Full text

It was felt from the earliest conceptions of "TO FLY" that the film would be incomplete without a sequence depicting not only our present ventures into space, but portents of the future, as well. The whole motion picture represents an outward expansion from Early America and that motion had to continue through the rest of the film.

The space sequence was originally designed as a series of stills showing views of our solar system, galaxy and the universe. None of us ever thought that would really be acceptable and in our earliest pre-planning we decided that money saved on the other sequences would be earmarked for a more ambitious space sequence.

Unfortunately it is not possible to make a large-screen film on space without automatic comparisons to "2001". Needless to say we didn't have the millions of dollars necessary to equal or surpass "2001", but we still had the comparisons to live with, unfair or not. The volumes of material written about the production of that picture were a starting point in our research. We found that there were many things they had worked out by trial and error that we could incorporate or expand upon. For example, we were forewarned that our most difficult task would be the filming of starfields that looked correct both in 70mm IMAX and later in 35mm release. Many times our 35mm printdowns showed no stars whatsoever, while the corresponding IMAX negative looked normal. When the stars looked normal in our 35mm print-downs they looked like basketballs in IMAX. Obviously the answer was a compromise of very narrow limitations. This was a problem that plagued us throughout the filming of the sequence.

(LEFT) Associate producer Jeff Blyth, shown with the 'Starship', which was designed, then approved by the Smithsonian Institution, then constructed by MacGillivray/Freeman Films from many model kits. (RIGHT) Blyth adjusts the bellowed closeup lens on the IMAX camera. The geared Worrall head was used for smooth motorized pans and tilts.

The other major thing we learned from "2001" was the use of movement. Not only is there constant movement of spacecraft in that film, but the background starfields are constantly moving as well.

We had some very severe limitations in the manner in which we could shoot the space sequence. The most conventional full-screen optical effects are not possible in IMAX, not even acceptable dupes. The camera would not single-frame or run in reverse which meant we couldn't use stopmotion photography of models or registered mattes, etc. All of our effects had to be accomplished on the original camera negative. We could, however, reliably film as slow as 3 fps., and we shot most of the scenes at 6 and 12 fps.

Diagram of the 'Discount Slit-scan' rig used to create outer space effects for 'TO FLY'. Individual pieces of artwork photographed on Kodalith were matched to create an 8-inch by 30foot piece of art. Using an overhead projector with a specially designed track, and motorized mechanism, the artwork was pulled very smoothly across the glass surface of the projector. The images were projected onto a very large reflective white card. The IMAX camera, with fisheye lens, was set close to the edge of the card and perpendicular to the axis of the projector.

Another serious problem was light: we couldn't get enough of it. We lost one stop with our pellicle viewfinder. Our close-up lens was a speedy f/5.6 and occasionally it was used with a polarizer and a bellows extension. We couldn't force-develop our negative because we were very concerned with maintaining consistent black densities. On top of that, we sometimes, in the case of large planetary models, had to film at T/16 for enough depth of field.

These factors influenced the storyboard greatly. Each shot was designed in terms of how it could be photographed in IMAX. That very rarely meant the best or most convenient way to photograph it. Each scene was detailed very carefully in the storyboard and to insure that we were communicating our exact intentions to the National Air and Space Museum, each shot was broken down into two-second intervals. It showed the full progression of elements during the shot. We used the new color Xerox machines to quickly make each frame from elements that were painted on clear eels and backgrounds.

Our shots broke down into three general categories. There were a number of scenes that required studio facilities and planetary models. Another group could be filmed at our offices on a home-made animation stand using 2-dimensional artwork. The rest of the shots were of an abstract nature and each required its own particular conditions (see accompanying article on laser photography).

Actual full-size IMAX frames from the space sequences of 'TO FLY'. (ABOVE LEFT) In the newly designed 6-panel optical formation, the protective ring of the 2nd stage engines on the Saturn rocket falls back to earth. The scene was photographed by a 16mm camera actually mounted on the engine stage. (CENTER) The discovery ship explores Jupiter and its moon, lo. The scene was accomplished with models photographed in four separate camera passes. (RIGHT) The ship passes in front of Saturn on its way to the outer nebulae and galaxies. Photographing this scene required both front and back-lighting. (BELOW LEFT) To simulate nebulae and exploding stars, special combinations of paints and chemicals were used. (CENTER) The spacecraft, built specially for the film, was designed as a long-distance 'discovery' or scientific explorer craft. (RIGHT) One of the re-created weird and wonderful sights of outer space.

