In 2019, Nhat Phong Tran told us about the work of Pixomondo on the second season of THE ORVILLE. He then worked on EUPHORIA.
A few months ago, Phi Van Le told us about his work on MIDWAY. He then worked on FOR ALL MANKIND and WATCHMEN.
How did you and Pixomondo get involved on this show?
Tran: We’ve been collaborating on this show since Season 2, and it’s been a pretty good ride, so why don’t we do it again for Season 3?
How was the collaboration with the showrunner, the directors, and VFX Supervisor Jay Worth?
Tran: We mainly communicated with Jay Worth, who is the creative voice of the show’s visual effects.
Le: Jay was able to get feedback directly from Jonathan Nolan, which streamlined a lot of the ideas back to us. The overall collaboration was fantastic.
What was their expectations and approach about the visual effects?
Tran: “Invisible effects.”
Le: Pixomondo created the futuristic Los Angeles we see throughout the season. The expectations were to make it feel real and believable. I suppose that is expected on most projects when it comes to VFX, but Jonah and Jay wanted people to believe that this is actually Los Angeles in 30 to 50 years. Therefore the environment became a combination of what historically currently exists and what could exist a few decades from now. Production supplied some early concepts for future Los Angeles, and we ran with it.
How did you organize the work between you?
Tran: Phi and myself have been pretty good at understanding each other’s responsibilities and perspectives, which gives us a synergetic advantage over other VFX supervisor/producer duos.
Le: Nhat just can’t get enough of me!
Where was filmed the various parts of the series?
Tran: Los Angeles, Singapore, Spain.
How did you work with the art department to design the futuristic environments?
Tran: We received artwork from the art department, which gave us a pretty good idea of Jonathan Nolan’s vision of the future world. There was no expectation of being provided with artwork for every single building of a multi-million person city like Los Angeles, so we studied the designs and then interpolated and extrapolated the style to form a cohesive look.
Can you explain in detail about the creation of these beautiful environments?
Tran: Building large scale cityscapes often run into the limitations of repetitiveness. Creating full building assets in CG gets proportionally more expensive with the number of buildings we see. In some shots, we easily added 50+ buildings. There is no way we can economically hand-model each building. There is also no way we can reuse a pre-set library of buildings 1:1 without seeing significant repetition. So the beauty lies in controlled randomization of pre-modeled buildings. We were very strict internally about having a concept artist lay out the silhouette of the skyline, which to me, is equivalent to a musician laying out the rhythm of a piece. It is imperative at this stage that the rhythm is aesthetically pleasing and exciting. We wouldn’t want 4/4 time, but something more like a Waltz or ¾ time. Once that is defined, we get our skillful artists to randomize and alter our pre-modeled buildings. The result is a cohesive design language maintained across the future buildings’ general look while providing enough alteration to prevent repetition.
How did you create the various shaders and textures?
Tran: We used pretty straightforward layered shaders that were procedurally modulated to create variation.
What kind of references and indications did you receive for the environments?
Tran: We received plenty of references drawn by art departments and from architects that inspired a lot of the designs. One reference came from architect Bjarke Ingels, while Singapore’s Marina One was from Ingenhoven Architects.
Le: We wanted the look of the environments to not only be futuristic but to look believable. If WESTWORLD Season 3 is only about 30 to 50 years in the future, we wouldn’t see BLADE RUNNER or THE FIFTH ELEMENT-type buildings and vehicles just yet.
Can you tell us more about the lighting challenges?
Tran: When you work at large scales like the Los Angeles metropolitan area, you start running into macro-type illumination phenomena. For example, the Raleigh scattering component from the Mie and Raleigh scattering model becomes apparent. In addition to that, the smog creates a large gradient that affects diffusion through the volume. Almost like the whole frustum of the camera is slicing through the gigantic sub-surface scattering.
How did you populate the cities with crowds and vehicles?
Tran: Some artists had much fun placing and animating those. It was a fun project to take on because each one of those – let’s call them “agents” to speak of in massive terms – had its own story, including why someone/something was going somewhere or moving towards.
Did you use procedural tools to create some parts of the environments?
