When we came to make a decision about what film format to use for Tron, we felt very strongly that we needed to shoot it on a larger negative than traditionally done. And it came down to a choice between 65 millimeter and VistaVision for the whole show. But the availability of 65-millimeter cameras was far better than the availability of VistaVision cameras, so we made the decision to shoot the entire show in 65 millimeter. And to keep a certain consistency, and as long as we had rented the cameras, we felt that we might as well go ahead and shoot the real world in 65 millimeter and not in 35 millimeter. It was kind of one of those nice to have things, and nobody objected strenuously to it, and also it would be the first time since Ryan's Daughter that a film had been shot entirely in 65 millimeter. And I think you can see the results. I mean, it looks wonderful. In the monitors in Lawrence Livermore, one of the things that Triple-/ had to do was create lots of imagery that appeared on monitors, and that imagery had to be shot and created long before we got to production. So, there was a lot of planning in creating all of the monitor imagery, and Triple-I did a great job on that. Spent a Iot of time on this shot. To accomplish the effect, what we did was get a 4-by-5 still camera and photograph Jeff Bridges in his position after he's been zapped, and then immediately moved him out and took another photograph, which was the background by itself. So, as you pull away chunks of Jeff, you see the background behind him, and then to put the laser and the grid and all that on top of him was basically the effects animation of John Van Vliet. When Flynn is de-rezzed and pulled into the computer, we go through one of the most interesting sequences in Tron, which is the real world to electronic world transition. This sequence was created by Robert Abel & Associates, primarily under the direction of Kenny Merman. It was a sequence which I had designed, knowing that the way that Robert Abel & Associates was making these computer graphic images with the Evans & Sutherland computers, and, really, using vector graphics to create this particular look, would give us a look that would be unique just for this transition. The three-space transition, the movement through all these binary bit patterns and this polygonal landscape, was done by making multiple passes through a traditional animation camera that was pointed at a high-resolution, vector graphic, Evans & Sutherland computer screen, and making multiple passes, frame by frame, using different colored filters, coming back, making multiple passes of rewinding other filters, until you finally end up with this, which seems to be solid objects. But it's really made out of lots of tiny, tiny lines put together to make solid blocks of color or objects. Oh, man, this isn't happening. It only thinks it's happening. When Flynn says, "This isn't happening, it just thinks it's happening," it's a key line, because it means that the reality that he finds himself in now, not even he can fully believe exists. And if anyone should appreciate and understand this alternate reality, it's him, and now he finds himself trapped in it. All right, now we see Sark standing on the bridge, and all of a sudden, he is enclosed with these shrouds of light as he begins to have his conversation with the MCP. The database for the MCP was a human figure that we had created at Triple-/ called Adam Powers and was originally on the Information International sample reel. And if you look at that sample reel, you'll see a juggler character who was juggling balls. Well, that face of that character is the face of the MCP. So, the first time that you see him, he is a polygonal drawing of a face. And that's basically the underlying database of the face. So, it's made of polygons. And those polygons, we play them out on the Triple-I computers as the line-drawing polygons, and made 12-and-a-half by 20-inch stills, high-con stills of those. But we created the mouth positions for the vowels and the syllables so that you could take these interchangeable transparencies and lay them down and make him Say, by whatever order you put them in, whatever you wanted him to say. 'Cause he was voicing a lot of different dialogue. Then those were backlit, and then we applied an effect to those line drawings of putting a steel mesh screen over the taking camera, and it made it have that much more, kind of, complex look. And then we also animated the exposure occasionally. Early on in the film when I started working with Steven, we did a lot of experiments to work out how these characters were created. The thing that we finally decided was that the characters needed to have this energy inside themselves. They are obviously in this electronic world. Now, these costumes were unlike any costumes anyone had ever created for a picture before, in that they were costumes designed to have effects treatments done to them. They were white with black drawing or black lines over them. All of the black elements on the costume were turned into circuitry which could be backlit and light could be pushed through there. We originally shot a 65-millimeter image of these people, live-action photography of them on these black sets. Then from that 65-millimeter film, we created some photo-rotoscope machines, which basically could project the 65-millimeter film down to large pieces of film, which were pre-punched with animation punches. This film was created by Kodak for us, and we would project down with these photo-rotoscope machines, which would hold this film into a vacuum frame and make