Interview by Don Shay
For more than a year, the international film community attentively followed reports of two rival James Bond films in simultaneous production — each slated for a summer of 1983 release, and each hell bent on striking gold at the boxoffice. Octopussy, thirteenth in the phenomenally successful series begun two decades ago by producer Albert R. Broccoli and his then associate Harry Saltzman, was the most easily pegged. Starring Roger Moore in his sixth outing as Ian Fleming's redoubtable 007, Octopussy could be expected to follow the series' tried-and-true formula of juxtaposing exotic locales, beautiful women, spectacular stunts and glib dialogue over a plot that — in many cases — seemed almost an afterthought. Less predictable was the direction the alternate production would take. Never Say Never Again, the first Bond film to be made outside the Eon Productions fold since the atypically spoofish Casino Royale (1967), was ostensibly a remake of Thunderball, the only Bond novel whose film rights were not owned by the Broccoli organization. Top-notch production values, of course, were mandatory, and producer Jack Schwartzman set about assembling a team of seasoned professionals headed by director Irvin Kershner. Schwartzman's major coup, however, came in his casting of James Bond. More than a decade after walking away from the role than had threatened to irrevocably typecast him, Sean Connery — still considered by many the quintessential Bond — would return once again to the role that had made him an international superstar.
Primed for a boxoffice showdown, the industry and the media followed each production with interest and anticipation. Ultimately, however, the expected confrontation failed to materialize. Shortly before its scheduled release, Warner Brothers announced that Never Say Never Again would be held over until the fall, in essence capitulating to Octopussy and sparking rumors that the maverick production was in trouble. As it turned out, nothing could have been further from the truth. Never Say Never Again, arguably the best Bond film ever, opened to enthusiastic response and the largest boxoffice receipts of any fall release in film history.
Though the cornerstone of Never Say Never Again was to be a renewed emphasis on story and character, a certain larger-than-life quality was mandated — and ably supplied through the rousing action sequences of second unit maestro Michael Moore and, in a more subtle sense, the understated optical trickery of effects supervisor David Dryer. Dryer, a 1965 Phi Beta Kappa graduate of the University of Southern California film school, spent three years as a film editor at John Urie and Associates before transitioning to The Haboush Company as a director of commercials. He next moved to EUE/Screen Gems where he spent twelve more years directing commercials, award-winning industrial films and cable television productions. In 1981, Dryer was engaged by Douglas Trumbull and Richard Yuricich to complete the work they had initiated on Blade Runner but were forced to abandon in deference to Brainstorm. Overseeing most of the actual effects production, and working closely with director Ridley Scott, Dryer earned an Oscar nomination for his first outing as visual effects supervisor. He then returned to commercial direction at Bluebird Productions — handling such major accounts as Nissan and C. W. Post — before being recalled to feature work on Never Say Never Again.
How did you come to be involved with Never Say Never Again?
Originally I received a call from Doc Erickson, who was one of the production executives on Blade Runner. Doc had gotten involved in the — at that time — unnamed James Bond film with Sean Connery, and was in Europe scouting locations. And it was he who recommended to Jack Schwartzman and Irvin Kershner that I be contacted regarding the visual effects. I read the script one night, and then had a meeting with Jack at his home the next day and outlined basically the kind of effects work I saw taking place. Apparently he liked some of the things he heard, so he suggested that I fly over to London for a couple of weeks. There was going to be a big preproduction meeting with Kershner and we could discuss in detail what Kershner had in mind and what I had in mind. So I went over to London in August of last year, and we proceeded to break out what we thought were the visual effects shots based upon the script they had at that time. The upshot of it was that I was hired on as the visual effects supervisor for the film with the agreement that I would also do some of the live-action second unit work. Initially, I was also to do the title sequence but I ended up bowing out on that. I just didn't feel there was enough time to do justice to both.
How was it that the effects work ended up being done at Apogee?
Since the film was being shot at various locations throughout Europe, with studio work at EMI Elstree, their first thought was for me to do the work in London. So during that initial two-week period, I looked at what was available over there and I concluded that it would have been a nightmare even to try. Although they do some wonderful effects work in England, I really didn't see any full-blown large-format motion control facilities. There were a couple of 35mm four-perf facilities, but nothing on the scale of what I'd been used to working with. It's really unfortunate — and surprising, in a way. If anyone does open up a large-scale effects facility over there, they're going to get an awful lot of work — especially with the dollar advantage of shooting live-action in London.
When it began looking as though the effects would have to be done here, Kershner suggested ILM — which was natural, since his last picture had been The Empire Strikes Back. ILM was particularly attractive because they have a policy of guaranteeing the cost per shot. And, of course, they're very good. So we approached Tom Smith (ILM's general manager) with the idea of doing Never Say Never Again, but it seems they were tied up with a little film called Return of the Jedi and didn't have a whole lot of time for other projects. They figured about the time we were due for release, they could probably start working on story-boards with us. So that closed the door pretty quickly. The other two obvious alternatives were Entertainment Effects Group and Apogee. I would have been very comfortable at EEQ, because of my experience there on Blade Runner, but the problem became one of available equipment. We were about a month away from principal photography, and even Richard Yuricich concurred that at such a late date, and with so little warning, it would have been difficult to mount a 65mm team in Europe. Plus, we'd have had a problem just getting the film processed, since there's only one lab in the world that handles 65mm, and that's at M-G-M in Culver City. So my final recommendation was to go with Apogee and their VistaVision format. Bob Shepherd and the guys over there had been very helpful, and had worked up a really detailed bid for us — which, as it turned out, proved to be pretty speculative, since as the project went on, everything changed. Hour to hour, it was a whole new ballgame.
Even though a final deal had not yet been made with Apogee, I left the first of October to join the main unit, which was already shooting in the south of France. Assuming that my recommendation would be followed, I rented a rackover VistaVision camera and began shooting plates. I was very fortunate to be assigned an assistant cameraman named Maurice Arnold, who ended up being my entire unit for most of the time. After a few weeks, Maurice managed to get Samuelson's to rework two of their old VistaVision cameras to our specifications. They weren't much more than just shells at the start, but by the time Samuelson's finished with them, they had pellicle viewing systems, motorized controls so we could run at various speeds, and they would take Nikon lenses. Unfortunately, we lost one of them to Greystoke, but the one we had worked reasonably well, with minor exceptions. One of the main problems was that even though the pellicle viewing system allowed you to look through the lens, it's basically a beam-splitter situation inside the camera and you lose a third of a stop exposure because your light is going through that extra glass. And it was very dark. We'd be stopped down to about f/16, and all you could see through the viewfinder was a little dot in the center of the frame. I ended up sportsfinding most of the shots. Later, when Apogee officially joined the project, they sent over a mirror-shuttered reflex camera. It had variable frame rates and a nice bright viewing system, and when that came over to Europe, everyone fell in love with it — including me. Finally we could see the edges of the frame!
