Follow DavidBennun on Twitter

order British As A Second Language from order Tick Bite Fever from

Movie Visual Effects

[The Observer, Life Magazine, 1999]

ONLY HALF an hour after I pass Darth Vader in the corridor on the way out does it occur to me: that was the Darth Vader. Masked bodybuilder with the voice of James Earl Jones or no, Darth Vader is principally a figment of George Lucas's imagination. Most of us, if we found such imaginary figments loitering in our hallways, would be marked up as prime candidates for the booby hatch. “This is Darth. He's a seven-foot asthmatic leather freak. And have you met Harvey the rabbit?” But not George. At George's sprawling Industrial Light & Magic operation in San Rafael, a ferry ride north of San Francisco, the manifestly unreal is made manifest. So if any Darth Vader can be called the Darth Vader, that was it. Or him.
  It or him? This confusion is what George, and ILM, and the entire visual effects industry depend on. The fact that Darth is mounted motionless and for all I know stuffed atop a modest plinth makes him no less alive in my recollection. Which is, it's true, the recollection of a small boy circa 1977. Two decades later, Darth's successors face a far more sophisticated audience; equally willing to suspend their disbelief, but with eyes primed to spot the join where fantasy meets. . . further and more fantastic fantasy. The supernatural, the super-real and the super spectacular continue to soak up most of the time and effort spent on special effects. And that's a lot of time and effort, a thousand employees' worth at ILM alone.
  In the Nineties, the Special Effects Movie has become a genre of its own. You know what to look forward to. Large things will explode. Smaller things will behave in unlikely ways. Other things which have no business being there at all will stubbornly insist on being there a lot. What you may not realise is that nowadays digital effects lurk where they are least expected. Art movies. Costume dramas. The kind of cinema whose audience would rather swallow its own eyeballs than sully its retinas with the likes of Armageddon . ILM helped out Woody Allen on his last two features, Celebrity and Deconstructing Harry . Elizabeth had special effects up its Tudor wazoo. Even the emerald Oirishness of Waking Ned was touched up by hands more accustomed to fashioning bulbous clouds of orange smoke.
  John Berton is clearly a little sorry that he doesn't get to work on films like these. Berton is too good at the big stuff. He was once employed at the Ohio Supercomputer Center, on scientific visualisation films: “I had to choose between helping man understand the universe and scaring teenagers. I chose scaring teenagers.”
  Berton joined ILM as a computer graphics animator on Terminator 2 , and has since performed ever more substantial functions on Jurassic Park, The Mask, Men in Black and The Mummy. There is something of the visual effect about Berton himself, in the way his eyes and his grin bulge outwards from an improbable ginger barnet; but this is simply the way his huge enthusiasm for movies reveals itself. With Berton, the switch is set permanently to Enthuse.
  “Visual effects have become intertwined with all of film making now,” Berton enthuses, “with popular films that aren't seen as being blockbusters, things like Elizabeth and Waking Ned. It's very difficult to recreate a scene with a camera moving close into a group of protestants being burned at the stake, so you do it with the highest level of technology that you can bring to bear. A phone booth falling off a 350-foot cliff and hitting the beach - you could shoot it, but it would be a lot more difficult to get exactly what you want, so you use computer graphics.
  “Look at Forrest Gump,” he enthuses some more, “and you see how the techniques can create images that don't jump off the screen at you, but at the same time are incredibly unique. Many people to this day believe that Gary Sinise has no legs, and that's because of visual effects, because they are so convincing. Every film that's out there now is using this visual effects work in some form or another. It's become a part of editorial, a part of composing the frame.
  “The Phantom Menace is a tremendous example, where every shot is composed in the computer. The layers of battle droids and tanks and explosions and shields and grass fields, all these things are under the control of the film-makers. And now something that was developed as a method of creating amazing things on the screen to serve the needs of entertainment films, is actually becoming an inherent part of the way that we make movies.”
