Many of the individual model designs needed to be duplicated several times for use in separate camera set-ups or shot showing tight formation. Also, duplicate ships were needed in shots which required a miniature to explode.
In view of the duplication necessary, we set about determining what methods would be best used to produce the individual pieces. We considered vacuum forming, injection molding, and silastic glove molds. Eventually we used all of these methods to speed miniature production. The parts for each duplicate ship were also produced in a variety of materials; foam with a surface coat was chosen for those ships which were to be exploded. Foam was chosen for this application because it is low-density and shatters easily. These characteristics were needed to make our explosion shots more realistic. The shattering nature of the foam allowed us to use much smaller, slower-burning explosive charge, and the low density of the pieces caused them to move more slowly, both contributing to the scale of the explosion.
With this brief and generalized description of each of our main areas, perhaps their interaction could most understandably be outlined by following an individual shot from conception to completion.
A TYPICAL SHOT
First, the motion, size, position in frame, and frame counts are determined by examining the black and white WWII DOG FIGHT footage. Our example will be a Mustang moving from left to right and away while we pan with him as he rolls and begins to dive. A Zero enters frame lower right in pursuit, firing its guns at the Mustang. Our camera position tracks with them as they both cross the horizon and move down toward the surface with the camera, platform/viewers point of view on their tails. This black and white footage is now translated into a storyboard drawing describing the actions of the WWII aircraft in terms of our miniatures. The Mustang will be replaced by an X-wing fighter, and the Zero will be replaced with a T.I.E. fighter, firing lasers instead of bullets. The light sky in the black and white will become stars against black and the grey horizon in the black and white will become the Death Star surface.
The X-wing miniature is mounted from the nose mount so that it can roll about its own axis, yaw about its own axis, and crab to the side—in an axis perpendicular to the camera track motion. The camera, with its eyepiece mounted, is placed so that the ship is in the position indicated for the head of the shot.
Now the camera/subject motion programming begins. the dominant axis of motion is programmed first. In this case, the ship's motion away from camera and down in frame is established. With this portion of the program running, the "following pan" is put in. In this case, rotation of the ship on its vertical central axis will give the viewer a perspective change indicative of a pan, although the camera itself may not pan at all. With the crab and yaw motions on the ship running, and the track and tilt motions on the camera running, the ship's roll motion can be programmed to give the camera platform a feeling of motion.
All of the other motions of both camera and ship will be run while adding some roll to the camera, thereby giving the viewer the feeling that the camera platform is banking in order to follow the ship being photographed. These programming functions are being performed with the motion control system operating at approximately one-half real time— meaning that the motions seen through the eyepiece are at one-half the speed that those same motions will appear on the screen. The system can operate in a real time mode for programming, but for rapid subject motion the drive systems simply cannot move the camera at the speeds necessary for real time. I might add that even if the drive systems could operate that fast, it would be particularly dangerous, since if this shot were programmed in real time, it would require that the camera dolly and boom, move 40 feet from a standing start and come to a stop in just under six seconds.
Now that we have our move complete, the digital information (that is, the map of distances, accelerations, and speeds for that move), is transferred to a cassette for future use. We use frames-per-second as our increment of measure in the control system.
Example: The system running half real time is running on a 12-frame-per-second time base. We next reduce our time base from 12-frames-per-second to one-frame-per-second. This is really an exposure time of one-second-per-frame, but because pulldown of the film in this system is independent of exposure, the calculator automatically figures in the time required for pulldown, and compensates motion speed appropriately.
Because of this isolation of exposure and film advance and the long exposure times we are using to allow depth of field control, we have available shutter durations as long as 340. This long-duration shutter, though ideal for elimination strobing in extreme subject or camera displacement, must be used judiciously.
With our system, motion of camera and subject continues during exposure, just as in live-action photography. This is where the 300 shutter creates a problem. When the ship being photographed becomes small and its motion-per-frame is greater than a quarter of its overall size, it becomes a smear. In this situation, there is a happy medium of shutter-angle setting which makes the ship image apparently sharp and reduces strobing, minimizing that annoying syndrome as much as is possible. Once our shutter angle is chosen, we can deal with focus. The follow-focus device is built into the camera, and includes a tilting lens board for tailoring depth of field on tight shots. It runs off the same time base as the motion control motors. It can be initialized (establishing base point for focus calculations) and will change directly with the camera moves on the track. Or it can be individually programmed in cases where the camera subject distance is determined by something other than the camera track (if the camera boom is used or the subject is mounted on another track which moves it nearer to or farther from the camera during the shot).