Fortunately for us, we were able to work with George casey of Graphic Films for the studio portion of our filming. They were able to provide their excellent planetary models and facilities as well as their years of expertise in this kind of photography. While they had never filmed directly in the IMAX format they were able to adapt many of their techniques to our film.

The most difficult of the shots made with Graphic Films was Jupiter as seen travelling across the surface of lo, one of its moons. In the distance our starship passes between them headed for deep space. It required four separate passes of the negative through the camera, each time being rewound to heads and rethreaded to a common start frame, lo's surface was photographed first, moving slowly past the camera on tracks. Its upper limit of travel was then scribed on a piece of glass in our matte box. We then changed set-ups to film their Jupiter model, again scribing its boundaries on the piece of glass. The ship was a twoinch transparency (of a three-foot starship model) rear-lighted on a stationary glass frame. In this case it was the camera that moved on a motorized gear-driven dolly. Once more the limits of that component of the shot were scribed on our matte box and then we finally exposed a starfield that took into account the placement of each element.

While at Graphic Films we also filmed a sunrise as seen from space. We used their beautiful Earth model and motorized dolly tracks to create the effect that we were leaving Earth orbit from the dark side and moving far enough away to reveal the sun. We used a 5K with a snout for our sun, aimed directly at the camera lens. It wasn't bright enough by itself to create a rim of light around the Earth's rim so we employed a 1OK set at right angles to the Earth and it was goboed off to create a fairly well-defined rim. We worked very closely with Mike Collins, the former Apollo astronaut and now Director of the Air and Space Museum, for his opinions as to the effect of the sun in space. He should know.

We also filmed our starship model in a variety of passes to be used to unify the entire sequence. It serves almost the same function as our 747 earlier in the film and, in fact, I designed the model to have the same color scheme. It took about 150 hours to build from many model plane, tank and rocket kits. It was based on some fairly wellknown ideas about future forms of propulsion, such as advanced ion drive and magnetically-contained nuclear reactions, such as deuterium/helium-3. Mike Collins was able to provide many important inputs that affected design and construction and was a willing resource throughout the entire film.

As part of our arrangement with Graphic Films we were able to use their facilities to generate a number of still photographs that were then used in later scenes. We used their large Saturn model, for example, so that it could be filmed on our animation stand as a large transparency. This was partially due to the retouching required to remove traces of support for the model's rings. The storyboard showed the starship passing directly across the face of the planet, which meant that we had to use an animation technique.

Our stars were generated on highcontrast negative sheet film (Kodalith) on which we could easily matte clear areas for the planets. Saturn was a 24"wide transparency that was matted into a continuous 60" ? 24" kodalith starfield. The ship was a photographic print (and was therefore opaque) that was mounted on a large sheet of glass. Saturn, its starfield and the ship were filmed simultaneously, both top and bottom lit, so that the ship would be self-matting over the stars and the planet. It took considerable effort to balance the lighting for the planet transparency and the ship. All the top light was polarized (as was the camera lens) to eliminate surface reflections and maintain a black starfield. To make things more difficult the Saturn transparency and stars moved behind the ship. Had we been able to single-frame we could have filmed it as two separate and registered passes, one top lit, the other bottom lit. We had to make the 60" transparency/starfield move smoothly behind the large sheet of glass. We were working with a 30-field opal glass, so we had to contend with reflections over a wide area. The art moved so slowly during filming that the backlights caused it to buckle from the heat and it had to be replaced.

We acquired an Oxberry auxiliary peg bar which we motorized with special gears and a zoom motor and control. It took quite a bit of adjusting to get the peg bar to move smoothly on its original track, let alone the much longer one we needed for this shot. We spent many hours getting it to traverse the entire length of the stand without any bumps or jiggles. Saturn and the starfield were attached to the peg bar and the sheet of glass with the ship on it was suspended as close as possible to the surface of the starfield without interfering with its movement.