a continuous tone positive print of each frame of the film. Then these continuous tone prints were taken to a light table, it was a vacuum light table, where they were contact printed to high-con film to make a number of high-con positive and negative images. So that you basically have for every character a large cel and you have high-con positives, negatives, and a continuous tone positive. Then these high-con elements were hand-inked and painted to isolate the circuits on the body, the whites of the eyes, the whites of the teeth and any other circuits that we wanted to treat as a separate exposure. The characters are more often than not... The live-action characters are shot on an all black stage. When there is a set, the set is also black, but is measured out to conform with what we're seeing in this artwork. So that if a character appears elevated in a shot, like this shot, there was an elevated platform for him to walk on, but it didn't look at all like the set. Then we would composite these actors over paintings, transparencies, and once that was done, we would add the light and the color separately. And to simplify it, you can describe it as a sort of perfect blend between live action and animation in that we took live-action film, photographed it in a way that we could break it down to individual frames, then blow up those frames into large slides or transparencies. And we had 75,000 of these, which seems like an appallingly large number, but it really isn't if you compare it to an animation film. And because we were at Disney, they were not overly swamped. That's an actual Frisbee, by the way, and those are actual Frisbees on their backs. We had a excellent Frisbee coach, Sam Schaiz. I like the fact that the deadliest weapon in Tron is a Frisbee. A Iot of effects animation in this sequence and in the film. And that is the animation that makes the glows, and as the Frisbee gets brighter, and you see the reflections of it on their costumes, all that has to be done frame by frame. This is hand-drawn animation that, although it is drawn, a negative is made of that, and it is placed over a light source and then re-photographed, and the ability of the effects animators was such that we were never waiting on the effects animation on the show. They always performed very well. It was never a problem. They did very few redos, and that's because they had had experience doing this beforehand, whereas everything else that we were doing, outside of the effects animation, was the first time through. So, that had a much tougher and steeper learning curve. In the holding cells for the game grid, those are backgrounds that are entirely hand-drawn by the background department, again using Rapidographs and line drawing and airbrushing and then turning those into high-cons. But those drawings are all drawn to match the actual physical sets, which were built so that when someone passes behind something, or leans on something, those are actual physical sets that were built. But again, the sets were just black on black. They're as if they were made of black velvet. Part of the interesting thing as a cinemagraphic problem that was presented to Bruce Logan was that he had to shoot, unlike anybody had ever shot before, sets that were entirely black with white line drawings and white characters running around on these sets. Bruce Logan's job in photographing these people was very difficult because, unlike most photography for most films, you try and get as much chiaroscuro in the picture as you can. You let there be a lot of dark and you create shadows and you create this moodiness, which a cinematographer takes great pride in. In this film, during the sequences in the electronic world, basically, he had to light them so that we could see as much of the costume as possible with as little shading as possible because all of the shading and all of that were done by hand by making different masks and airbrush elements that were used under these costumes in post. The ring game was an interesting technical exercise. The set itself, again, was black flock paper with the rings drawn on this paper with tape. The actors had to realize which rings were there and which ones were not as they acted out the sequence, imagining that they were hundreds of feet above the ground. One of the inspirations of Tron is the movie Spartacus. And there's quite a few similarities to the persecuted people who had to fight in the gladiatorial games. This game, of course, was inspired by Pong and jai alai. I think one of the interesting parts of Tron was the synthesis of new games that were created. The design, for example, of the glove that's being worn here, we took a traditional jai alai glove and then rebuilt it and made it out of foam, added other elements to it to give it a more technological quality, and then again, I put the designs over the outside of that to make it blend with the rest of the costumes. Shooting in 65 millimeter, from a director's standpoint, is a lot of trouble. The cameras are huge and bulky. The format requires an enormous amount of light to fill that negative, so if you are shooting Lawrence of Arabia or Doctor Zhivago and you've got lots of snow and big exteriors, it's fine, but in low-light-level situations, it's very troublesome. The depth of field is sometimes as little as a half an inch, and you find your cameraman is asking you, "Now, which part of the eye do you want in focus? "Do you want the front of the eye or the back of the eye in focus?" Or if the head of the actor is not square to the camera, they ask you the really insane question of, "Which