  To illustrate Berton's point, you need look no further than the current hit Fight Club. A lot of things have been said about Fight Club - that it's vicious, amoral, downright immoral, neo-Nazi, vile, an affront to society. What hasn't been mentioned is the brilliance, or even the existence, of its visual effects. Civilisation could not have been brought down without them. Fight Club's visual effects supervisor Kevin Haug, when I track him down, is working at the smaller of Warners” two Los Angeles movie lots, on a film called The Cell, which he summarises as “Silence Of The Lambs, MTV-style.” The quietly-spoken, khaki-clad Haug does not strike me as an agent of anarchy or a closet Hitlerite. Not unless his skill at subterfuge outweighs even his gifts with a digital paintbrush.
  “It's not really that Fight Club is an effects movie at all,” he murmurs. “In fact the studio, Fox, never really recognised it was an effects movie. I don't have a title credit. It was only 25 effects shots at first, but it ended up being about 80. Although most of them are what I'd call invisible. Or ‘body and fender’ work. There was a big ugly scratch on Brad Pitt's face in one of the takes [Director] David Fincher liked best. That had to be taken out or reshot. It was one of those calculations - how much would it cost to reshoot that, and how much to fix it. The scratch took “em weeks to fix but it still worked out cheaper.”
  Haug is justifiably proud of sequences in Fight Club which don't feel like VFX, although when you stop to think about it - usually a long time afterwards - they could hardly be anything else. One is a writhing, polymorphous sex scene which features Brad Pitt and Helena Bonham Carter. Or rather, it doesn't. “It's animated, CG [computer generated] people, all naked, all close-up, all believable.” So believable that a Variety review gave the director of photography credit for it, despite its never having been photographed.
  “Our naked bodies were covered with white dots,” Bonham Carter has explained. “For 12 hours, Brad and I assumed positions from the Kama Sutra. The computer did the rest. It made us do anything it wanted to.” Another fancy manoeuvre follows the camera around Edward Norton's kitchen into places where it can't possibly go. “You fly up over the top of the stove,” says Haug, “and go down underneath the grill and fly around the burner, and then you fly up and behind the refrigerator, and then out from underneath it - all of which is possible in CG because the camera has no size.
  “The problem with CG is often you're chasing that photo-real look, and it ends up looking too perfect, it doesn't have the grain, the peccadilloes of the dyes that are on the film. This technique is off at a tangent from where the rest of the CG community is going - if you will, a sort of third-world way of doing it. You go out and shoot pictures, and they are photo-real by their very nature. You give them to the CG people, who use the images to make the digital models, and then map the images back onto them. It's very much the same way that satellites make maps of the earth. It's the same exact technique that's been around since the First World War. We did actually test to see if it wouldn't be cheaper to build oversized sets, but it just became a monstrous thing. A full crew working on set all the time is more expensive than the CG.”
  Which scotches one myth about visual effects; that they exist only for their own sake, swallowing vast budgets that might be better spent on, say, a dozen arthouse films with nothing startling in them at all, apart from the heroine's propensity for taking her top off every few minutes. Effects can cost a lot; but then movies cost a lot. Faced with a choice between a computer program and a giant kitchen, the bean-counters who purportedly run the movie business will tend to favour the former. Directors might not always feel the same way, but that's their lookout. It is, after all, the director's job to impose his own perspective on the movie no matter which tools - or which vendors, as special effects contractors are known - are to be used.
  “It's happened in the past,” says Haug, “with big facilities like ILM, the supervisor will have a stronger vision than the director does about the shot, and it'll end up looking like his vision. And it'll be beautiful. It just won't look like it came from the same movie.” He cites Escape From LA - not an ILM title - as “the classic example of the way not to cast vendors. Every vendor, lacking any real strong input from the director, did what they thought best. Some did better, some did worse, and every time you changed a shot you were in a different world.”
  “Sure,” grins Berton, when I ask him about this, “there's a danger that effects houses will dominate a movie - if you let “em. Technology is an incredibly inviting trap, and it will lead you where it can go, if you let it. If you don't fight against the pathways to which the technology is best attuned, you end up with boring stuff. You can never say, well, this'll be easier for us to do. That's crazy. If you never think beyond it you're going to be controlled by the technology and you're never going to have good pictures.”