In order to make the Starship 'move', the camera was actually moving on a vibrationless geared track. Here Blyth adjusts the ship on its motorized, revolving pedestal, while technician from Graphic Films adjusts the dolly track motor. The two-minute space sequence in the film took four-and-a-half months to pre-plan, test and shoot.Zoom motor controls with specially modified gearing were used to make slow, smooth, repeatable moves on the animation stand. In this case, the moon photograph moves into the frame across the backlit background of the star field and a transparency of the Earth. Since the moves were precisely repeatable, tests could be made in 16mm before shooting the final scenes in IMAX.

We wanted to add some other movement to the shot, as well, so we mounted the camera on a Worrall geared head and replaced the pan wheel with a specially-built gear. This, too, was driven by a zoom motor and it added a slight pan to the shot. Because we were filming this particular scene at 5 fps and the camera was only panning a few inches on the artwork, the zoom motor and gears were barely turning during the shot. We had tried several motors and control devices for ultra smoothness and we were pleasantly surprised at how well the zoom motors worked even at such slow speeds. In a couple of shots we used them to control tilt movements, as well. Most important of all, the geared head and zoom motors gave us something we could repeat with accuracy. We could afford to test extensively in 16mm for speed of movement with the assurance we could duplicate the movements later in IMAX.

The one shot that we thought would be the most elusive was a Saturn rocket taking off. We had filmed the roll-out at Cape Canaveral of the Apollo-Soyez mission earlier in the year and then returned for the summer lift-off. We used two remotely-started cameras mounted only 1200' from the base of the rocket. What we had going against us was the failure of other film crews on three previous missions. Something always went wrong with the cameras, usually traced to the humidity and condensation that forms in the camera (and on the lens) because they are left outside for over 24 hours. In our case it was the worst time of the year (July) with monstrous thunderstorms rolling across the cape and humidity that had you soaked by 10 a.m. We built shelters for the cameras and had them enclosed in tight plastic bags with a worklight that was left on all the timein the hopes that it would reduce the relative humidity. We had an opportunity to place one of our cameras on the pad a day early so we thought it would be a good test. The next morning, and our last chance to work at the pad, we found the camera so wet inside that a piece of test film was decomposing in the gate. It was totally jammed, and took three hours to tear down and clean. We could do nothing but lubricate every surface and part that touched the film, reasoning that some high-speed cameras splash a little oil on film without causing too much damage. Quite frankly, we weren't too confident about our chances of filming the last Saturn rocket ever. The next day we watched the lift-off from the VAB, over 3 miles away. Hours later we were allowed back on the pad and found the cameras had both run perfectly.

We tried very hard to come up with some effects that would engulf the audience completely and came up with what we called our Discount-Slit-Scan. It cost about 1/1000th what it did in "2001" and was relatively easy to set up. It had absolutely nothing to do with slit-scan photography other than to look very much like the effect.

We created artwork on 8''x10'' sheets of white paper. The art deliberately contained grey areas designed to "break-up" when printed onto high-contrast film. We made Kodaliths of each piece and they were then carefully matched together to create an 8''x30-foot piece of art. Using an overhead projector with a specially designed track and motorized mechanism, the artwork was pulled very smoothly over the glass surface of the projector. The images were projected onto a very large reflective white card that faced the lens of the projector, and was set up about ten feet from it. Focus was maintained across the entire white card and the extreme density of the Kodalith film maintained solid blacks that would not "read" on the negative. The IMAX camera was set up with a fish-eye lens very close to one edge of the white card and perpendicular to the axis of the projector. The camera faced the far edge of the card, creating a "horizon". The images moved towards the camera and were deliberately distorted by the lens and the slight curvature we imparted to the white card. It gave the effect of very distant moving shapes that stretched out and exaggerated as they came closer and then rushed by the camera. This was done to create one plane, covering half the IMAX frame. We then flipped the camera over, rewound the film and repeated the effect for the other half of the image. The 30-foot-long artwork had gels of different colors attached directly to its top surface (so as not to interfere with its movements through the tracks on the projector). We experimented several times in 16mm to get the best effect of color and speed. We also had to make a giant rewind to take-up and pull all the artwork, and it, too, was driven by a zoom motor.