eye do you want in focus? "I can give you the front eye in focus or the back, "but the other one's gonna be blurry." Now a lot of these shots where you see actors talking to each other and we're doing over-the-shoulders, the camera couldn't hold focus for the blow-ups to be made, and I had to shoot the actors on separate passes. So, in a shot like this, where you see all three actors talking to each other, it wasnt filmed that way. I filmed them separately and they were composited. And there's quite a few shots like this. Whenever you see them walking around and they're separated by more than a couple feet, those are all separate shots, and then the actors are composited. So, it's very difficult for the actors because not only do they not see the environment they're in when we're filming, all they see is an all black stage, but they don't even see the actor they're talking to. Forming of the Lightcycles, again, is almost entirely done by hand-done animation done by the effects animation department in creating the way that these cycles form around these characters. We built an object that the actor could sit upon, and it was literally a mechanical shape that was the seat and the handlebars, so he could sit down and it would thrust his arms forward and pull him down into that locked position. So that everything that he sits upon and touches, it was, again, drawn by the animation department, and not until you see the final completed cycle, which is actually a CG/ rendering of the cycle, is any of it done by computer. The Lightcycle sequence was done by MAGI. Their way of creating an object were to take basic geometric shapes, cones, cubes, spheres, cylinders, and make an object by collaging those particular pieces together and creating an object. And that's how the Lightcycle was created. All wide shots that you see are computer-simulated. All of the shots, other than the very tight shots of the figures inside the canopies, are computer-simulated. The shots inside the canopies are actually hand-drawn artwork of parts of the Lightcycles, and the animation that's happening over the Lightcycle windshields is hand-done animation to give them a sense of speed. But virtually every scene that you see of the Lightcycles is entirely computer-generated. And there's not even effects animation in those scenes. If there's an explosion when a Lightcycle hits the wall and a tire bounces across, I think those were basically all CGI. Syd Mead worked really hard on designing these motorcycles so that they would incorporate the characters. But if you look closely at them, you'll see that the second half of the bike is flattened and sort of two-dimensional, and that was done because the computers couldn't handle too many compound curved surfaces. So, we restricted those curves to the wheels and the windscreens, and then the rest of the bike was simplified. The ability to move the camera through 3D space with these computer-graphic-looking landscapes is just great. The Recognizers are a sort of King Kong. There's a little head on top of that gate structure... Suggestion of a face, but it, sort of, got lost. The Recognizers were created by MAGI-Synthavision. As I mentioned, there are graphic vector lines, red lines outlining all of these objects, the same way with the tank. The tank was another unique design of Syd Mead, who is a futurist, a fabulous designer. Once Ram, Tron and Flynn have escaped the Lightcycle grid and are off through the canyons being pursued by the tanks, we cut inside the tanks and see another example of a Syd Mead set that was built as a three-dimensional set, again with black background, and all of the elements on there graphically put on so they could later be treated. So the camera, you can see, is moving through scenes in ways that no physical camera or no model shot could possibly do. The animators that I worked with to create the choreography for all of the CG/ sequences were Bill Kroyer and Jerry Rees. But to communicate all this information to the computer technologists, the people that are sitting at monitors at that time, took a new language which we had to create. So, what we did was, first of all, we had to think of each sequence as a real physical reality. Not only would they draw the point of view that they saw as an animator that we would work out together, that was the story point that Steven wanted to make, and also the point of view that we wanted to take. But after we would draw the original storyboards in a traditional, kind of, storyboard manner, we would have to go back and draw a top view, side view and front view of the objects, where they were in time, where the camera was in time, and what the camera's point of view was. So, we really had to define everything to the CG/ technologist in a three-world, three-dimensional space. And that was the first time that that had ever been done. They must've gone right past us. We made it...this far. Now, all of this, this revolt, it's all being led by the user who's gone in the system, Flynn. The Tron character and Ram character, they would have toed the line and gone through the software the way they're supposed to. We'd better, Null Unit. Null Unit. Get the computer dictionary out. Look up "Null Unit." What does that mean? In this sequence, you can really see some of the flaws. I don't really mean the flaws, but the imperfections in the cels, little bits of dirt that pop on and off. Yeah, but they're few and far between considering. Yeah. Come on, you little bugger. Come on. Look at that. A lot of pops and a lot of glitches in there that we would always Say, "Well, that's what