WHEN movies started out, it was enough that they moved. From the very beginning, they were a visual effect in themselves. We tend to forget this when we put special effects into its own category. The distinction is merely a technical one. All films are illusions. That's the point of them. Yet we don't consider it “unreal” for an actor to speak prepared lines, or a director to manipulate time and space. Not the way we consider it unreal for a big lizard to tread on New York, at any rate.
  Modern special effects have followed the same path as more traditional film-making, starting off as spectacle, then being harnessed in the service of plot and atmosphere. But special effects are by no means a recent development. Georges Méliès used them, with outstanding results, in his classic films of the 1900s, of which the most famous and most copied is 1902's proto-psychedelic Le Voyage Dans La Lune (translated, with prescient aptness, as A Trip To The Moon.) Méliès excepted, the earliest special effects wizards were exactly that: magicians who transferred their illusions to the cinema screen. Their lineage can be traced back at least as far as the 1790s, when the conjuror Etienne Gaspard Robert, who billed himself simply as Robertson, staged phantasmagorical magic lantern shows across Europe.
  This is significant firstly because it suggests that visual effects are not a new and corrupting influence in film, but a tradition which predates even photography; and secondly because current computer based effects are heavily indebted to that tradition. “I'm particularly fascinated with the history of film as illusion,” says Berton, with, if possible, even greater enthusiasm than before. “I've always been much more interested in what you can put up on the screen that you can't see in reality, whether it's something as simple as ghosts or something as concrete as spaceships in outer space. I look at Méliès as the beginning of what fascinates me about cinema.”
  Berton is also a fervent admirer of Ray Harryhausen, whose stop motion animations in films like Jason And The Argonauts and One Million Years BC took visual effects further than anyone would, or could, until Lucas came along. Pride of place on the wall of ILM's lobby goes to a huge, signed still of a Harryhausen monster (signed, that is, by Harryhausen; the monster has a distinctly illiterate cast to its features.) When Harryhausen gave a talk to a packed house at ILM, Berton was delighted to discover that a mob-handed CG fight scene from The Mummy echoed Harryhausen's own approach. Technology may advance, but it's of little use without human ingenuity; and in visual effects antiquated methods are often brought to bear on new problems. Witness Kevin Haug's Great War approach to the voyage around Edward Norton's white goods. Or for an even more extreme example, take a look at The Matrix.
  The Matrix received an unfairly hostile reception in the UK, from critics who may have taken its mystical pretensions a little too seriously. True, Keanu Reeves trundles through the movie like a stuffed beagle on castors, but only when he acts. Thankfully, he is called on to do very little of that. What he is called on to do a great deal of is hover through the air, up walls, between bits of flying lead and so forth, in a breathtaking mode of temporal mangling known as “bullet time.” The Matrix's action scenes are nothing short of astonishing, and the outfit responsible is Manex, a rising young VFX company located in the former Alameda naval base, on the other side of San Francisco from ILM.
  At ILM, the first of the big-time effects houses and still the yardstick by which the others measure their success, everything is orderly, timetabled with digital precision. Surprises are frowned upon. At Manex, all is flux. My arrival, at the agreed time, seems to be no more or less expected than that of a hang-gliding yak. Unfazed, the desk staff track down Dan Piponi - “he's around here somewhere” - digital supervisor on The Matrix. Piponi's background is in mathematics. He explains how scenes were shot using “volumetric capture', with still cameras, up to 180 of them, arranged around the action via laser targeting and fired by an electronic system designed for that purpose alone. Thus the viewpoint circles the slow-mo action in a smooth but dizzying fashion which a movie camera could not begin to duplicate.
  At its core, this technique bears a remarkable similarity to the one devised in the 1870s by Eadweard Muybridge for the purpose of tracking motion in horses. He would arrange his cameras along a racetrack, with strings set to trigger the shutters when snapped by the animal's legs. This was effectively the birth of motion photography. Muybridge came up with his zoopraxiscope, a revolving disc, as a method of turning his photographs into a projected moving image. Piponi - like Muybridge, British born but resident in America - used a computer to align The Matrix's shots,

