Weta: The masters of illusion
As the final Hobbit chapter looms, Nikki Macdonald goes behind the scenes with Weta Digital's visual effects wizards to find out how far computer animation has come since the original Gollum.
Real apes don't cry. At least, they don't produce tears.
So, when Weta visual effects supervisor Dan Lemmon wanted to make his digital apes cry believably, he called a friend who is an eye surgeon.
Together they filmed Lemmon putting droplets of ultra violet fluorescent dye into his eyes to record the shapes the liquid made and the way it pooled around the lids.
"We wanted to make sure that as tears started to well they welled in the right way," Lemmon says.
"What he didn't tell us is that it stings like crazy. You want to cry."
It's just one tiny example of the extraordinary lengths Lemmon and his Weta Digital team go to to ensure their digital characters are believable, whether it's The Lord of the Rings' Gollum, Planet of the Apes' hero ape Caesar, or villain ape Koba, or The Hobbit's dreaded dragon Smaug.
Much of the latest movie in the Apes franchise, Dawn of the Planet of the Apes, was shot in a Canadian forest in the rain. To ensure the digital apes looked at home in the real-life dripping forest, the Weta animators plotted the path of the real raindrops and twitched the digital fur where the droplets fell. Where the rain hit the chimp's bare muzzle, they simulated trickles of water tracking slowly down the face.
"You don't necessarily see all the detail," Lemmon concedes.
"But you feel the detail.
"You have to be a little obsessive, I guess. I think I probably have some kind of condition and I think a lot of the guys I work with do as well. The women as well. They all work incredibly hard to get their work as unbelievably real as possible. Or I should say believably real.
"We ride that line between making things grand and fantastic and spectacular, but still feeling authentic and real and believable. It does require a lot of analysis and attention to detail and breaking down what it is about the real thing that makes it look real.
"Where can we make it different to make it look fantastic and where do we have to stay true to what makes it look real?"
Behind the scenes at Weta is an army of real-life elves who are masters of illusion.
There's no room for half-arsed jobs or the kind of person who swings easily from perfectionist to slapdash. Or the kind of person who thinks it's anything other than perfectly normal to arrive back from a week in Los Angeles at 11am to deliver a presentation to media in Sir Peter Jackson's personal cinema two hours later.
Because there Lemmon is in media photos, being honoured in LA with actor Andy Serkis at the Hamilton Behind the Camera Awards. And yet here he is, in a director's chair at the front of the sumptuous Camperdown Cinema, talking apes, dragons and blue cat people.
It's fitting that Lemmon was honoured alongside Serkis - the relative unknown brought to New Zealand as a character actor to voice Gollum in the first chapter of The Lord of the Rings. Serkis has helped drive the development of the motion capture process that lies at the heart of Weta's quest to make the fantastic believable.
The technique, which Weta has helped pioneer from Gollum to Avatar's blue Na'vi to Smaug, involves actors portraying digital characters on set alongside real- life actors and animators translating that performance, transferring their smallest gesture or facial expression on to their digital double.
Acting in weird, grey skin-suits with a pox of white spots covering the body and face has become resolutely mainstream, with Serkis now directing Jungle Book: Origins which is expected to feature performance capture from A-listers Cate Blanchett, Christian Bale and Benedict Cumberbatch.
But does simply mimicking an actor's performance take the romance and fantasy out of traditional animation?
IN THE kitchen adjoining Weta's Miramar motion- capture studio are relics of cinema history - movie posters for the original King Kong and Son of Kong.
But there's nothing vintage about the technology in the gymnasium-like space next door.
Isaac Hamon's grey skin-suit is fitted with 53 reflective markers whose movement cameras will plot from all angles. Like a wind- up toy, he raises his arms and wriggles his hips, helping the computer calibrate his form into "generic man" - the digital humanoid that will be used as a reference for the chimp Hamon will become.
As Hamon dons arm- extending crutches and effortlessly slopes up over the table and along the green line taped on the floor, his movement is captured live by the 84 cameras watching from all sides. Within seconds, the bank of computers behind transforms him into an ape on screen. Before our eyes, generic man shrinks and his back becomes bowed in a quick- fire reversal of the classic ape-to- man evolution drawing.
But don't be fooled into thinking the computer is doing all the work. It is Hamon who turns man into ape, with his easy four- legged gait and the distinctive shoulder roll of a genuine chimp. It's an awkward, full-body workout. Believe me, I tried.
The hero actors also get helmets with face cameras recording their facial expressions and the movement of the white dots painted on to their faces. As you'd expect from an outfit obsessed with perfection, both the helmet and the mask used to stencil the dots are individually moulded using a 3D-scan of the actor's head.
Somewhere in the bowels of Weta is an entire library of Andy Serkis heads through the years, from the original Gollum of 2001 to The Adventures of Tintin's Captain Haddock to The Hobbit's Gollum 13 years later.