happens in an electronic world." When we started there were going to be no differentiations between the flesh tones and the rest of their uniform. But at a certain point they looked, well, not very good. So as a result, that added, approximately, 120,000 extra frames, extra elements to the shot, so it did grow in many aspects. The cave sequence where Flynn, Tron and Ram finally re-energize their selves with this liquid energy was a very interesting technical problem to solve here. In the sequence in the cave when the water is being handled by the actors, literally, frame per frame, rotoscope animation is isolating the water from the body so that it can be treated with a different filter and a different exposure. And again, this is an example of how light is used to portray motion or energy, as Tron drinks and you see his circuits light up and they become energized. The set itself was a complex geometric shape, which was designed by Peter Lloyd, and we built into this set, basically, water channels, and the water itself was reflecting light sources that we put in angle so they would reflect to the camera, and the water was in black tanks so that all we're really seeing are the highlights on the water. Yeah, but the biggest problem at that time was do we fill this with colored water or clear water? Had to do tests, you know. - Right. That was your problem. Do we put milk in there and make it purple? I think that what Flynn is surprised now, ironically, to see that there's parts of this mirror world that are more alive than he anticipated. So, it's not just the harsh computer reality, there's something living about it. It's a very complex shot, again, with all the elements. Probably about 30 different elements, 30 different separate exposures for each frame. Normally in a special effects movie, you get a very bad bottleneck effect in that all these things have to be composited through one or two optical printers. Now we have digital compositing machines. But by putting it into a manufacturing system like this, where it became like an animated film, we could use 14 or 15 animation stands, and we could use a slew of effects animators and ink and paint people to do all of this work simultaneously. As far as I know, we still have more shots with human beings composited into an artificial environment than any other movie. I believe there's 1,100 special effect shots in the film and 900 of which have human beings composited in them. And that number is just very, very large. Just the organizational task alone was monumental, not even considering the creative side of it. For every frame you would have an additional five to 15 cels that isolated the different colors and the different... We had body mattes, we had face masks, continuous tones. You made print backs on top of print backs. So, those 75,000 original cels grew to over half a million. I think we ended up with something like 600,000 cels, all of which had to be kept in order. We had to pull trailers, literally these large house trailers, kind of, industrial trailers onto the lot. We ran out of space and we ended up with 10 trailers that would house all these cels and had to be organized and sent over... 80% of them were sent overseas and had to be numbered and then painted and kept in order. At one point we thought if we had 1,000 scenes, and this was around Christmas time, the film was going to come out later that summer, and we had no idea of how we were going to get it all done in that short a period of time. And we thought, "Well, it's summer vacation. We have two weeks. "We'll get college students, 500 college students in a room." We really believed this might happen. We discussed this for about an hour and we Said, "You'd have 500 students in a room. "We'll teach them how to do inking and painting and rotoscoping, "and they only have to do two scenes each. "And so they do one scene a week. "At the end of that time, we'll be done, "and we'll just go and shoot them on the animation stands." It didn't work out. So, we brought on Arnie Wong, who was an animator. We put him in charge of supervising Cuckoo's Nest, which is a ink and paint service that was in Taiwan. Approximately 80-some employees in a single room. And what we did is we went through and we made a videotape of every situation and what to do in that situation. So that if an inker over there, who didn't even have to understand English to do this, could go to a TV monitor, roll to this particular problem and see exactly what you'd do in that situation. And then he was there to answer questions that were unusual. And the most interesting thing, and one of the things that I'm particularly proud of with this technique is that in spite of what a pyramid it was to build, we managed to get all of this post-production done in six to nine months. And that is using a technology that we had developed. It had never been done before and we developed it and used it on this picture and delivered on time. And that was only possible because of this manufacturing technique. It's interesting the computer animation iS the simpler part of the set. - Yes. Ironically, one of the things that was a creative philosophy that we enjoyed and were proud of was that we were taking computer animation and letting it stand on its own. We weren't trying to make computer animation mimic reality. And the job was then to make reality, the actors and the sets, look like the computer animation. We used to say, "Well, if you've got lemons, make lemonade." Everybody else, and certainly since this point, has been going nuts trying to make computer animation mimic reality