Photos: An air launched cruise missile drops from the bomb bay of a B-l, extends its wings and fires its engines. / Effects cameraman John Sullivan adjusts his lighting on the large-scale B1 underbelly mockup used for the missile launch shots. A separate key light — representative of the sun — was motion-controlled to move up and down on a stand to impart a suggestion of aircraft roll to the locked off shot. / Inside the bomb bay was a fully-articulated rotary launcher for the missiles. To facilitate photography, both the miniature and the camera were inverted. / Sullivan prepares for a head-on view of the cruise missile. Even in the launch shots, the missiles were photographed separately from the B-l underbelly. / Modelmaker Ken Swenson airbrushes the camouflage paint job on the smaller-scale B-l used for all the flying shots.
Photos: Fore and aft views of the missiles cruising over the English countryside were generated at Apogee using motion control miniatures — photographed primarily by cameraman Doug Smith — matted over aerial background plates shot from a helicopter by effects supervisor David Dryer. Apogee's reverse bluescreen process — developed by Jon Erland — was employed to extract high-quality mattes from the shiny, light-colored miniatures, which would otherwise have created major "blue spill” problems with traditional rear-light bluescreen.
Apogee also ended up sending over their front projector, since all of the large format process projectors in England were tied up on other projects, too. Krull, for one, was using a lot of available equipment, and they went quite a bit longer than expected. I think Eon Productions at Pinewood had some front projection gear — but since they were producing the other Bond picture at the same time, obviously they didn't want to talk to us. We even ended up having to build our own front projection screen. We ordered the material direct from 3M, had it checked for matching, and brought a crew in to piece it together. It was a screen made up of one-foot-square and six-inch-square patches, all set at diagonals, all with edges going away from the center of the frame — so there were no shadowed edges. And the guys did a terrific job. We did have one near-disaster with it. It was quite large — about 43 feet wide and 27 feet high — and rather than use a hard surface, we decided to put it on stretched canvas so it could be rolled up and moved around. Over the canvas, they put strips of criss-crossed paper before laying down the 3M material. At about the time the paper was applied, the Christmas holidays came around, and during the break the heat was turned off and the change of temperature created all kinds of little wrinkles and warps. I had a fit about it, but fortunately they were able to sand it down and eventually it turned out to be a terrific screen. In fact, David Tomblin — who was the first assistant director on the main unit, and who had been second unit director on the first two Superman movies — said ours was the best one he'd ever seen.
Could you be more specific about some of the work you did on location?
Well, they'd already started shooting by the time I got to the south of France, and I really hit the ground running since I was scouting locations at the same time I was shooting. Basically, I was there to do the plate work, but I was also involved in doing live-action work, establishing shots, aerial shots and pickups. Mickey Moore was doing the primary second unit stuff, in terms of the chase sequence; and I was shooting plates with his unit and trying to interface with the first unit. The script was being rewritten all the time and there were many mornings when I wasn't quite sure where to go. So Maurice and 1 would just go off and do whatever I thought was needed. We got hell for it a few times, but it's all there in the film.
I also shot some of the stuff with the boat — which belongs to Adnan Khashoggi, a billionaire arms dealer who uses it to conduct his business outside of territorial limits. It's the second largest privately-owned yacht in the world, and it costs $50,000 a day just to keep it running. I inherited the job of talking with the captain, who was a very interesting Italian fellow. Originally, he thought they were only going to be there for a day or so, but we actually had them scheduled for a whole week's worth of shooting. Ultimately, we compromised at two-and-a-half days. So that was pretty tight. In fact, the opening aerial shot of the yacht — where the helicopter POV crosses the bow and then comes in on the landing pad at the stem — was shot on the final day as the boat was already underway to Italy for refitting. They wouldn't actually let us touch down on the deck, but we came in as close as we could, and then cut away to a low-angle shot of Largo in his helicopter, which was just sitting on the ground with blades spinning.
After that, we went on to Nassau for three weeks where I shot some more plates, including some of the aerial footage for the cruise missile sequence. Finally we went back to England for about ten weeks, where — among other things — we did all the process work and shot the rest of our aerial plates.
Were your aerial plates shot from a jet or a helicopter?
A helicopter. Actually we used a couple of them. The plates we did in the Bahamas were shot from a Jet Ranger. One of them was for a shot that didn't wind up in the movie, but the idea was that it was supposed to be a missile POV as it skims along at six hundred miles per hour about four feet off the water. For that, Peter Alwork, my aerial cameraman, used a Continental Camera mount for a VistaVision camera and positioned it right below the helicopter. To get the apparent speed we needed, we had to shoot at four frames per second, and the helicopter pilot did a terrific job of holding it steady, with a real easy banks. Later, over Penzance, in the south of England, we did a similar move over the castle on St. Michael's Mountain, although on that one we used an Augusta 109 with a VistaVision camera mounted upside down in the nose wheel well. This kept most of the camera body out of the wind and our 15mm lens right at the axis of the ship. We had a full clean view because it was out in front and we didn't have the blades or the belly in our way. Plus, the Augusta is a three-blade ship, so we didn't get as much yaw chop as we did with the Jet Ranger; and it would do 150 knots very steadily. We got some very good plates with the 109.
We also did the opening shots, and a lot of aerial plates for sequences that never got processed. One, for instance, was a scene that was going to be in the film after Largo is killed and Bond and Domino are flying away in the Coast Guard helicopter. Bond has Largo's little video communication device that you've seen earlier in the film, and Blofeld is still back at Spectre headquarters saying; "Come in Largol Come in Number One!" And Bond was going to come up on the screen and say: "Sorry, you've got the wrong number. It's 007." At that, Blofeld jumps back and is scratched by his cat — which had been established to have poisonous claws — and then Bond takes the communicator and dumps it out the window. The plan there was for us to do a motion control tumbling shot of Blofeld s face screaming on the communicator as it's falling off into the abyss of water down below — which then cuts to the girl diving into the pool in the final scene. Somewhere along the line, it all got taken out, but we were down there shooting plates for a lot of things like that.
Did you use the inertial navigation system Apogee developed for Firefox to record motion control data on your aerial plates?
No we didn't. For the relatively few number of shots we had to do, we decided there would be too much time and expense involved in sending that system out on location with us. It would've been nice, I suppose, but we figured if we planned the shots carefully, incorporating horizon and witness marks that we could use to track the movement manually, then that would work out just as well — which is what we did. When we were shooting, I tried to pick areas where there was enough ground detail so that we had good reference points, and then that footage was later tracked through a microscope at Apogee. David Bartholomew tracked each frame. There was a plotting system so that coordinates for the various witness marks could be written out from frame to frame, and then programmed into the motion control computer when the miniatures were being shot. I would suspect that if you were doing a film like Firefox, you wouldn't want to have to go through that kind of process for a hundred and fifty shots. But we only had three, so it was really no problem.
Was John Dykstra involved at all on your project?
