any regular "flat" movie. There is just enough evidence in this claim to be provocative. This 3-D "effect" can be achieved in one way by projecting the "flat" image as an aerial image hanging in space in front of the screen. There are techniques for doing this either with or without auxiliary viewing devices (such as "glasses"). It follows that while the image is "hanging in space" there is no roundness or stereo perception of the individual elements in the picture. It is a flat picture,
Another method that has been somelivhat commercialized uses two frames of Jhe ,picture simultaneously projected and viewed by the polarizing technique. Complementary filter techniques or prism glasses techniques could also be used. This method relies on lateral movement of objects in the frame. It utilizes a basic principle of binocular vision (for stereo perception); that is, the left eye and the right eye must have appropriate, displaced viewpoints. Therefore, each and every individual object that moves laterally in the proper direction will appear in 3-D. The interpupillary distance is, of course, dependent on the distance the object has traveled during the time the two viewpoints were recorded. Obviously, if the camera dollies laterally, all objects will allow stereo viewing, unless some object or objects cancel this movement or reverse it to an improper direction, which will result in pseudoscopic viewing of the object or objects. A lateral dolly in the improper direction could create pseudoscopic stereo in essentially all objects in the frame. This can be very disturbing.
It is axiomatic that all action is at least one frame out of synchronization and interocular distances are a complete gamble!
Extracting stereo viewing from a flat picture is most unsatisfactory. This is an understatement.
A two-image system that does honestly meet the many challenges of stereoscopy is being presented by Stereoscopic Productions Inc. It is a single-camera, single-film strip, single-projector system that is simple to use, possesses great flexibility and exceptional versatility, is essentially foolproof and free from eyestrain. It solves the many problems of stereoscopy in a simple and economical manner-and is compatible with current production and exhibition hardware and techniques.
This two-image system (the two images being simultaneously exposed) places one image above the other, with the image centers displaced vertically apart so that their center distances are further apart than the center distances of the identical two image film areas available. see FIGURES 1, 2, 3, 4, 5 and 6. This results in a wide space or "bar" between corresponding stereo image pairs.
FIGURES 7 and 8 show how the "bar" is used when the corresponding stereo pairs are separated by other stereo pairs-a very useful arrangement for some applications.
With the formats suggested by FIGURES 7 & 8, the left and right stereo image centers may be vertically offset, either further apart or closer than the image centers of the available film space.
This space or "bar" is extremely useful in both photography and projection. In projection it is most valuable in helping to separate the two projected images when they are "adjacent" to each other and in preventing the "bleeding" of each image into the other on the screen, with a minimum of vignetting. It also gives immediate image identification on negatives, positives, dupes, etc. and is an obvious splicing area, thereby avoiding accidental image transposition, which is very disturbing to a viewing audience. In addition, it makes any misframing of the film in the projector immediately obvious, since the two incorrect images will be seriously misaligned vertically. Such an error is immediately obvious. Not only will a frame line be projected, but the vertical misalignment will make stereo fusion essentially impossible! The projectionist can immediately correct this error by reframing the picture, which is a very simple procedure.
This offset image placement is a patented feature.
Currently used 35mm professional motion picture cameras that are provied with full apertures or Cinemascope apetures can be used without modification. This includes studio relfex cameras and hand-held cameras.
Inter-ocular control, stero window's control, focus control, exposure control (continuous with an iris), and a selection of focal lengths is possible. A small matte box is desirable, but not essential.
The relfex cameras presented quiter a formidable challenge to the concept of a 3-D attachemnt that would fit onto aslready existing lens mounts-and require no other modifications to the camera. Since reflex cameras are so very popular, we had no choice but to accept the problem.
We have met this challenge with a design that is applicable to all film sizes. We have a prototype for the 35mm professional camera field. As previously stated, a full camera aperture or a Cinemascope camera aperture is most desirable. The design of the attachment is such that a matte box is desirable, but not essential. The attachment has been used very successfully without a matte box or shading of any kind.
A "double pulldown" 70mm format should have incomparable impact! It appears attainable and practical, especially for specialty presentations. Either an eight-sprocket or ten-sprocket pulldown is very attractive. Currently a 70mm eight-sprocket format is being screen tested by Film Effects of Hollywood in cooperation with Stereoscopic Productions, Inc.