perfectly. And I found that the limitations of computer graphics at the time were the most exciting thing. If computer graphics... If computer animation is no longer different from reality, maybe we've lost something in that. Certainly you gain special effects technology and you can do certain things, but it's the limitations, I find, to be the creative challenge. I think at the time we were using four computer animation companies... Yes. -... which were probably the only animation companies that existed in the country at the time. Yeah, I had been visiting some of these companies for two years before we started making the movie. Maybe even longer than that. And I used to show up at their doorstep and Say, "One day I'm gonna make this movie. "You know, we're gonna do this and this is gonna be great." And they'd say, "Yeah, yeah, yeah." I'd come by every six months and say this is really gonna happen, and I think they were more surprised than anybody else when we really did this movie. And they got to show their stuff. The way the de-rezzing effect was created, for example, when Ram passes away and he's in the cabin of the Recognizer, there's a combination of the original photography of the character, and then that is overdrawn with literally hand-done, line-drawing animation done by the animation department. And between that animation and light exposures, you can make it just, basically, run off, dissipate and fade away. Also, upon viewing this again, for so many years, you tend to kind of lump it all together visually in your memory and we forget, I forget, how much detail, how much layering of texture was put into this film. - Mmm-hmm. Ai! these shots are all completely storyboarded. Even the electronic world and all the simulated shots were all on storyboards. There must have been thousands of storyboards. Yeah, it was very detailed. Because rendering times in computer graphic imagery, the time it takes for the computer to draw each frame, are high. They're even high by today's standards. It takes sometimes as long as an hour or more for each frame of film. Probably the most complicated CGI images that were in Tron were done by Information International. The Solar Sailer hangar, the Solar Sailer, its formation, the walls of that environment, that's all CGI. As far as Cindy Morgan's involvement, she was very brave to get involved because a lot of actresses Said, "What am I going to wear? "You're going to put what on my head? "I've got to have a helmet and headgear "and wear all this spandex?" And that scared a lot of actresses away. Yeah, it was very hard to get anyone to take us seriously. You'd call people up, they'd come in for casting sessions, and Steven would do his best to present the film, and they'd look at you askance, think you were crazy. You'd run some video on them, and they just didn't believe it was going to happen. And as a result, it was very, very difficult. And I think that was one of the last major parts that was cast. Yes, it was two or three days before the first shot or something. Yes. - It was very close. And one of the people we tried was Deborah Harry. Right. We screen-tested Deborah Harry. The Bit was created by Digital Effects Incorporated, and we didn't have the time to choreograph a CGI Bit for every scene. So, what we did was created a series of stills that could be cell flopped, and these transparencies were created by Digital Effects so that the Bit could be rotating and have these different pulses in it, and then when it wanted to express itself, we flipped to the next sequence of stills, which would make it become more spiky or change its shape, and literally those were cell flopped and then flown around by moving the animation camera on the object to give it its motion from left to right or up or down or wherever it moved, we got it closer to you. That was all put in by moves on the animation camera, on these stills that were being cell flopped. These characters were very interesting. I especially liked the one that looked like a vacuum tube. Other programs... - Other programs and... ...in the system. The Recognizer sequence is another set that was built based on designs by Syd Mead. The interior of the Recognizer, as the interior of the tanks, was all a physically complex shape that the actors moved around on with white line-drawing vector material over the surface of it, and isolated animation coming back and colorizing and animating those elements. I think one of the most successful pieces of computer choreography in Tron is the whole Recognizer sequence, when the Recognizer hits a bridge and becomes multiple pieces and Flynn pulls them all back together with his energy and the choreography of the way those parts all fall back into place and tumble. The thing that people don't realize about computer simulation, especially at this time, is there were no programs that imitated the effects of nature on choreography. Every piece and every part of every computer-simulated object had to literally be choreographed frame per frame by an animator. When the Recognizer moves along and bounces off the ground floor and the pieces separate and then come closer together and have that real, elastic, rubber-banding kind of quality to them... Simple things in choreography... I mean, when an object goes around a corner, does it just swing around the corner or does it have back animation? Does it weave left and right? Does it back animate before it moves forward? Those are the things that the animators brought to this and that the computer-simulation people did a terrific job of interpreting.
28:30 · jump to transcript →