But whether it's talking apes or fire-breathing dragons, the work begins long before the actors get suited up. The visual effects team builds digital models of the creature's face and body, based on real musculature. For the Apes franchise, that meant hours studying chimps at Wellington Zoo.
When Cara the chimp underwent ear surgery, animation supervisor Dan Barrett got to watch.
"I stroked a chimp, which was pretty cool."
Motion capture had its origins in early Disney movies, when animators used a rotoscope to draw over filmed footage of ballroom dancing to get realistic movement. For Weta, the real revolution began with sticking Serkis in a grey suit on set to play Gollum, Lemmon recalls.
"The thing we noticed was that when he sat in a scene with the other actors, the other actors got better as well. They had somebody they could play off.
"In the past for digital monsters, you might have a tennis ball on the end of a stick and wave it at the actors and go, 'This is a monster, be afraid'. In a character like Gollum, who had a lot of subtlety and conflict in his performance, playing to a tennis ball on a stick wasn't the best way to get good performances out of everybody."
For the original Gollum, there was no whizzy technology - animators simply used Serkis' movement and expressions as a guide. Since then Weta has refined the process so it can now accurately capture motion - and emotion - on location, as it did in the Canadian forest for Dawn of the Planet of the Apes.
Barrett says the ultimate would be to stick electrodes on the actors to measure actual muscle movement, instead of just plotting how the skin or body moves in space. But for now that's just a crazy dream.
"I think for a long time there is going to be artistry involved. There is no magic button."
NOWHERE is that artistry more obvious than in the facial modelling department, where the feelings conveyed by uniquely human expressions are translated on to inhuman faces.
It was Avatar director James Cameron who coined the term "performance capture", as distinct from "motion capture". Performance capture, he said, was about capturing not just the motion of an actor's performance, but also the emotion.
Lead facial modeller Alessandro Bonora watches as a three-headed Serkis rolls his eyes, squints, whoops and snarls to the camera. His extra two heads are reflections in mirrors capturing his face from different angles.
It's Bonora's job to design the digital ape's face to reflect the actor's features and to translate Serkis' emotion into ape-like expressions. When the actor playing Koba changed between the two ape movies, the ape's digital face was also adapted.
"It's like when you go to a cartoonist. The more prominent your features, the easier it will be. If you are a very Ken-doll, really handsome guy, the cartoon is going to just look strange.
"That's why performers like Andy [Serkis] or Toby [Kebbell, who plays Koba], when you see their faces in person, there's lots of features you can exploit.
"For very human expressions, we have to look for ways to translate that on to the ape to evoke the same emotion. There's a bit of artistic freedom on that but there's a lot of research on apes as well.
"We like to keep it one-to-one, because the actor has put so much soul into the character."
There are 240 different muscle controls to contort Caesar's face, including four just to tweak the eye creases. Much of Bonora's job involves adding imperfections to make the character look more real.
Barrett admits that having to follow an actor's performance exactly can stifle an animator's creativity. "Being invisible is almost the goal."
But the benefit is that the actors become the stars, which should make it easier to attract great talent.
Following the acclaim for Serkis and Kebbell's performances in Dawn, Barrett expects more actors to want to work on the planned third Apes movie, which Weta hopes to be involved in. Nothing is concrete yet, but Dawn director Matt Reeves is working on a script.
And there is still room for animators to bring their own creativity and experience. When animating a newborn ape, Barrett drew on the experience of watching his three children come into the world.
"There's that moment when they first come out. They are just pissed off: 'I was so comfortable'."
ALVY RAY SMITH, the co- founder of animation powerhouse Pixar, believes the next big thing in computer- generated imagery will be replacing actors with avatars that are indistinguishable from real humans.
However, he is not convinced performance capture is the way to achieve that.
"If you hire one of the great actors of the world, they will put up with this putting dots on their face for a few years but I think in general they won't do that. So what we need to come up with is some intelligent interface between extremely talented human actors and their avatars on the screen, and that's a tough problem. It probably requires a certain level of artificial intelligence."
Smith, who visits New Zealand next week, says Weta Digital has been at the front ranks of quality digital movie production since its inception - and he expects them to stay there.
There's a common misconception that Weta has been dependent on the Lord of the Rings and Hobbit trilogies.
In fact, Lemmon and Barrett were so busy on the Apes movies and Superman film Man of Steel they didn't even work on The Hobbit. And with Middle-earth finally wrapping up next month, the team will not be twiddling their thumbs.
As well as the next Apes movies, the Weta animators are expected to work on James Cameron's planned Avatar sequels and Jackson's slated Tintin sequel. There's also Batman vs Superman.
In the meantime, they'll be refining the detail of their digital doubles so audiences can keep believing that blue cat people do walk around in outer space - and that apes really do cry.
- The Dominion Post