He was primarily on Dune. We did have an agreement, though, that he would act as a consultant when needed, especially in my absence. I flew back and forth from Europe several times, but for the most part, the guys at Apogee were pretty much on their own during that period. So John would take a look at things from time to time, and advise the various departments. Actually Bob Shepherd was the real spearhead at Apogee.
Most of the model construction was done while I was away. Ken Swenson was in charge of the model shop at first, and he and I said hello and goodbye to each other as he went off to work on Dune just about the time I was getting back. Then David Beasley took over. Actually, the transition was pretty smooth. We were in production ahead of Dune, though there was some overlap. For the most part, the model shop was pretty much on my project full time. Then, as they finished working on Bond, they'd segue into the miniatures for Dune.
Could you talk a bit about the models?
Well, since we were dealing with real aircraft and real missiles, we of course wanted everything to be as accurate as possible. At one point we even tried to get help from the U.S. Government, but they asked to see a script and unfortunately someone sent them one. Not surprisingly, they had major problems with several key story points, and so we didn't get any help there. But we did have a guy named Brendan Alimo as our technical advisor on the film. He's an author and had a lot of very interesting information on the B-l and the cruise missile — nothing classified, but he follows all that very closely and had a lot of material all in one place. So we relied on him a great deal, and he was very helpful in keeping everything as accurate as possible. In fact, at one point during the shooting, I flew back to the United States for one day — to meet with Apogee and see how things were coming along — and I happened to have our plans for the B-l and the cruise missile and the hovercraft sandwiched in between some suits and jackets in my suitcase. So I'm going through customs and the inspector says to open my suitcase — and there inside are all these plans. The guy lifts them up and says, "Will you step aside please." And it was a bit tense there for a while, trying to convince them that we were making a James Bond movie and that these were plans for miniatures.
What size models were you dealing with on this show?
It varied with the shot. The full B-l bomber was about three feet long, and we had a larger six-foot closeup section for the belly shot where the bomb bay opens and the missiles drop out. The bomb bay doors were controlled with stepper motors, and - as they plan to do on the real B-l - there was a rotary launcher inside that indexed around so that when the first missile is dropped, the next one then rotates into position. Of course, when we photographed that, the bomber and the missiles were shot on separate passes — and just making the missile look like it actually fit in the rotary launcher involved a lot of tricky programming. The three-foot B-l was used for the wider shots. It too had motorized bomb bay doors, and then, rather than just having darkness up in there, we had a still photograph of the rotary launcher, shot from the same perspective, stuck up inside so there'd be some detail. You can't really see it, but we thought it was a good idea to protect ourselves, just in case.
The cruise missiles were basically two different sizes — one three feet long and the other eighteen inches. The wings on the smaller version were motion controlled so they'd open out when launched. The elevons — or tail fins — worked via stop motion, with eighteen separate positions, as the missile dropped away from the plane. It was still a motion control shot, but during those series of eighteen frames, we had to go in and articulate each elevon by hand. The larger model was more detailed, so we'd try to use that one whenever we had closeups. The smaller one we used whenever the missile had to start at apparent infinity and fly by us, or conversely fly past us out to infinity. Those were done on a seventy foot track. To get the same effect with a three-foot model, we'd have had to tear out walls and build a four hundred foot track. So the smaller model was definitely preferable. Before 1 came back from England, there was even a nine-inch miniature made. It was used on a couple of test shots, but it just didn't have the necessary detail and focus capability, considering how close the pass-bys needed to be.
When you got back from Europe, had any shooting actually been done at Apogee, or were they still working on the models?
I got back right before Valentine's Day, and at that point, they had done some element shooting on their own — but only one shot of the missile going away from us actually wound up in the finished film. Doug Smith was in charge of the photographic end of things, and he'd already shot this one particular setup as a beauty light test and an exhaust test, even though we'd never planned on that particular shot. We'd planned on the missiles coming at us, not away from us, but we said: "What the heck. We've got a plate for it. Let's put it together." So we used that as a testing piece — primarily just to see how we would fade our mattes in and out and what density we wanted to run things at — and ultimately, it ended up in the film.
Photos: To suggest the missile's ability to track its own trajectory and adjust against a predetermined flight plan, POV shots were overlaid with mock computer grids generated by animator Harry Moreau. / All but oblivious, two children play on the beach as the hijacked cruise missiles head out over the ocean. Engine exhaust effects were achieved on a separate pass by feeding a gas line into the motion control miniature — producing a blue flame which was then filtered to a reddish-orange.
What kinds of photographic techniques were employed?
It was pretty much standard motion control photography. We were doing model work on two different stages, with Doug programming and shooting on one and John Sullivan working the other. One of the things we found, though, was that we had to cheat a bit to make reality look real. The B-l, for example, actually flies more like a fighter than a heavy bomber, but when Doug Smith did his initial black-and-white move tests for the B-l shots, they just didn't look right Basically the thing just looked too much like a fighter; and even though it was probably close to being right on, both Kershner and I felt it ought to be more lumbering. Somehow it just didn't seem to project an appropriate sense of mass. So we kept working on the program and refining it — slowing it up a bit and giving it the lumbering look of something heavy and ominous coming through the air.
Photos: Poised at opposite ends of the gaming table, Bond and Largo prepare to play ",Domination" — a supposedly holographic videogame. In actuality, all of the material seen on the central screen, and on the smaller individual monitors, was generated in postproduction via model photography and animation. / A wire-frame globe was carefully prepared, with land areas covered by a sheer knit fabric which glowed blue under ultraviolet light. Colors were produced by successive passes of red, blue and green light — all of which could be combined to produce white. / Working over global projections traced onto halfspheres, Apogee model-makers soldered together wire outlines of the continental land masses and then transferred them to the wire-frame globe as templates over which the sheer fabric could be stretched.
We had a similar problem with the cruise missiles. Even though they have wings and look a lot like airplanes, they don't fly like airplanes at all. Tor one thing, they fly just subsonic and very close to the ground; and they have an inertial guidance system which periodically reorients itself to insure that it's on its proper trajectory. As a result, the missile can be flying along and then suddenly make an instant correction — very jerky. We examined some real footage of cruise missiles flying; and frankly, it looked phony as hell. But that's the way they fly. Since there's no one inside, there's no need for nice smooth movements. If it looks phony in real life, though, you can imagine the problem of trying to make it look convincing in miniature. You're constantly fighting audience preconceptions of how things ought to be. So we had to try and find a compromise where the missile would fly smoothly enough to look real, and yet still have those quick responses. Doug Smith's original program was just as smooth and slick as can be, but then we went back in and added extra vibrations and things to create a greater movement against the ground.
Did you use the same matting system Apogee developed for Firefox?
Yes, but with a few variations Jon Erland has come up with since. The basic reverse bluescreen process involves the use of a special phosphorus lacquer which is applied to the model right over its final paint job. When photographed under normal beauty-pass lighting, the lacquer is invisible. But during its matte pass, the model is lit with ultraviolet light and glows blue — or red, depending on the phosphors involved. That pass, photographed against black, can then be used to generate whatever mattes are necessary. The major advantage of the process is that it allows you to photograph a shiny object — like an airplane or missile — without the accompanying blue spill you'd be likely to get if you used a traditional background blue.