The film areas available sometimes invite the use of anarhorphic lenses. A vertical squeeze appears attractive in some situations. For example, a 2-to-1 vertical squeeze ratio used with "the standard 5-sprocket 65-70mm cameras and projectors presents the same heightto-width effective aperture ratio as would special 10-sprocket 65-70mm cameras and projectors. The film areas available indicate a wide screen aspect ratio of approximately 2.38-to-1.
The vertical squeeze is also of interest in some uses of the 35mm, 16mm, Super 8mm formats, etc. A horizontal squeeze can be of interest on occasion.
Various film area formats and corresponding stereo pair placements are indicated in FIGURES 1 through 8. Those illustrated obviously do not limit the potential arrangements. Transposition of these arrangements within all film sizes may, on occasion, be desirable.
A standardized offset of the images makes optical printing an attractive facility. For example, shoot in 35mm and "blow up" to 70mm or reduce to 16mm and/or Super 8mm. Shoot in 16mm and blow up to 35mm or reduce to Super 8mm, etc. This is a very versatile system.
The Eyemo camera pictured has been modified to one form of Stereoscopic Productions Inc. stereo formats and has been a very useful camera. The effective focal length of the lenses of the camera shown is approximately 22mm. Rhomboid prisms are used, as shown in FIGURES 13 and 14, to place the images in the format of FIGURE 15. This design has patented features.
Superimposing of these two stereo images in projection has used various formats. Five of them are illustrated as follows:
1. A separate lens for each aperture or simultaneously projected image. see FIGURES 9-a and 9-b.
2. A single projection lens for both apertures (images), and recombining them on the screen with the use of mirrors. see FIGURES 10-a and 10-b.
3. A single projection lens for both apertures (images), and recombining them on the screen with prisms. see FIGURE 12.
Each image is, of course, filtered to the effect that the left eye sees only the left-eye image and the right eye only the right-eye image, i believe that filter combinations that have been used are polarizing filters, red and green filters, and magenta and green filters.
4. A useful but little-used viewing technique for motion picture presentations is the use of prism "glasses" on the viewer. This is an old technique in use for years in the parlor 3-D viewers of still pictures so popular many years ago. Using this prism viewing technique the two images can be projected either side by side or one above the other. The prisms can be adjustable and rotatable to accommodate different viewing distances.
5. Zonal viewing systems for viewing without glasses have been used and are being further explored. A restricted head position (laterally) is required. Other requirements and restrictions are serious handicaps to this type of viewing.
In all of these projection formats it is most desirable to follow general industry practice and provide a matte (in the projector) for each image. This matte is the final control of the image size and it also gives clean edges to the projected images. For double-image 3-D projection it is obvious that one matte slide can be used with two matched and separated aperture openings. In some situations, an alternate method, such as a proper black matte printed on the film, would be satisfactory.
In connection with large screen presentations, the film area available for each image determines the maximum screen size. The limiting factors are either the image magnification and resultant acceptable screen definition or the amount of light that can be transmitted by the optical system, assuming, of course, an optimum light source and that the picture areas are restrictive apertures.
The 35mm technology is probably the most highly developed, and current equipment (modern 35mm theatre projectors) indicates a potential screen width with superimposed stereo images, of approximately thirty feet. A 70mm format with the same image presentation indicates a potential screen width of approximately 60 to 70 feet. The above sizes have allowed for the recommended theatre screen brightness standards, as measured through all necessary filters and assuming a screen gain factor of approximately four-which is customary on many non-depolarizing screens.
There are many ramifications to the two-image approach to stereoscopy. I would like to believe that I have presented the more prominent and/or practical of the approaches that are within the limits of our current practical technology.
Simplicity and economy are powerful factors. I believe that this simple and economical two-image system of Stereoscopic Productions Inc. is superb in its ability to meet the challenges of motion picture stereoscopy.
By WINTON C. HOCH, ASC…
Questia, a part of Gale, Cengage Learning. www.questia.com
Publication information: Article title: Challenges of Stereoscopic Motion Picture Photography. Contributors: Hoch, Winton C - Author. Magazine title: American Cinematographer. Volume: 55. Issue: 4 Publication date: April 1974. Page number: 426+. © American Society of Cinematographers Feb 2009. Provided by ProQuest LLC. All Rights Reserved.
This material is protected by copyright and, with the exception of fair use, may not be further copied, distributed or transmitted in any form or by any means.