Issue link: https://digital.copcomm.com/i/1527843
When Cassandra invades Paradox's mind, we see her ˆngers emerge from his face in a very surreal fashion. Tell us about cra"ing these unsettling visuals. This was the biggest technical challenge for us. We know our way around full photoreal humans. We can do full screen with perfor- mance, even dialogue. You would never know it's a digidouble. We've got 30 years of evolution on our pipeline to make that absolutely work. But now we had to shove a hand through the middle of one of these faces and break everything down. Everything we've done is to make it photoreal—to make the bone structures and the muscles and the skin all work in a realistic way for a normal human. Now, we're going to have to take all of that and really push it and distort it in ways that are not like a human at all. She's pushing her hand through the skull, it intersects through the skin, and it comes out through a nostril, but ignores the teeth. The logic in it is com- pletely bonkers. But we needed people to believe that she was doing this thing, and we wanted everyone to feel like that was the last place in the world they wanted to be: in Paradox's shoes with Cassandra rummaging around the inside of his head, just in an awful way. We literally had one comic book frame of what they wanted it to look like, and we knew it was a full dialogue sequence. Paradox was actually talking; Cassandra was chatting away. We knew it would be close up and the shots were going to be long. We knew our digidou- ble was good, but all the dynamics, facial controls—everything had to be completely rebuilt from scratch. We had to invert our whole process. Usually we put all the simulations on the backend because you want everything to be working: the animation, the dialogue...Then you run the simulations over the top of it to get the skin, the little wrinkles, all that nice stuŸ. But we knew that if we ran the simula- tions at the backend, the client would never see what they were go- ing to get until right at the very end. I wanted to get into a position where I could present, as early as possible, what was happening to his face, her œngers, her hands. Our hands are fully CG, fully animated, but what was happening with the face is very much a mixture of Matthew's performance—this brutal intrusion into his physiology—and then how he is reacting to that, because he didn't know what the animation was going to be like. We were using cues from his other performances to know what he might do when this happened. We had to get the big performance beats in the animation so we could fast iterate that and get feedback. All those massive de- formations were now controlled by the animators, because a lot of that could never be simulated. You could only do it with an animator having some fun. We could get them to sign oŸ on all the really big stuŸ, and then we could give it to the simulators and eŸects artists to make it all link together. A˜er we made it feel unreal, we needed it to feel real again. We knew very early on that it was going to work, but we also knew it meant 15 years of pipeline work had to go in the bin and be rebuilt in six months. Which programs and tools were most vital to your team's VFX work'ow? We used pretty much the full gamut. For the main visual eŸects, we very much relied on the big three, which are Maya, Houdini, and Nuke. For things like visualization, we're deœnitely using a lot more Blender now. That's been fully incorporated into our pipeline. We're used to having multiple vendor shows, but now shared ven- dor shots are much more commonplace. With Marvel, it's a normal thing now to be sharing not just shots, but assets as well. So over the last ten years, we've slowly worked our way back from highly pro- prietary tools, because it makes it very di•cult to work with other vendors. Now we're using much more open-source tools and exchange formats because we know we're going to be working with other ven- dors a lot. We can share assets much faster and rely on the skills of the artists to get us through the unique scenarios. We love challenging shows. We very much put ourselves forward and say, "If you've got something that's really tricky, we're up for it." We've built a team who have the skillset to do new and interesting things. We want to be there challenging those very fringes and push- ing the envelopes because that's what we built a team around, and that's what that team wants to do. We love doing that and we love doing it well. Do you have any advice for up-and-coming VFX artists? I was a head of department for 2D for many years at Framestore, so I was looking at showreels—for compositors, in particular—probably by the thousands. Showreels are very important. We're not looking for people being clever and pushing the boundaries on their show- reels. We're looking for quality and consistency. It doesn't matter if you've done the most fantastic shot in the world if it's not done pro- fessionally. If you've done a hundred boring blue screen shots, but they look really well put together—we like that. So it's not about showing oŸ. It's about showing your profession- alism. The ³ashy stuŸ will come with time, because it always does. In the meantime, we've got 100 shots, and 90 of them are bread-and- butter, need-to-look-perfect shots. We need to impress our clients with how clean, tidy, and professional we are. ted like that. ¢