The Oscars: Burke Talks Deathly Hallows: Part 1
TB: I think when we got a lot of the Houdini fluids working and that gave us something we could all latch onto. And we were using facial capture data to drive the Voldemort faces and that plugged in quite well. It was all about trying to integrate them into this big, swirling mass. But there was no real eureka moment -- it just pushed us to the line.
There were other very big, complex sequences: the seven Potters motorbike chase, the sheer volume and size alone was a difficult one to deal with because it involved so many CG environments. Again, I'm not sure how many people realize how much was CG and how much was real. From the moment they take off Privet Drive, a partial set within a set extension and then into full-CG cloudscape environment and then they crash back down into the tunnel, which, for the most part is CG, but some of the traffic is real and some is CG, and full CG environment outside. It was quite a mammoth job for everyone.
BD: Even though you're dealing with a fantasy world on the outskirts of our real world, it has to look believable and consistent.
TB: Yeah, absolutely: you've hit the nail on the head. And that's what we've strived for. It can't be a fantasy world because the audience has to believe the wizarding world exists. So if you go too fantasy with the design of the buildings, the environments and the effects, then it would remove the believability of this parallel world. So we've tried to make the effects real and grounded and have a sense of volume and presence. You can immerse yourself in the story, and with Part 1, the wizarding world was invading the muggle world. So that was another great shift and to see the cross-over.
TB: For production purposes, we've treated it as one big film. So, we started our previs on sequences in Part 1and Part 11 simultaneously. In 2009, we were actually finishing off the sixth film but got the previs up and running so we could develop some of the big sequences for Part 1 like the motorbike chase. But we were also prevising a lot of the major battle sequences in Part II because the shooting schedule dictated that some of those fell a lot earlier than sequences that had to be filmed for Part 1. For instance, we didn't shoot the Horcrux scene until May 2010. Of course, we've been developing R&D for Part II since back in 2009, when we started the digital build of Hogwarts. Because, in order to achieve everything we needed for the massive fight sequences and the level of destruction, and to have flexibility later on, we decided, for the first time in the series, to do away with the miniature and turn to a full-CG environment of Hogwarts. So that's been ongoing for more than two years now and has been an absolute godsend. There have been so many changes in design and in shots and sequences and how the shots are being used. But because we have a CG model, we've just been able to go in and re-previs, re-block out. And, with the level of detail of our model, we can pretty much fly into any part of the school. So we've got complete freedom.
Bill Desowitz is senior editor of AWN & VFXWorld.