equalise the lighting and, painstakingly, fill in the missing frames.
  The backgrounds were digitally modelled with a technique created at UCLA's Berkeley by a team including technical designer George Borshukov - a technique not unlike the one employed by Haug on the kitchen scene in Fight Club. “They used the campus,” says Piponi. “All they had to do is take 20 photographs and they had the software to build a CG model and paste the photographs onto the buildings.” This method was applied to The Matrix's backdrops, and is now being used by Manex to create entire digital cities of photo-realistic quality.
  “We wanted to use that technology for the first time in a movie,” Piponi says, and this urgency is understandable. As Kevin Haug remarks, “You always want to be the one who's stepped out further. The Matrix was a big impetus for the sex sequence in Fight Club. Frankly, one effect that they did turned into Gap ads before they ever had a chance to finish the movie. We didn't want that to happen to us.” After all the talk of “bullet time” and “jello time”, this white-background look was sardonically labelled “dead time.” “There were so many ads coming out with this effect in,” Piponi grumbles. “But ours is a lot more complex. I mean, their stuff probably only took a couple of weeks. Ours took a year and a half to do. Really damn hard.”
  In a business this competitive, getting beaten to the punch is an occupational hazard. A company which fails to fanatically pursue new methods might as well give up. At ILM, John Berton is characteristically excited about a breakthrough in “procedural systems”, whereby a large number of highly complex synthetic effects - the duplicated passengers in Titanic, for example, or the bandages flapping from The Mummy's arms - can be devised on a computer relatively quickly, then tweaked individually. Manex have their photo-real digital landscapes. But the most eagerly trailed special effect of all, the one which is presumed to hold the profoundest implications for movies as a business and cinema as an art form, is the artificial actor. Once the technology has advanced far enough, we're told, then special effects will be able to dispense with the living and resurrect the dead. The sky, in other words, is falling; no actor's livelihood, and no star's memory, is safe.
  Kevin Haug is sceptical. “The photo-real actors thing is still a way off,” he believes. “You could do it now, it's just a matter of whether or not anybody would pay you to do it.”
  Dan Piponi points out that stunt-work already uses virtual humans, or at least parts of them. “People have been replacing heads in movies for some time. Stuntmen doing dangerous things have actors' heads on them. I won't mention which movies, because one guy claimed to the press that he did his own stunts. But my opinion on digital actors has changed very recently. I've absolutely no doubt we'll be doing it within a couple of years. I do think that people will not be able to tell.”
  Piponi's boss John Gaeta, visual effects supervisor on The Matrix, takes a broader view: “Most people know about visual effects from very commercial, high-profile pictures. But visual effects need to be looked at as a movement in art. When things seem to appear, it's because as a collective consciousness it was about time for things to come out. With virtual humans, you could make a perfect shell inside of the next few years. But that is a far, far reach from actually having any projection of character or personality. There are as many - or as few - animation directors who could actually pull off directing the personality to be placed within such a thing, as there are good actors. I get asked often, in a panicky way, ‘Are actors going to get replaced?’ No. They're not. It's going to be a handful of people in the community who will be able to deliver good examples of this.”
  John Berton's thinking runs along similar lines. “Computers are finite systems and reality is not. I've always been a great believer that the human spirit is way more complicated and way more important than anything you can calculate mathematically. It's a fairly deep philosophical argument, but what it boils down to is that even a really talented animator working with a computer system is probably not going to be able to create a performance that's the equal of a great actor. I think that perhaps one day we will create a tremendously convincing digital actor, and people will be amazed that such a thing is not real. But it still won't be Humphrey Bogart. It still won't be Marilyn Monroe. It still won't be Laurence Olivier. Because these people are who they are, and their particular traits, talents and skills are something which cannot be simulated. Maybe we will create a Laurence Olivier performance that couldn't be done by Laurence Olivier, because Laurence Olivier's performance needs to be mapped onto a three-headed alien in Star Wars Nine or whatever. That's where it's at. It's not about making Marilyn Monroe. It's about making something that's as good as Marilyn Monroe that isn't real. As far as how that relates to putting actors out of work, it doesn't. I think actors are always going to be. . .”
  Out of work?
  “Yeah, heh-heh-heh. But remember, computers aren't very powerful, all things considered. You have to constantly come up with ways of taking the pathetic amount of data that we can use to create pictures and forcing it to create something that's going to work for the film. It has to serve the movie. It has to serve the story, it has to serve the image, it has to serve the concept of each film. That's the trick.”
  In a business based on trickery, it's surprising to hear that the trick is so wholesome. Still, Berton's point is a fair one. The visual effect is a creature of its master. It will be as good or as bad as Hollywood makes it. If you don't like it, there's always The Blair Witch Project. And if the people who made that don't hire a visual effects supervisor on their next project, I'll eat their woolly hats.