In addition to the reverse bluescreen matte, sometimes we'd put up a red screen behind the model to give optical an additional image to work with. In some cases — as with the flying hovercraft, or capsules as we called them — we'd also put up 3M front projection material with little lights around the lens and pull a contrast matte. The whole idea was to give Roger Domey and his people in optical as much flexibility as possible in assembling these images. Sometimes one type of matte worked best. At other times, another one might be better. Sometimes we'd use more than one and sandwich them together. Jon Erland now calls it Mix-a-Matte. Whatever worked best And we were constantly swelling or contracting our mattes to try and find the right fit and look. Matting light-colored objects, like the ones we were using, against light-colored backgrounds is a real nightmare — especially with the kind of spindly lines and detail we had on the capsules. And the problem is that if you actually get a matte that fits perfectly, it looks wrong — because there is no such thing in the real world as a fine hard edge between one object and its background. Tor one thing, any object is always reflecting the light that's around it. So most of the time, 1 was fighting to have everything pulled in, and we were actually DX-ing the matte lines as much as possible to get that kind of blur that happens in reality.
Of course, whenever you shoot miniatures, you always try to match lighting direction and lighting quality. In addition, on our beauty passes, we started putting colored cards behind each object that would match generally with the live-action background. So if there was a blue sky up above, we'd use a blue card with maybe some white mixed in. To a certain degree, that gave us some semi-reflective lighting or color reflection on the object itself. What was even more successful was to take the actual background plate and, on an additional pass, set up a rear projection screen behind the model and project the plate and let the object itself reflect a certain amount of what was behind it. So in dailies, we'd have our beauty pass, shot against black, and then this reflection pass with the plate running behind it. In the final composite, the matte would remove the plate area, leaving just the reflection on the model, which would then be wedge-tested and double-exposed onto the beauty light pass.
What did you do for the engine exhaust effects on your cruise missiles?
Those were actually flame effects done on stage as an additional pass. We had a gas line feeding into the miniature, which produced a burning blue flame that was then just filtered to orangish-red.
How about the cruise missile POV shots where you're superimposing a computerized grid of the terrain?
The idea there was to create a representation of the missile's "brain." Brendan gave me a lot of information about how the cruise missile is programmed to follow the terrain. It has its program, and then it has the real world, and it's constantly trying to make the two coincide. So we tried to graphically demonstrate that continuous updating process by cutting to frozen eighteen-frame cuts done as continuous tone black-and-white images with a blue tint. Then, on top of that, we double-exposed a vector grid — rotoscoped and animated by Harry Moreau — as if the missile were reading where it was and adjusting its position.
What sorts of problems did the hovercraft present?
Well, for one thing, they were very spindly, and therefore difficult to matte. It was nothing to have thirteen to fifteen optical takes before we got one we could live with. I'm a real maniac about matte lines, as is Kershner. Anything that had even a hint of a matte line, and he was on the telephone — and the telexes were going back and forth, too. So after a while, everyone got wise to the fact that we were just going to have to do these things over and over again until we got them right. So we kept pulling and squeezing and doing everything we could to make those damn matte lines disappear. What was doubly frustrating on the hovercraft shots was that we worked really hard to get rid of the matte lines, and then when they were duped in the final print-making stage, they developed what looked like matte lines, but were actually halating grain patterns dancing against the edge of the surface.
Plus, Doug Smith is convinced that the first shot in that sequence — where the shells open up and the booster falls off — is the most difficult motion control shot ever done. We realized that with so many things going on at once it would be very risky trying to shoot everything separately and then matte them all together. So the guys at Apogee put together a very complicated rig with offset arms to split the missile open and lift the capsule out and have it all tied together in one piece of machinery. The hovercraft, which was about two feet from top to bottom, also had very detailed figures — about twelve inches tall — with heads and arms and necks that moved. It took Doug about a week and a half to program the whole shot.
Then we had the nightmare of the vector jets. A C02 pass would give us the basic blast, but the speed would be all wrong at the one-second frame rate we were shooting. So we went back and did miniature smoke projection passes — using full-scale smoke effects shot at 120 frames per second out in the parking lot. Then we put a little white card up next to the hovercraft, and set up a projector on a motion control pan and tilt head to follow the capsule's movement. That gave us a separate piece of film we could then burn in over the beauty pass — which really helped give the smoke and mist the proper feeling.
How did the "Domination" game between Bond and Largo come about?
If you stop and think about it, every Bond film has a game sequence of one sort or another, and usually the first confrontation between Bond and the villain is always some sort of game. So either Kershner or (screenwriter) Lorenzo Semple came up with the idea of using a videogame — which was then expanded out to be a holographic video game. At first, Atari and Mattel and some of the other videogame companies made presentations of their own designs for the game. But Kershner was never happy with those — nor was I. If anything, they just went too far. One of the companies came up with a game that existed in a dome which surrounded the players, and they had battleships and submarines and all kinds of things going on. My feeling was that the game would be so complicated and confusing that no one would be able to get a handle on what was going on. So when I got over to London, I made my own presentation of what 1 thought the game should be — and Kershner bought it. There was a writer there at the time, and we spent about three solid days of that initial two-week period just flushing out the premise and embellishing the game — coming up with the idea of losing more than just points, of actually getting an electric shock when you fall behind. And, in theory, the game actually works. If you could produce 3-D holographic images in space, it would be a very exciting game to play.
The basic idea of the game is that you have a country in free space within a frame over the table, and then each player has a video monitor with a cursor that shows the positioning of his own lasers and missiles. You move your joysticks around in this dimensional space and you fire your weapons. If your laser hits the city or territory you want, you then capture that territory. Initially, the idea was that you would go from city to city and draw a line; and as you circled an area, that area would become yours and you could double the amount of points for a city. The trap was that as you were ready to encircle your area, your opponent could tell where your cursor was and he could fire a missile and blow you away. In postproduction, that basic concept changed a bit. If anything, we got too complicated. The original demonstration on France makes more sense to me than the later games where the images are stretched out and distorted. I feel that had we limited it to a map in relief that existed in a recognizable form — without being stretched or foreshortened — it would have made a lot more sense to the general viewer, for a long time, I fought to have every aspect of the game fully understandable to the audience. But then, toward the end, I gave up on that with the realization that you can't understand everything about a video game the first time you see it. You have to play the game to figure it out — which is what happens in the story. Bond gets beaten for three games and then finally starts to get the hang of it.
Was the live-action for this sequence shot at the studio?
Actually, it was shot at the Rothschild's home in England, called Waddesdon Manor. There was a great old French rococo sort of room there, and we pulled their furniture out and put the table in, with very little additional dressing. It was just what Stephen Grimes, the production designer, had in mind. One of the concepts that I brought to it was that the table itself, rather than being some space-age futuristic thing, be an old gaming table made with polished woods and a fine leather top. Everything we did in that sequence was to be an extension of Largo's personality — and that seemed to fit the character.