The blue backing is now covered to expose just the area around the X-wing. This is done to minimize blue reflections on the miniature. This leaves a lot of extraneous equipment in the shot: lamp stands, gobos, etc. These will be eliminated later by garbage mattes made in our rotoscope department. With the lighting set, focus and move programmed, we shoot this X-wing element on black and white negative material and process it in our black and white lab. This serves several purposes, it can be viewed immediately to check move, lighting, focus, etc. Its other purpose will be apparent later.
Because the X-wing is seen from the rear in this element, the photography on final color negative will be made in two passes—one to record the X-wing miniature with its key/fill light, and blue backing, and a second pass on a separate piece of negative to record the X-wing's engine light effect. The reason for this double pass is twofold. The built-in practical lighting, in this case the engines, is not bright enough to balance in exposure against the X-wing body and wings. The second pass allows us to increase the exposure and add some filtration to enhance the color and the flare of the engine light effect. Also, when using the blue screen matting system, it is very difficult to process a good fitting matte for the engine flare when the element must be matted over a grey value background. This double-pass approach on a complex move such as this is only possible because of our system's unique electronic and mechanical precision repetition of camera/subject position on a frame-to-frame basis.
We have now generated three pieces of film which exactly duplicate the motion of the X-wing miniature and its engine effects frame-by-frame. The black and white negative, which we have already processed, gives us a black X-wing image on a clear cell background. The color negative shows the X-wing and blue backing, and the third negative has the X-wing engine against a black field. A camera report is then made up for the processing of the color negative elements and set aside for delivery to the laboratory.
We now prepare the T.I.E. ship miniature for photography. Normally, when one of the ships in a shot is firing lasers, the miniature has a laser emanation point provided in original photography. This point is usually a small light at the tip of the ship's laser cannon which lasts for one frame when a laser fires. In this case, however, the lasers are fired as the T.I.E. ship is going away from the camera. Hence, no laser emanation point will be required on this shot. The T.I.E. ship is mounted and we begin programming the second element of this shot. Because our camera platform, or viewer's frame of reference, is required to make some moves in this shot, we will use a portion of the program that we used to photograph the X-wing. We load the digital information off the storage cassette into solid state memory in the control device, and we are ready to proceed. We are describing with the T.I.E. ship basically the same move that the X-wing performed, but it will enter a frame later in the shot than the X-wing because it is to appear that the T.I.E. ship is pursuing the X-wing. The only axes of motion that we must use from the previous program are the camera nodal point, roll, pitch, and yaw. This is necessary to maintain the camera platform/viewer frame of reference integrity. This motion will match the camera motion of the X-wing element on a frame-for-frame basis, so that no matter what position or independent motion we give the ship, the two shots will have the same angular displacement (nodal point camera motions) and, therefore, give the appearance that the camera platform/viewer frame of reference rotates and pans independent of the individual ship motions.
We will change the track motion, the crab motion, and the boom vertical motion slightly to give the T.I.E. ship's move a subtly different character. We will move the track position in tighter to make the T.I.E. ship larger. We will reposition the crab move to the left to have the T.I.E. ship stay behind the X-wing and we will boom down to position the T.I.E. ship higher in frame. The model crab, rotation, and yaw will remain the same to give the appearance of tracking the same moves that the X-wing is making. We now generate the same film elements on the T.I.E. ship that we generated on the X-wing with the exception of the engine element. The T.I.E. engine effect is provided by LE.D.'s (Light Emitting Diodes), which record in the original key, and blue backing element photography.
First we photograph the T.I.E. ship move on the black and white negative material and process it. At this point we have another use for the black and white negative element which we made of the X-wing's move. We can "bi-pack" (lay together), the black and white negatives of the X-wing and the T.I.E. ship in a registration viewer. Because the images of the ships are black, and the backgrounds are clear, we can see the motions of the ships relative to each other and the composition of the shot at projection speed. Having viewed this test, we can now modify the T.I.E. ship move if necessary, or go ahead and photograph the T.I.E. ship move on color film.
Once the T.I.E. ship is photographed on color negative, we move on to the background in this case, stars and black sky—and the surface of our mechanical planet, the "Death Star". This shot requires us to dive on the surface from high altitude in pursuit of the X-wing and T.I.E. ship and pull out just above the surface. We have four scales of Death Star miniature to work with and, because of the extreme perspective change we have to go through to follow the X-wing and T.I.E. ship realistically, we must use the second largest scale. This scale has the relief necessary to look realistic in the move, and is small enough in scale to provide the high-altitude to low-altitude diving portion of the shot without requiring a very great subject-to-camera distance in order to achieve the proper size change.
[ continued on page 3 ]