Tran: We used CityEngine to proceduralize deep B.G. buildings.
Which one was the most complex to create, and why?
Tran: Los Angeles. Building all our city augmentations true to scale and geographical layout was very complex.
Can you elaborate on the creation of the vehicles?
Tran: The show’s art department provided the design of the vehicles themselves, however, we still wanted to tell the backstory and mechanics of each vehicle. Phi and I are Los Angeles-based, so knowing the city, the climate, etc., helped a lot in this case. When we textured and shaded the vehicles, we spent much time laying out where we’d see accumulated weathering and streaks. Since it does not rain a lot in L.A., it’s likely that the streaks we see on ground vehicles are caused by moisture from the earth that get thrown upon the cars. For the aerial vehicles, we took into consideration the bi-directional travel opportunities as well as the condensation streaks of the beautiful morning dew we often see over the city, up to a certain altitude. The streaks and textures also have a more prominent layer of dried-up salt from the ocean.
How did you handle the challenge of their reflective aspect?
Tran: A reflection map that is a rough representation of the scene with enough dynamic range is usually sufficient, as long as light sources are properly projected to create the necessary amount of parallax we see in the reflections.
Which sequence or shot was the most challenging?
Tran: Possibly the creation of future L.A. There were many design studies on how to create functional buildings that follow the safety codes and incorporate ideas on how to maximize the floor area to get the most out of it for the optimum quality of life.
Is there something specific that gives you some really short nights?
Tran: Getting the 8K LED wall footage that needed to have a C.G. futuristic Los Angeles skyline out in time for the live-action shoot with a cast and crew was something that could keep you up and rob you of an hour or two of sleep. Not because you’d be concerned about time and render constraints, but because you actually can’t easily review a 50’ curved LED wall footage at your office, so the concern is more that you don’t have a clue of how it’s going to look like on stage!
Le: Also, there was a particular sequence in episode 307 that proved to be quite challenging towards the end. It was a sequence that involved « infinite » cryopod chambers jammed inside a room. The creature of this sequence coincided with the mandatory COVID lockdown, where Pixomondo had to completely transition from everyone working in the office to everyone working remotely due to government mandates. After about a week, we were able to adjust and resume back to « normal » quickly. That was a bit of a stressful week but we couldn’t have been more proud of the final product!
What is your favorite shot or sequence?
Tran: Certainly, the reveal of Los Angeles. It elicited a strong, powerful, emotional response of uncertainty. We see the world outside of WESTWORLD – it looks beautiful, grand, and kind, and it is home to better people. But it is also the type of world that is capable of creating a powerful, desire-fulfilling, sometimes violent theme park like WESTWORLD.
Le: My favorite sequence is in the first episode of Season 3, where Dolores is flying through Los Angeles on an airpod. Pixomondo helped supply VFX for the background footage used in this scene.
What is your best memory of this show?
Tran: It was a late night, and we were working on creating the first version of one of the bigger establishing shots of the future L.A. cityscape. The moment we saw it playing back in our office theatre with all the bells and whistles, I felt a collective understanding that this could very well be our real future, if not the future of our children. It was an exciting moment to get that possible sneak peek into a potential world and have the whole room of artists and production feel lucky it could be ours.
Le: I am a fan of the show, so seeing the work come to life with a new episode every Sunday was probably the “best” memory for me this season.
How long have you worked on this show?
Le: We started looking at concepts and bids sometime in June 2019.
What’s the VFX shots count?
Le: A little over 400 shots.
What was the size of your team?
Le: Over 160 people have had their hands on the project throughout the season.
What is your next project?
Tran: THE ORVILLE Season 3.
Le: There are a few shows I am bidding on that I can’t mention at the moment, but hopefully soon you’ll be hearing about it.
A big thanks for your time.
WANT TO KNOW MORE?
Pixomondo: Official website of Pixomondo.
© Vincent Frei – The Art of VFX – 2020
Nhat and Phi were so great to work with. Along with the rest of an awesome team! This was a tough show but so rewarding! Also pretty honored to see 2 of my shots featured in this piece 🙂