It was actually a very complicated sequence. Kershner and I would work nights and sometimes weekends, storyboarding the thing from beginning to end. And the sequence varied a little bit every time we worked together. One of the things that soon became apparent was that the three-and-a-half days allotted for shooting that location were not going to be enough to do all the Panavision shooting — the tight shots and the dialog sequences — and then bring in the VistaVision crew to shoot the effects plates. The only way we could get around the problem was by working together. So, anyway, we had our two VistaVision cameras running sync sound simultaneously with the first unit's Panavision cameras. Ideally, they should have been blimped; but they weren't, so we did the best we could, building baffles and covering them with blankets. As director of photography, Douglas Slocombe was in charge of the lighting and I would take my f-stops from him — and negotiate sometimes. Although the final game imagery hadn't been settled upon, I had Apogee make up some rough artwork to put on slides, and I had two Carousel slide projectors set up so that while we were shooting the actors we could dissolve these out-of-focus images back and forth on their faces to look like there was interactive light hitting them from the game. And that sells pretty well. They're never quite the right images, but it's nothing anyone is likely to notice.
Photos: During the live-action photography, only an empty frame existed over the table. In anticipation of the holographic images that would eventually fill the area, colored light patterns were projected onto actors Sean Connery and Klaus Maria Brandauer to suggest an interactive lighting effect. / Material for the individual video screens was produced entirely in the animation department, supervised by Cameron Striewski. During principal photography, a white screen was employed to facilitate a simple matte extraction. Then, once the viewscreen animation was completed in postproduction, the game table was placed on the Apogee stage and the camera aligned exactly as it had been during the live-action take. The composite animation material could then be projected onto the screen and recorded, with appropriate keystoning and reflective effects.
How were the images on the screens created?
There are actually two different categories. First would be the video screens that are immediately in front of each of the players. The material on those screens was produced entirely in the animation department. We had an animation director, Cameron Striew-ski, and he worked closely with Harry Moreau to produce all of the material that would go into those screens. With the map and cursor positions and the missiles, plus various graphics elements and the points running up at the top, it got pretty involved — but we didn't want to lay that on the optical department, which was already busy doing all the other stuff. As a result, there were a lot of multiple passes on the animation stand to get all that information together on one VistaVision plate.
At the time we were shooting the live-action, we took great pains to measure off every single angle to all of the important points. We knew what lens and what camera we had used, its height from the floor, its inclination, plus any roll it might have had. Then we measured off the distances from the camera lens to the four comers of the big frame where the holographic image would be, and to the comers of the little video screens the players would be watching. Since the video material wasn't ready before principal photography, we used a white screen on the set so we could easily identify the area that needed to be matted in. Then later, after we produced the VistaVision composite animation plates, we set up a screen at Apogee and projected that plate material onto it. This screen was set in the table to match its original perspective in the live-action shot. As a result, any subtle keystoning in the image would be appropriate to that particular angle, rather than just burned in at a straight-on angle, which always drives me nuts when 1 see it. We also tried to duplicate the subtlety of 525 raster lines by using a screen that was made from a modelmaker's material for wood-sided houses that has a lot of very fine lines etched into it. As it happens, for the given height of our screen, there were 520 or so of these lines, which we enhanced with a very Fine pencil point along each one of the grooves. You can see the effect in the VistaVision plate, though it doesn't really translate in the final optical. But my theory is that if all those things are apparently correct, you get the right subtle feeling, no matter how far down the line you are. In addition, we took the white screen material from the original live-action photography and used five percent of that, double-exposed over the image, so that whatever reflected lighting there was on that screen from the room around it was also on our optical. It helps to give you the feeling of a real surface being there, instead of a cut-out image that looks almost like you're seeing into another room.
Photos: Since one of the countries had to rotate in space during the game, a wire-frame representation was constructed. Harry Stover, Mike Speaker and lllyanna Lowry solder the faceted segments together. / Nonrotating countries were rendered on glass. To generate an appropriate image, however, a three-dimensional buck was built to match the receding perspective of the game, with a faceted approximation of each country's geography drawn on its walls. / The buck was assembled over the game table and photographed in flat perspective, then traced onto a large pane of glass covered with black paint. Following the tracings, Harry Stover and John Shourt carefully scribe away the paint, producing an array of interconnected lines which could then be backlit with laser light to achieve the desired scintillating effect. / Jim Dickson, director of photography for the game sequence, ties off a fiber optic bundle connected to the rear side of the panel representing France. When lit, each fiber optic bit signified a city on the map. / Individual facets were colored with gels, and laser beams were rotoscoped by Harry Moreau.
Were you able to pull simple mattes from the white screen area or did you have to go in and rotoscope?
Well, we didn't plan to do any rotoscoping, but no matter how much you try to explain everything during shooting, there's always someone who ends up having his thumb in front of the screen. Although 1 was behind the camera on much of the plate photography, in this case I wasn't because I was running the slide projectors for the interactive lighting. So I wasn't in a position to tell, necessarily, if there was a problem. And sure enough, Kershner and I would be sitting in dailies the next day tearing our hair out because somebody's thumb would be going over the screen. So whenever that happened, we had to go in and roto the thumb and make a little traveling matte for the screen area. Nobody notices — but if we hadn't done it, they'd have noticed.
What was involved in creating the imagery on the big screen?
The concept for the big screen was that of a three-dimensional image in space, projected by means of a holographic projector into the area over the table. The initial idea was that we would build dimensional wire models for everything, and shoot them under motion control. The basic concept I had in terms of execution, though, was that I wanted to use laser light rather than incandescent light to illuminate the objects, so that the image would have a scintillating holographic look to it. There's no such thing as a continuous clean line in holography. It seems to have a vibration to it, as a laser light does. So we used laser light almost exclusively.
Did you use a scanning technique to light the models?
As it turned out, we didn't always use models. But yes, in some cases we used laser scanning. More often than not, though, we used interference effects, either projected onto a dimensional object — which was painted with 3M paint to make it highly reflective — or we projected the laser patterns through etched glass. Mark Gredell operated a krypton laser for us — commonly called a white light laser, even though it's not a true white light. It has spectrum lines in it that represent the major primary colors. You get the interference patterns by taking that beam and projecting it onto a reflective medium of some kind that, in effect, is taking that beam and lensing it thousands of times. We used a mirrored surface that had a drop of acrylic on it — about the size of a penny. The acrylic's not perfectly level. In fact, it has a lot of little bubbles in it. And every time that surface bends or distorts a little bit over the mirror, it is going to refract the light image differently. We also put a little bit of oil onto the acrylic, so that as the beam of light hits it, it is hitting each one of those tiny bubbles and the oil itself and lensing it and coming back as thousands of little laser light images. In addition, the mirror slowly rotates, creating thousands of little radiating lines. You can lean your color toward one direction, but all the colors are in there and you've got a thousand little rainbows working simultaneously. This effect can be projected off of the mirrored surface — in fact, we used a lot of mirrors just to bounce the image around.