Christopher Townsend of ILM, sequence supervisor on The Phantom Menace, talks us through the sequence where Obi Wan Kenobi and Gui Gon Jinn meet Boss Nass, king of the Gungans, in the underwater city of Otoh Gunga. No, really.

“THE first thing we had to do was create the impression of water. We rendered some spheres with very highly detailed textures. We had some simple geometry for the actual ribs on the outside so when they moved in the water you could have a sense of parallax and therefore a sense of distance.
  “Boss Nass himself starts out as a digital wire-frame model, then a simple skin is put on to create what we call the plastic render. Once the model has been created it's given to an animator. We use a piece of software where you can create various shapes and morph between those shapes. At the same time the model's given to a view painter who puts the textures on. ‘Bump’ can be added, and wherever the view painter has painted white, it shows up as wetter area. Boss Nass had moisture on his teeth that would glint, and bits of spit that would fly out - we went and shot some goo being dropped off somebody's hand for that.
  “Then both the animation and that view-painted model are handed off to a technical director, who effectively ‘skins’ it and lights it and composites it into the scene to make it look as if it's really there. We had problems with this character, who was wearing a very thick, heavy cloak. So we had to do simulations of cloth - a very heavy geometry, to make it sag and wrinkle and stretch in all the right places
  “We've got three dimensional fish swimming by, we have reflections of Boss Nass in the bubble at top left. A lot of it is not consciously seen by the viewer, it's sensed, so you think, yeah, it looks like it's underwater.
  “The actor who was used for the voice and the reference was Brian Blessed. They shot him on stage with Liam Neeson and Ewan MacGregor, so the animators and technical directors could watch how he moved, his facial expressions, the way his jowels moved. Effectively it is Brian Blessed that we're seeing on screen. If you think back to other movies, one of the classic examples done here is Dragonheart, with Sean Connery, and if you imagine Sean Connery as a dragon, that's how he would look. It's the same with this.”

Digital Domain, the special effects house behind Apollo 13, The Fifth Element, and much of Titanic and Armageddon, created the title sequence of Fight Club, in which the camera moves from the depths of Edward Norton's brain out through his head and beyond. Digital artist David Prescott talks us through it.

