By TREVOR HOGG
By TREVOR HOGG
A true mark of innovation is when one can look back two decades later and still be impressed by what was achieved given the technology, tools and resources available at the time. Say what you will about The Matrix Reloaded (2003) and The Matrix Revolutions (2003), but it is amazing the number of top-level visual effects professionals who have emerged from them, such as John ‘DJ’ DesJardin (Man of Steel) and Dan Glass (The Tree of Life). Adding to the complexity of The Matrix Reloaded and The Matrix Revolutions were the logistics of having both shot at the same time in sequential order and released six months apart from each other.
While John Gaeta (The Matrix) was the Overall Visual Effects Supervisor, the real and virtual worlds were divided respectively between DesJardin and Glass. “I was always fascinated by the real world as a subject because of this notion it had become a dystopian nightmare,” DesJardin notes. “The idea that we were going to go to Zion and see all of the aspects of it, not just where they live but the machines that keep it running and the big temples. Then the fetus fields and right down to Machine City. Those are my favorite things. The battle was a nail-biter to get that done. I recently came across a shooting assessment that I made and delivered to the producers for the how we were going to shoot the guys waging the battle in the APUs [Armoured Personnel Unit]. It was a big motion-control effort. But I will say, when I learned the story of the film, one of my favorite moments – and couldn’t wait to get a handle on to make – was when Neo and Trinity fly up above the clouds to get rid of the Sentinels that are clinging to the ship, and you get to see the sun for the first time in the real world. It’s a great idea, and I love the way it came out.”
“The street scene took awhile because we had 50 doubles for Agent Smith wearing printed masks; along with them we built mannequins from the cast of Hugo Weaving. The doubles were in the background, and in front of them were two mannequins that they could move left and right. When Hugo brought his kids on set, they were slightly horrified! There was 151 of dad there!”
—Dan Glass, Visual Effects Supervisor, The Matrix Reloaded & Revolutions
The Matrix Reloaded & Revolutions
Cutting-edge digital human technology was utilized to create the Burly Brawl and Super Burly Brawl involving the massive onslaught of Agent Smith clones. “I watched the Burly Brawl fairly recently, and the reason why it holds up is the split-screen work is all photography, and as you get into the more virtual work, for its time it was ambitious and pulled off some incredible things,” Glass remarks. “The Super Burly Brawl took the longest to shoot. There was a side thing where the Wachowskis didn’t want just rain. The raindrops had to be oversized. The special effects team was trying to figure out how to make this extra-wet, blobby rain. The street scene took awhile because we had 50 doubles for Agent Smith wearing printed masks; along with them we built mannequins from the cast of Hugo Weaving. The doubles were in the background, and in front of them were two mannequins that they could move left and right. When Hugo brought his kids on set, they were slightly horrified! There was 151 of dad there! And a lot of rain. It was grueling for us and I imagine as well for the actors.” The digital doubles of Agent Smith were not simply carbon copies. “You always try to bring some level of individuality so it feels more credible. The advantage of working in the Matrix for those movies was it was about a simulation, so it gave us some leeway,” Glass adds.
Spider-Man 2
“When I learned the story of the film, one of my favorite moments – and couldn’t wait to get a handle on to make – was when Neo and Trinity fly up above the clouds to get rid of the Sentinels that are clinging to the ship, and you get to see the sun for the first time in the real world. It’s a great idea, and I love the way it came out.”
—John ‘DJ’ DesJardin, Visual Effects Supervisor, The Matrix Reloaded & Revolutions
Set in New York City, Eternal Sunshine of the Spotless Mind (2004) explores what happens if a couple breaks up and go through a medical procedure to get rid of their memories of each other. “This was the best script I’ve ever read in my life,” states Louis Morin, who was at the time a Visual Effects Supervisor for Buzz Image Group and made suggestions about ways of erasing memories, such as having abstractions melt and disappear. “The producer said that Michel Gondry (Be Kind Rewind) didn’t want any visual effects supervisor on set. It was to be a free camera style of filmmaking and no lights, like Breathless by Jean-Luc Godard.” Camera tracking was hellish. “There was the ‘Pan from Hell,’ which is exactly the Breathless shot – and from the peculiar mind of a director who decided to flip the image so that the actor was walking into a flipped image of himself. We then had to marry the two together with tracking, morphing, and put in a telephone pole to help us out. The camera goes four times like that.” As Joel Barish (Jim Carrey) keeps going back and forth on the street, the details in the imagery begin to fade away. “We had to redo the whole store in CG to be able to erase everything step by step,” Morin adds.
Eternal Sunshine of the Spotless Mind
Then there was the preceding moment featuring a missing leg belonging to Clementine Kruczynski (Kate Winslet) and a falling car. “Michel wanted to have the first moment indicating that the memory of Clementine is being erased and he said, ‘Remove a leg,’” Morin recalls. “I said to him, ‘Nobody is going to notice that.’ We did it and nobody was seeing it. Also, Clementine didn’t turn her head at the right time, so we Frankenstein’d the shot by taking the head from a longer take, which worked well, but nobody was noticing the leg again! Somebody suggested that we could have a car fall down from the sky. Everybody thought it was ridiculous, but Michel said, ‘Let’s do it.’ We had to do that entire background and car in CG. At the end, everybody liked the idea, and it was powerful.” There were also subtle digital adjustments taking place. “Joel falls off of the sofa bed and reverses back into another shot of him on the sofa bed eating Chinese food with chopsticks. But the chopsticks weren’t working so we had to make them CG.” A major visual effect was the collapsing house. “At first Michel was talking about doing some optical iteration of the image. It wasn’t looking great. Then Michel asked, ‘Can we have a chimney collapsing?’ Upon seeing the test, he went, ‘Wow. Can we have the house collapsing?’ The house became entirely CG and was destroyed by using rigid body dynamics,” Morin says.