The problem, though, is that this kind of laser light is a nightmare to photograph. Since each little refracted image is basically a straight-line source, and since they are being reflected off in all directions, you have a situation where you're reading one f-stop at one point, and then a little ways away it's fifty stops different. So what do you go for? The fall-off is just immense. We went through a lot of black-and-white Polaroid film trying to balance the left side of the screen with the right side. So it was a lot of painstaking work. But it did produce an interesting effect and a scintillating quality for the game.
You said that you ended up not using dimensional models throughout?
Yes. The globe, for example, was a wire model, illuminated through a beam splitter on the axis with the lens. It was mounted on a gimbal mechanism and we set up a motion control program to have it rotate and roll — and we made three passes, one for each of the primary colors. Later, as the game progresses, we planned to build faceted wire models for each of the countries we were going to be using, and then project laser light onto them and photograph them in proper perspective to our original camera angles. But when we started breaking down all the elements involved, and began considering all the wire models that would have to be made to do the kind of image stretch that Kershner wanted, it would have been a tremendous job. I'd prefer to have seen a more recognizable image, but Kershner felt very strongly that the countries in their normal form would not be dimensional enough. He wanted to create a kind of space warp where they stretched out in front of your eyes.
At any rate, to take the image of the United States, for example, and slitscan it out into a sort of forced perspective tunnel, would have been very difficult the way we originally envisioned it. To do the four sides of that in wire models and then go back and fill in each facet with some medium — tracing paper or whatever — that would catch the blue or the red light as that facet was hit, would have been prohibitive. We just didn't have the time or the money to do it. As it was, there were over three hundred individual elements that went into that game. So we took the actual gaming table — which we'd had shipped over from England — and placed it in position to the camera to match the original live-action scene. Then, out of white foam core, we built a four-sided buck that created the desired forced perspective. And we started moving that around until we found the position that looked right for each particular angle, now, Cameron very carefully drew lines on each of the four sides of the foam core until he created the faceted map that he was happy with. Then we photographed that from each angle with a Nikon camera to give us a two-dimensional photograph of this three-dimensional shape. From there, we projected the photograph onto a flat piece of white card now positioned over the table. Somebody then traced all those lines onto the white card. We then took that and transferred the lines to glass sheets of equal size — each one painted black. Methodically, a team of three people scratched the black paint off the glass to create each line. A line that was closer would be thicker, and a line going off to the background would be thinner and thinner until it was hardly there at all. So we had twelve to fifteen of these "stained glass" panes which we then set over the table again. Then we projected a krypton laser light through the glass — so we would only see it as it came through each line. Finally, we rephotographed that in proper perspective from the front, making a lot of different passes since we couldn't get the laser to cover the entire area. It turned out to be a very laborious, time-consuming process. But it did beat building the wire models.
Photos: To incorporate cruise missiles into the game, the miniature missile used earlier in the film was coated with red and blue phosphorescent lacquer and photographed with ultraviolet light to produce either a glowing red or blue image, which could then be composited over the game. / Mark Cane and Mark Gredell examine the laser used as a light source for much of the game imagery. / Gelled light was front-projected onto the globe to create multicolored overlapping images.
It sounds as though it might have been easier just to animate it.
Well, a number of people have suggested that. And we did run some tests on that approach, but we just weren't happy with the look of it. In fact, it looked like we did exactly that. We even tried to back-project laser interference patterns on the VistaVision animation stand — just undulating blue lights through a window matte of lines. But at that scale, it ended up just looking like animated line drawings rather than having a holographic look to it. So we went with the large-scale approach, using matte painting sized glass.
And you actually photographed it over the game table?
Right. We took the table, painted it black and set it into position on the stage at Apogee. Then, every time we lined up a shot, we actually put whatever was there over the table. That way, when we illuminated the object with our laser light from behind, the light coming through the lines and the facets on the glass was also reflected down onto the surface of the table and would thus be double-exposed over the live-action image — which, of course, helped the illusion of the "holographic" image actually being there.
Were the specific details of the game pretty well worked out before principal photography began?
No, they weren't. We had our basic concept, but the finer points were still a matter of discussion. Prior to settling in on a final plan for the game, Kershner and I agreed that we should first get a rough cut of the live-action to see how it cut dramatically. At that point, we could see how many shots we had and where we were. Then, we could come up with a final decision as to the rules of the game and what kinds of things needed to happen. So Kershner and I worked on the cut and, once we agreed on it, were able to come up with a satisfactory blueprint for the game. Then he went back to England — since his hands were full put-ing the whole movie together — and I stayed in California to execute all the effects.
As it turned out, the game is considerably different from our original concept. The way it works now is that inside each facet is a city, and as that city flares up in brightness, it means it is vulnerable and you have about a second and a half to react with your joysticks and fire a laser to capture that territory. The first player to hit that city then owns it, and the territory around that city becomes either red or blue. You can't just randomly fire into the game and capture territories.
Once the facets were red or blue, were these done with gels on your windows?
Yes. We actually had two sheets of glass bipacked together. The foreground glass was our etched line glass, and the background glass was a ground glass sheet that would just image the light source coming to it. We started the first couple of takes with laser light and decided that although it looked more interesting, it was harder to see. So eventually, we went to a process of exposing gelled lights through a window matte for each facet.
Photos: Miniatures were used for the scene in which Bond and Domino jump their horse off the fortress walls. A hard-plastic model horse was purchased, then cut into pieces and reassembled with articulated joints. Illyanna Lowry adds some early detailing to the horse, which was subsequently covered with felt and outfitted with miniature riding gear. / Although the primary movement of the horse and riders was generated under motion control, animator Harry Walton was called upon to add subtle movements via stop-motion animation. Since the hand-animated portions of the scene were nonrepeatable, alternate beauty light and red phosphorus matte frames were exposed — then skipped out and separated in optical. / The fortress walls — being detailed and painted by Harry Stover and Catherine Hobel — were constructed in miniature. / After determining that additional aging and detail were required, David Dryer personally became involved in the on stage alterations. / Horse and riders perched atop the fortress walls.
Was the game imagery simply double-exposed over the live-action plate, or were there mattes involved?
Both. Sometimes we'd use partial density mattes to help us hold back an area; other times we let it go through a hundred percent.
How about the cruise missiles fired during the game?
The missiles we had to hold back. Otherwise, it just became too confusing against all the lines. The game missiles were the same ones we had used earlier in the film — shot with basically the same motion control techniques to look like they were in proper perspective to our game. The lighting approach was considerably different, however. By the time we got to the game, we'd finished all our main effects photography for the cruise missile sequence. So we took the missiles and painted one blue and one red, with the phosphorus coating we used for the reverse bluescreen. Then, if we wanted a red missile, we'd use the red missile against black; or if we wanted a blue missile, we'd use the blue missile against black. In essence, what would normally be a reverse bluescreen matte pass became both the matte pass and the beauty pass.
