MPC’s previs supervisor Duane Floch discusses three different POVs and lots of apes used to create the Colony Attack scene in Matt Reeves’ hit sequel.
With much attention focused on the spectacular animation and nuanced performance capture on Matt Reeves’ Dawn of the Planet of the Apes sequel, it’s easy to forget the massive action sequences involving a huge ape army attacking a desperate colony of humans. Not only are the visual effects spectacular, but they beautifully capture the emotions at the heart of the story - doubt, fear and hatred – which collide in a horrific and violent battle between the growing ape colony and the human remnants of a once vibrant city.
The complexity and sheer enormity of the Colony Attack scene make it easy to overlook the underlying story perspectives the director sought to convey. Central to Reeves’ visual aesthetic and storytelling sensibility was the sequence previs, handled by MPC. Led by previs supervisor Duane Floch, who handled previs supervision on the first Apes film, MPC’s team, split between Los Angeles and New Orleans, spent more than a year on the project handling both previs and postvis duties.
I had a chance to speak to Duane, who shared his insight not only on the Colony Attack sequence but on MPC’s recent move into previs, the differences in previs and visual effects production pipelines as well as the challenges faced on a film with literally hundreds of apes going to war.
Dan Sarto: So tell me about MPC’s previs efforts and your work on the film.
Duane Floch: About two and a half years ago, MPC made a decision to increase its industry offerings by providing previs services. The goal has been to take on projects and see them through, from previs to postvis and final, all within the MPC pipeline. We’ve learned some valuable lessons over the last couple years, most notably concerning building our own Windows-based previs pipeline, with the goal of seamless integration with our Linux-based visual effects Linux pipeline. The complexities of porting the tools from Linux to Windows…
DS: …it’s not that easy…
DF: …No, it’s not that easy or practical. Previs needs to be light on its feet. The vfx pipeline concept and structure, working across continents, sharing data and aspects of data, doesn’t translate well to the speed with which we need to work. With that in mind, we’re in the process of developing our own pipeline. It doesn’t always happen that we take on a show’s previs that is also going into MPC’s finals pipeline. Dawn of the Planet of the Apes is such a case. But we do have several projects in our previs pipeline that will be going through the finals pipeline with the goal that the work we do will serve as a first pass at the layout phase for the finals work.
I worked as the overall previs supervisor on Rise of the Planet of the Apes. I became pretty good friends with everyone on that crew. I worked alongside Ryan Stafford, one of the film’s [ROTPOTA] visual effects producers. We were in touch during the fall of 2012 and the new film seemed like a good fit, a good project for us to take on. It was a great project and was fun to work with the apes again. We were also able to work quite a bit on our pipeline.
DS: A number of months ago, I attended a Motion Picture Academy VFX Convergence program describing various aspects of previs within feature film production. The panel was great and had among others Ron Frankel [Proof], Daniel Gregoire [Halon] and Chris Edwards [The Third Floor]. During Q&A, someone asked if previs is such an important, growing part of the creative filmmaking process, how did the big visual effects companies ever let it get away and come under anyone else’s control [laughs]. It was a funny but interesting question. We understand how it’s a very different animal in so many ways. But with the growing emphasis on postvis and the importance of techvis and the increasing need for integration with visual effects, I’m not surprised a company like MPC is doing more and more previs to, in essence, own the entire visualization design and production effort.
DF: Exactly. What often happens with previs vendors is that the work, between the rigs, the scale and the methodology, maybe even the software, often ends up as something you can only use as a visual reference. There’s not a lot of data that can be extracted as a first phase of the visual effects work. That’s the intent of what we’re doing here. We’re not trying to grow into a huge previs house. We’re purposely being very selective in what we take on, though we have grown quite a bit over the last couple years. We’re trying to keep it manageable and keep our eye on the goal of folding what we do into a finals pipeline.
That dictates our workflow somewhat, certainly in one aspect, the reality-based way of working with cameras, that the scale we work with conforms to the scale all MPC visual effects operations worldwide work with. Vancouver, Montreal and London…we’re trying to move in that direction.
MPC’s pipeline is 20 years in the making. It’s a growing, living beast that every year continues to gain more depth. It does quite well what it is specifically designed to do. What it doesn’t do quite well is what we need to do in previs. I guarantee that’s true in every major visual effects house.
DS: I can’t imagine that’s a unique issue just for MPC. I’m sure it adds to the challenge of producing the previs in that you have to keep an eye on integration with MPC’s global production pipeline. So tell me about the scope of your work on this film.
DF: We started in the fall of 2012 with two teams. We had a team in Santa Monica that grew to 12 and a team of 16 people in New Orleans headed by my good friend Mike Neely. New Orleans began work January of 2013. We hit our stride a few weeks after that. Between the two locations, we probably did previs on 12-16 sequences. Some were small sequences, some huge. The Colony Attack sequence, where hundreds of apes attack the human colony in San Francisco, was massive. There was a boarded version we executed. Then there was the version Matt [Matt Reeves, the film’s director] wanted to shoot. So we worked with second unit director Brad Parker to put Matt’s vision of what he really wanted, which was off board, into play.
We worked on previs until roughly June of 2013. Then the postvis started trickling in. Once editorial and the production moved back to the Fox lot, like the first film, we embedded a team in editorial at Fox with up to six people. That continued for an additional four to five months.
One nice advantage this time around was because of MPC’s global infrastructure, we were able to work with MPC’s Bangalore facility, which among many other tasks, handled the lion share of match move and roto work. We could turn over plates from editorial, when we left at night, and come in the next day and look at a first tracking pass. That was usually good enough for the intent of the shot. We weren’t trying to get final tracks or roto. It was immensely beneficial to have the team at Fox not spending time drawing roto map lines and handling tracking themselves. The Fox team could focus just on addressing the immediate needs of the film’s director and editorial.