By far the most interesting of our effects involved what look like immense and colorful nebulae expanding and contracting in deep space. In actuality they ranged in size from a pack of cigarettes to a postage stamp. We filmed them using a bellows extension, a Hasselblad closeup lens and a variety of diopters. The camera was mounted on a very rigid stand pointing straight down, the fully extended lens about 14'' from the surface of a table top. The table and stand were very firmly braced to dampen all vibrations. An 8''x10'' plastic developing tray (although just about any flat container will work) was filled halfway with a measured amount of water. (This creates a constant focus because the surface is always the same distance from the film plane each time you set up). To the water we added a few squirts of jet black ink, mixing it in thoroughly. When the liquids settled to a smooth, ripple-free surface we added one drop of white enamel paint. It virtually explodes on contact, pulling itself apart and expanding rapidly into clusters and arms that break up into a million little white spots, all moving outward from the center.

We must have loaded and unloaded that tray of water/ink/paint 500 times and it never worked the same twice. That was the problem. There were many variables that affected the results, such as: the temperature of the water and paint, the relative temperature of the water and paint (cold seems to slow the reaction slightly), the size of the container, proximity of lights (we used two 1000's set up 2 ½ feet from the tray for an approximate f/8 stop), the thickness of the paint, the size of the drop of paint and, most importantly, the type of paint. No two white enamels reacted the same, even when they were the same brand. Even the reactions of one kind of paint might be different an hour later. We tried many different types and colors but ended up using only a few types of white which we could then color by adding gels to our lights. One interesting technique was to form a small wire into a shape, like a branding iron, and then dip it into the paint. Its shape had to be perfectly flat on the bottom but once it was carefully lowered into the ink/water it created some really beautiful patterns.

Our biggest headache, besides trying to repeat something for the cameras that we saw and liked while experimenting, was the speed of the reaction. It should be filmed at 96 fps. if not more. There are little motions created by currents within the expanding paint that you can't see with the naked eye but are picked up at 24 fps. Some of these currents add to the effect, but most don't. Since we couldn't film much above 32 fps. we wanted to slow everything down. We tried using mineral oil as our medium (the ink and some paints don't mix) and various other solutions, but the most promising (and the most surprising) was to add a layer of paint thinner to the surface of the water. It slowed down many of the reactions and created some interesting new ones. Some paints sank immediately to the interface layer between the water and thinner and began to drift and dissolve while other paints spread out across the top surface, leaving the bottom layer undisturbed and creating a 3dimensional effect. All the paints form a scummy residue almost immediately that starts to slow the reaction once it reaches the edge of the container. We found no way around that problem except trying paint that is "fuel-proof" for model airplanes. It forms such a thick residue that it can be picked right up off the water if you don't like the pattern. At least it saves the hassle of loading and unloading the container all the time.

The space sequence took 4 ½ months to pre-plan, test and shoot, all for a two-minute sequence at the end of the film. It would have taken a great deal more time and money had we not planned every phase of it from the beginning. It was a series of challenges above and beyond the problems encountered in the rest of the film.

NEW 2X1SIX-ELEMENT LENS EXTENDER AVAILABLE FROM A.G.E. INC.

A new, multi-coated, six element 2X lens extender designed to meet the exacting requirements of the motion picture industry is now available from Alan Gordon Enterprises Inc., according to Ted Lane, sales manager.

Lane says the extender is made with the finest optical glass available in order to obtain maximum detail and contrast. The six elements guarantee optimum resolution and the multicoating eliminates flare, reflections and ghost images while improving contrast and color saturation.

The 2X extender is available in the following lens mounts: Arriflex standard, Eclair CMS and Arriflex bayonet, which can be used with any Arri bayonet-mounted zoom lens and certain focusable prime lenses.

For additional information contact Lane at A.G.E. Inc., 1430 N. Cahuenga Blvd., Hollywood, Calif. 90028. Telephone is (213) 466-3561.