“When I talked to the people who were animating the shots of Spider-Man, I told them to imagine that he had his own cameraman, and the cameraman has to travel the same way as Spider-Man. As a result, you get a much more human or fallible version of camera operation that lends reality to it.”
—John Dyksta, Production Visual Effects Supervisor, Spider-Man 2
Departing from the normal routine from recruiting directors from within, Pixar collaborated with Brad Bird (The Iron Giant) to produce a superhero family adventure. In the process of making The Incredibles (2004), Visual Effects Supervisor Rick Sayre had to work out what he calls “open problems,” such as simulating the long, straight hair of Violet Parr, which was a key part of her character. “Violet is a teenage girl and her power is mostly defensive,” Sayre notes. “She puts up a shield or turns invisible because she wants to disappear. You will often see her with one eye. She is hiding behind her hair. That was important to Brad.” The existing hair simulation system had to be overhauled to allow for interaction. Explains Sayre, “One of our tricks before was to randomly connect some sets of hairs to other hairs of invisible springs; that would allow for a coif to retain its volume, but that technique doesn’t work with long hair because either the hair flattens out if there are no springs, or it looks like cement. We ended up embedding the simulation hairs inside of a volume, which was how they were able to couple their motions and collision responses to each other in a way that still isn’t as computationally expensive as every hair looking at every other hair. It’s as if they’re embedded in a block of invisible goo that is modulating these responses. We also used that block of invisible goo to infer some information that we used for lighting, shading and shadowing.”
A comedic sequence occurs when Fashion Designer Edna Mode demonstrates how indestructible her superhero suits are to Helen Parr by putting them through a series of extreme tests, such as a flame thrower. “It’s funny you mention that because it has an almost live-action approach,” Sayre remarks. “We hadn’t done a big effects film. What’s happening in the simulation chamber is done by a different team with a different set of techniques, even a different renderer, than Helen and Enda sitting on the other side of the glass. They are essentially on a set looking through the window at greenscreen where nothing is happening, pretending to react to all of this stuff that is done later and comped together. There was no way that we could have done all of that at the same time in the same system. The thing that caused our team the most headaches were the super suits, which were tight-fitting and caused simulation stability and collision fidelity issues. Because Edna is so amazing, her super suits have special visual properties. They’re shiny and had these surface characteristics where we would see these rendering artifacts coming from the guts of how Catmull-Clark subdivision surfaces got rendered in RenderMan of the day. At some point, we were using a different kind of subdivision surface or a loop subdivision and then reprojecting it. The super suits that Edna doesn’t make, like Syndrome, were easier to deal with because they’re more like regular cloth.”
Being able to make the protagonist crawl walls and swing through the air from buildings in a believable manner, and giving the antagonist mechanical tentacles that have a mind of their own, were a couple of many challenges John Dykstra faced as the Production Visual effects Designer on Spider-Man 2 (2004). “Sam Raimi (A Simple Plan) wanted to use as many practical elements as he could, so we pursued Doc Ock that way,” Dykstra remarks. “It is tough for puppets to defy physics because they are in the real world and want to work in real time. We went through and prevised the entire sequence, and did computer-generated imagery for those shots where we felt puppeteering was impractical. The arms of [Doc Ock] were a digital endeavor from the get-go. The art department worked with us in terms of the design, and we worked with the vendor to figure out the animation look in regards to the speed and mass and how the arms worked. We were defying gravity a lot in Spider-Man 2.” Realism was built into the CG camerawork. “When I talked to the people who were animating the shots of Spider-Man, I told them to imagine that he had his own cameraman, and the cameraman has to travel the same way as Spider-Man. As a result, you get a much more human or fallible version of camera operation that lends reality to it.”
The Incredibles
An extremely hard shot was the tight closeup of Doc Ock falling. “Trust me, that was torn from the artists’ hands by the time it was put into the film!” Dykstra laughs. “The idea was to have a moment where we actually featured a CGI character with emotional content, and the challenge was to do it in a way that you would be convinced that it was real, especially when you’re doing something with a real person. I suppose Alfred Molina could have done it, but I don’t imagine he could have been underwater for so long!” CG skin is always tricky. “Things like pores and inconsistencies in surface reflectivity often contribute to the complex and somewhat visually noisy thing that is human flesh.” In theory, Spider-Man is an ideal CG character because the material of the suit has a smooth matte finish and no hair or fur has to be simulated. “When there is an absence of natural phenomenon, you end up questioning the verisimilitude of what you’re looking at. It was important to improve upon the specular nature of the suit, and the way it wrinkled had variations in the texture of the surface of the body while it was in motion.” Spider-Man 2 occurred during a transitional period from analog to digital solutions. Dykstra states, “One of the things that we had to work on in that era was including world noise. We had to take the perfection of the computer-generated model and haul it back into the realm of the real world. Stuff like film grain and how it was reacting. Was it out or in focus? We had to study that to figure out how to apply it to the shots because it’s the filter through which you see the world.”