DS: For the previs, did you have to create most of your own assets or did the art department provide you with fairly finished assets you just needed to down res and put into your shots?
DF: Environments, props and the human characters were built from scratch. There were some sequences shot up in British Columbia where we got the actual geographic location and were able to lift the topographical data from Google Earth. From that we could recreate with solid accuracy the environment for some of the scenes they shot. There was a dilapidated old gas station the ape army walks by at one point. We got data from the area around gas station they were going to shoot and were able to put everything into exact positions. Then we set our camera angles from that. There was a dam sequence filmed up in Vancouver that the team in New Orleans handled. It involved cranes and camera placement. That was the closest to techvis we did on the film.
The apes themselves, some of the geometry came to us from Weta Digital via Cinedev. So we were working with Cinedev ape rigs. They started out in the very early stages of the show and had done a little bit of pitchvis for the studio. We sort of just inherited the ape rigs when we took on the show. We had cycles, a little bit of mocap for them as well as the stuff we created ourselves.
DS: So for the most part, was the goal of the previs to flesh out the story, to find the tone, to figure out what could be shot live? What were you ultimately trying to achieve with the previs?
DF: Matt Reeves had a very finely tuned visual aesthetic for what he wanted to capture in camera. So for example, on the Colony Attack sequence, shot composition, lens choices, depth of field and timing were all things we worked on to get the previs on the same page as his vision. The kinds of shots and lens choices he was asking for were very specific and helped facilitate the telling of the story.
Matt’s idea for the sequence was that it would be told from three different perspectives. One is Caesar’s son Ash, who is gung ho at the beginning of this battle. He’s looking to kick some ass. Then there is Koba’s perspective, who has run amok with rage against humans. Then there is the perspective of Dreyfus, played by Gary Oldman, who leads the humans waiting for this massive ape army to attack. Matt’s idea was to be very clear from which perspective we were seeing things. His aesthetic for POVs and establishing geography was that at no point should there be a question who’s POV you were seeing. At no point should there be a question as to where you were in the set.
It really helped sell that sequence. There was an arc to it and that arc was dependent on knowing where you were and from whose POV you were seeing the scene. It was very interesting to get into that level of detail in a sequence with literally hundreds of apes. There were probably over 500 apes involved in that attack. So you’re telling the story from within that army as well as from the perspective of the human colony. That visual sense has grown so much on me that know it’s just part of my psyche...when I’m working on a film, to be hyper aware of that POV use of the camera.
DS: In a sequence so intimately tied to the director’s aesthetic, what type of access did you have to Matt as you were creating this previs?
DF: It became pretty clear early on when Matt came on the film, with script rewrites, location scouts, work in New Orleans, then the Vancouver shoot, that once shooting started, access to Matt was going to be very challenging. Our executive producer Julian Levi and I flew up to Vancouver to meet with Matt about several sequences we were working on. We would catch him in the mornings, or at night after a day of scouting. I flew down to New Orleans many times to meet with Matt and Mike, to make sure we were on target with what Matt wanted to see. As the spring went on it became apparent the Colony Attack scene needed to be reorganized. So when Brad was brought on to shoot the sequence as Matt’s second unit director, we worked together on a daily, sometimes even hourly basis, shaping that sequence into what Matt wanted.
Brad was like an extension of Matt here in Santa Monica. When we were on location, I went and walked the set several times to get a sense of the physicality, where things were going to be, how far away they really were. It’s one thing to look at a set in 3D. It’s quite another to walk through a physical set, especially one of that size, an entire block in downtown New Orleans, and wrap your head around that.
We had had access to Matt on other sequences. But at that point, the Colony Attack sequence shoot was on and the first unit was doing their thing and Brad was doing his thing with second unit shooting. I felt fortunate to have that much access to the film’s second unit director.
DS: For the main sequences you prevised, like the Colony Attack, was the final work you provided pretty much how the actual sequences were shot and finalized? How closely was the previs followed?
DF: From what I’ve seen, it was really close.
DS: You guys had to work fast and furious, turning stuff around sometimes in a matter of hours. What were the main tools you used on this project?
DF: We worked in Maya. Our previs pipeline building effort is quite young. It really began when we started working on Apes. We had a way of outputting from Maya with forced naming conventions, a file and database structure as well as versioning that was taken care of for the artist. There was no possible user error. That was critical from an organizational standpoint. Additionally, we were in the process of handling crowds, rigging and animation imports and exports. In the Colony Attack sequence, with hundreds of apes, we were transferring a lot of animation cycles, some mocapped, some keyframed, onto our ape characters. We did geocaching on some of the characters. The more background the characters were, the less of a rig they needed. So we were able to create some really light background characters.
DS: Looking back, what would say were the main challenges on the film?
DF: It can be challenging to have previs being handled in two different geographic locations. Early on, Mike and I adopted a daily, sometimes two and three times daily phone call and cineSync session just between us, let alone between us and production. It was a challenge just making sure our workflows were moving in the same direction so that when assets like cycles, characters, environments and effects were updated by either team, it became plug and play for the other team.
It was definitely a great project. Matt has a great dramatic sense for keeping that low grade tension throughout the movie but still tell a lot of large and small stories within the larger arc of the film.
I keep coming back to the Colony Attack sequence. The number of apes in that sequence, just managing that one aspect…not just managing them, but managing them in a way you could extract story arcs out of all the insanity going on. It was a fun challenge.
Dan Sarto is Publisher and Editor-in-Chief of Animation World Network.
Dan Sarto is Publisher and Editor-in-Chief of Animation World Network.