“FROM the production side, they had the idea of the areas of the brain they wanted to see. They wanted to see dendrites, the cortical bone and things like that. We wanted to keep everything 100 per cent to scale. For reference we used medical illustrations, and we also went down to the University of California in San Diego, and saw how they do electron microscopy, and how they do the scans. We modelled our stuff to that.
  “We had to figure out how we could fly through this intense, dense forest of geometry. Some of it is modelled with math and with noise. It's a system called Iso-surfaces, it says, define a noise field in three dimensional space, like a sinusoidal wave, and whenever the wave goes through zero, I'm going to put a surface there.
  “We did what we call digital location scouting. We built this massive, massive piece of geometry, it was absolutely huge, about 2.5 million polygons, somewhere around 500 times bigger in all direction than the actual scene we see. In the same way you would roam around a cave with a camera looking for a good shot, we went into this geometry and drove around it for half a day, basically looking for a nice location, and once we found it, we basically clamped it off and deleted the sundry areas.
  “Another aspect was, we used L-systems, which stands for Lindenmayer systems, after a mathematician from the sixties who defined an algorithmic rule set for the generation of plants. We grew the dendrites this way.
  “And then the lighting was interesting, we did what's called an SEM look, Scanning Electron Microscope, it looks like a reversed x-ray. Obviously it's Hollywood, so we wanted to add a bit of blue light in, and to add in some slight noise in the colour.
  “Then we go into the hair follicle, and we pull out. There was a ton of really, really close-up photographs taken of Ed Norton's face, and a 3 D scan of his head. Judith Crowe supervised this part. Once they were out of his hair, they created a shader for the tiny little pores of skin, and all the hairs, and the sweat bead that rolls down, all the way along the nose.”

Dan Piponi of Manex, digital supervisor on The Matrix, explains how the fight scenes were created

“WE STARTED with previz - pre-visualistion - using crash test dummy-type figures to map out the camera positions for the volumetric capture [see main piece]. There were anything from 120 to 180 35mm still cameras. Then we brought the actors in on wires to act out the movements, and set off the cameras.
  “The next part was to align and smooth out the pictures, paint out the wires and fill in the gaps in between frames. Some scenes were stretched from only eight stills. It wasn't that there was any technical limit to the number of cameras we could use, it was just the maximum number that could physically fit into the space available. If somebody's hand moves behind something, the computer doesn't understand that it's been hidden. Wherever it went wrong somebody has to laboriously paint it on the computer. That was just the foreground. It took about a year to do for a good shot. Then we had to apply motion blur. Otherwise it just looks like a bunch of pictures flashed up one after than the other. It makes you sick, and it's horrible.
  “Then we took photos of the set, from which we created a 3 D computer model, and mapped the photos back onto that for a photo-real look. The entire set is CG. Even the lighting is virtual - the set never looked like this. Then the scenes were combined with the set. Almost everything you see is CG, even the helicopter that appears in the rooftop scene.”


“CG done well is better than real. That's the point. Otherwise you wouldn't do it, you'd be stupid to do it.” - Kevin Haug

CG: computer generated
DISPLACEMENT MAP: Method of applying varying contours to a digital surface
MINIATURES: solid models, old-fashioned but still used when convenient
POLYGON: the building block of 3-D modelling geometry, like a pixel in 2 D
PREVIZ: pre-visualistion; the planning stage of a visual effect
PROCEDURAL SYSTEM: computer program which allows large-scale effects with many moving parts, such as a flock of birds
RAY-TRACING: virtual lighting technique that follows light on its predicted path
SHADERS: digital paint and texture applicators
SHOW: a movie, or that part of it to be worked on by a visual effects team
TRANSLIGHT: giant transparency used as a backdrop
VENDOR: visual effects contractor on a film, often highly specialised. The visual effects supervisor may hire several of these for any one film
WATER: a continual visual effects problem. A good water simulation is a source of great pride
WIRE-FRAME: 3-D modelling outlines constructed from digital “wires”
PHOTO-REAL: the visual effects Holy Grail; CG material of photographic quality and appearance

All material on this site is copyrighted to David Bennun and may not be reprinted or reused without permission.

Back to Features
Back to Main Index