What technique did you use for producing the laser beams in the game?
The laser beams that they're firing were all rotoscoped by Harry Moreau — white line beams against a black background that were then bipacked in the optical printer. Pretty standard stuff. What we did differently, however, was we shot a full field of laser light to bipack with the black-and-white roto beams. That would then matte through the line, resulting in a line area that scintillated like true laser light. In fact, we found through testing that that gave us virtually the same look as a photographed laser beam and much more simply. If we'd tried to use real laser beams for every angle we wanted, we'd still be there doing it.
We used the same technique for the laser beam from Bond's watch. The live-action plates were shot in London — mostly with a stand-in, although Sean was in a couple of the shots that didn't wind up in the film. There were a number of practical effects, including little spark charges and a glowing lightbulb that we rheostated up as the laser beam supposedly burned through the metal. Then the charge would go off and break the manacle. The actual beam we did later at Apogee, using a laser plate bipacked with a window matte — just letting the light through that line area. Two optical passes. One pass sharp, and another pass with some diffusion — pretty much the same way lightning effects are done. Then Harry Moreau went in and added some more animated sparks, just to help it along.
How did you go about doing the shot where Bond and Domino jump their horse off the top of the fortress?
Basically, the script called for our hero and heroine to escape on horseback from a horde of Arab slave traders, and the only way for them to get away was to jump off the tower and into the water down below. Well, if a thirty-foot fall is terrific, and a fifty-foot fall is even better, then a 150-foot fall has to be better yet. So we all agreed that what we needed was a 150-foot leap off the fort walls into the water. Needless to say, though, it quickly became apparent that we weren't about to get any stunt people, or a horse, that would be willing to actually do that. So they shot the lead-in live-action where Sean and Kim Basinger start to go over the wall, and then Mickey Moore and his unit set up the end shot where a couple of stunt people and a horse actually plunge thirty feet into the water — and we had the job of filling in the middle.
We had always talked about making it a very wide-angle shot to show the scope. If we made it a tight angle, no one would really know what was happening — and, quite frankly, it would have given away what we were doing. So we proceeded to build the fort walls in miniature — about twenty feet across and about six feet high — with a sky backing. Everything in the frame, in fact, is a miniature. We considered doing it as a matte painting, but we figured that a miniature would give a better sense of dimension to it. The horse and riders were also miniatures — though on a considerably larger scale. The horse was about fourteen to sixteen inches long, and the figures were about ten inches high. Those were shot separately, also using the reverse bluescreen process — although in this case we found that the blue phosphor did not work well on the felted material that was used for the horse hide. For some reason, the red phosphor gave a much more even reproduction overall when illuminated with the ultraviolet lighting. The fall itself was photographed by John Sullivan as a combination of stop motion and motion control. The streaking movement and a certain amount of yaw and pitch were programmed in as the figures fell through the scene, while the movement of the horse's limbs and the articulation of the riders was all stop motion. So a frame would be shot for the motion control, and then a matte frame would be done right after it, since simultaneously Harry Walton, the animator, would have to go in and move the figures a little bit. So we just ping-ponged all the way down the shot, pulling the matte and the beauty light pass off alternate frames.
I'm glad it worked out. Quite frankly, that was the one shot I was most concerned about. We pushed and pulled the red matte against the blue matte. We went all the way up to take 13 in optical before finally settling on take 11.1 don't like hard matte edges, and I kept sending Roger Domey back to soften them up; but finally he said, "I agree with you that your matte line looks better in VistaVision, but by the time we reduce this to 35mm, it's going to mush away and you'll lose everything." So, we ran a test and he was right. That's why we went back to take 11. There were other headaches, too. We'd sit and watch the VistaVision dailies over and over and it didn't look right. Finally, I decided that we had to have something else that gave some motion to the scene, because — even though your eye goes to the horse — if nothing else is moving, something seems wrong. So up in the upper left comer there are two tiny seagulls flying along that Harry Moreau laboriously animated. Everyone said no one was ever going to see them, and they're right. Nobody ever does. But as soon as you do things like that, the shot starts to come alive. We also added some smoke coming out of a building, and some smoke blowing off a wall in another part of the frame. I think people see all that stuff subconsciously.
Photos: The rescue of Bond and Domino from the water was staged in the Bahamas, then augmented later with a matte painting of the fortress and shoreline rendered by Lou Lichtenfield. The submarine, also, was painted—with simple slot gags executed on the matte stand to suggest rocket flashes. / To establish the fictional Palmyra setting and its relationship to the sea, separate live-action plates were photographed — one in the gardens of the Rothschild estate, and another overlooking the harbor and Largo's yacht. The two plates were then linked together by another Lichtenfield painting of the interior fortress walls. For reasons of pacing, the elaborate matte shot was ultimately deleted from the final cut.
Is the shot of Connery and Kim Basinger just preceding the jump a front projection shot?
That does have a peculiar look, and I'm not sure just what they did there. It was an afterthought. Kershner felt that they had to show that it was Connery, and so that was actually done as an insert in postproduction. As a result, it was a matter of trying to shoot a matching connective shot two continents away and four months later. Since the shot before it was sun-flared, I guess they decided to tie it all together by blowing it out with a lens flare, and I think they even distorted it a little bit — some kind of zoom lens dolly shot, I believe.
After they land in the water, there are a couple of shots showing the submarine firing on the fortress. Did those involve models, as well?
Mo. As a matter of fact, those were matte paintings. The live-action portions of the chase were shot at two locations in the south of France. One was an old citadel at Villefranche where the auction block and the big gate were shot. Most of the rest was done at an old fort in Antibes, about thirty miles to the west of Mice. But since nothing we found had exactly the right geography we wanted for the jump shot, the establishing shots were either miniatures or paintings. In the case of the long shot of the submarine firing on the fort, the live-action — which has the hero and heroine in the water and the boats coming to pick them up — was shot in Nassau. The fort walls were painted over an existing shoreline, and the submarine was also painted. The rockets were done with a fairly simple slot gag.
I can't recall Apogee having done matte paintings in the past.
Well, they didn't have a matte painting department as such. In fact, originally I contacted Matt Yuricich — with whom I'd worked on Blade Runner — and he was going to do it. But then our date pushed back a little and we ran into a conflict of schedules, so he had to back out. At the time, Apogee was involved in Dune, and they knew there would be matte paintings in that project, so they were talking about bringing in Lou Lichten-field — who's done a lot of work for Warner Brothers through the years — to set up a matte camera stand and be their painter in residence. So they suggested Lou and I agreed. He's very good, he's very fast, and he's very knowledgeable.
What kind of matte camera operation did they set up?
It's a VistaVision matte camera stand. Paint on glass. They can do a conventional matte shot where it's combined in-camera with the light pass to expose the interpositive or — in the case of the Bond film — we used separations. For all the live-action, we found we were able to control our color better with separations than with an interpositive. The painting shots we had were completed right on the matte stand; but if necessary, they can do more complicated matte passes that cycle through the optical department. They've also recently set up a beam splitter with a second camera/projector. With that, they can take a plate that jiggles and rephotograph it with pan and tilt on the camera so that the horizon and other witness marks always land in the right places — which really helps if you've got some bad plates to work with.
There were three other matte paintings planned for the film which ended up not being used. At some point in their execution, I'd get a phone call from England saying; "Cease and desist on that shot. It's no longer in the movie." Some of them we'd already completed, some we hadn't. The first was a shot of Bond and Domino on the deck of the boat, and the camera's looking past them toward the fort. The shot just before was a point of view through Bond's binoculars, which I shot at the fort in Antibes, and which is in the film. Then it was going to cut back to the longer view where the fort was a painting with a live-action ocean plate in the lower part of the frame. Since there were actors involved, the final shot had to be a bluescreen composite, and the live-action plate for that was done in England, with the artists against our front projection screen onto which we projected the blue using a front projector. It was a pretty bricky matte, I especially with Kim Basinger's long blonde hair, and we fiddled around with it quite a bit. We also planned to add some smoke and birds and things to the painting, but before we completed it, we got the word that it had been cut. Another one that was cut was a reverse angle shot of Largo as he opens the gate to reveal his palace and tells Domino that it's her new home. We didn't get too far along on that one.
The one I was really disappointed about was a very nice establishing shot of Largo leading Bond and Domino toward the palace at Palmyra. There were two separate live-action plates in that shot. The first, with the three actors walking along, was shot from a high angle overlooking the formal gardens at the Villa Rothschild in the south of France. That was to dominate most of the lower portion of the frame. Around the perimeter of the gardens, Lou painted in the fortress walls — which, of course, weren't actually there. Then, up in the upper left-hand comer, we wanted to be able to see over the walls to the ocean beyond, with Largo's yacht anchored in the harbor. To get the proper perspective, I shot that part of the plate from the top of Loew's Hotel — right below the casino in Monte Carlo — as the boat was coming back in to dock after a day of shooting with the first unit. We radioed out to them and said, "Okay, slow down, stop, hold it there," and we rolled some footage. So we had two live-action plates, shot at different times and at different places, connected together by the matte painting of the walls. And it was a really nice-looking shot that oriented everyone as to where the fort was and where the boat was and where everything was in relation to everything else. But as it turned out, at that point in the movie, nobody really gives a darn about orientation. They want to get on with it. They want to know if Bond lives or dies. So ultimately, it was cut. It was a really nice shot, but it just wasn't that important.
Photos: Model shop supervisor David Beasley interfaces an articulated James Bond puppet with its attendant motion control system. Lifelike figures of Bond and Felix Leiter were employed in scenes featuring the miniature hovercraft. / Modelmaker Tom Pahk connects the puppet wiring into the capsule base. / Actor Sean Connery aboard a full-size mockup of the hovercraft, constructed for on-location closeups and used also on the stage to produce a number of front-projected composites.
Which of your front projection shots actually made it into the film?
Ironically, I don't think any of them did. A lot of the stuff of Bond in the hovercraft originally was to be shot front projection — the plan being to take that, cut it in with the miniature work, and they'd have their sequence. I shot all the plates for the hovercraft shots down in Nassau, with some nice puffy clouds in the background to give us some distance relationship. After London, we went to Spain and actually did the full-size shots of them landing, and there wasn't a cloud in the sky. As a result, all of our front projection takes didn't match anything. So those shots went away. Also, there were to be some shots of Sean on the motorcycle. We shot plates of the streets and things in the background, and then set up a gimbal-mounted rig for the motorbike. But the rule is that if you can do something for real, do it for real; and since they had a couple of shots they'd done on location that looked good, those were used instead. So our front projection material turned out to be strictly insurance. flow much time did you end up spending on the film?
A total of nine months. I was overseas for about four-and-a-half months, and the balance of the time was at Apogee. Between mid-February — when I got back — and the ninth of July, we turned out fifty-five effects shots.
Having worked in both 65mm and VistaVision, do you have any comments on the respective advantages or disadvantages of each?
In a perfect world, for my money, every movie would be shot in 65mm. It is clearly the best way to go. The usable frame area is considerably larger, and once you've seen 65mm dailies, you don't want anything else. The difference is night and day. But VistaVision is a good format. If it's joined to the original photography at the right place, and integrated properly, it holds up very well. And there are other practical considerations. For one thing, the fact that you can only get 65mm processed in one place in the world makes working with it overseas a real pain. Plus it's more expensive. To do the same amount of work in 65mm that we did in VistaVision would have cost us considerably more. So there are always tradeoffs.
For the most part, I'm pleased with our work on Never Say Never Again. Our intention was that if we all did our jobs properly, hopefully no one would know that there were visual effects in this film. We were locked into reality constantly. Beyond the basic Bond/Spectre thing, we didn't want to make anything look larger than life or phony. But what is larger than life? A lot of the live-action stunts in this film are a lot more spectacular and unbelievable than anything in the effects area. We were there to do things that either physically couldn't be done, or would have cost too much money to do, or would have required sets that just didn't exist. From a production point of view, if it wasn't feasible to build a particular set or shoot a particular shot at a particular time, then we were there as a stop-gap to plug in all those holes. It's a difficult position to be in, and much more demanding than the more usual type of effects work. On Blade Runner, I became as much an art director as anything else — and it was fun. We were creating images from scratch; and if we had a problem, we could be creative as to how to solve it. We have a matte line on something? no problem. We'll put a bright light in there and it'll glow and everybody'll love it. But you can't do that on a cruise missile in bright daylight. So we were constantly stuck with fighting the reality factor, Even the color of the cruise missile was a problem — mainly because it was designed by the government not to show up in the sky. So if we wanted to be authentic — and we did — we always had those kinds of things to deal with. Many of the reviewers have commented that Never Say Never Again is James Bond without all the gadgets and gags that have come to be associated with the series. Well, there's just as many gags and outlandish things in this one as there are in the others; and I think it's a credit to Irvin Kershner, primarily, that it doesn't seem that way. He insisted that no matter how far out these things were that everything be made to seem as plausible as possible. And it worked.
Photos: The submarine-launched missile splits apart to reveal James Bond aboard his hovercraft. / Rather than risk a mismatch by shooting hovercraft, capsule halves and rocket base separately, a decision was made to build a single motorized rig that would control and maneuver all four elements at once. Atop a ladder, Bill Shourt — who engineered and built the extremely complex mechanism — makes final adjustments prior to a film test. / Three successive exposures demonstrate the operation of the motion control device, which cameraman Doug Smith spent more than a week programming. Rocket and retro-jet effects were produced on separate passes, using smoke projections on white foam core.
[Source: Cinefex #15, January 1984, P.4-27. Copyright © 1984 by Don Shay. All rights reserved. Never Say Never Again photographs copyright © 1983 by Taliafilm and released through Warner Bros. All rights reserved. Effects unit still photography by Michael Middleton.]