In 2013, Chad Wiebe explained to us about the work of Scanline VFX on MAN OF STEEL. He then worked at Method Studios on THOR: THE DARK WORLD, GRUDGE MATCH, NIGHT AT THE MUSEUM: SECRET OF THE TOMB and AVENGERS: AGE OF ULTRON.
How did you got involved on this show?
Having recently worked on AVENGERS: AGE OF ULTRON, I was eager to continue working on Marvel projects and while this was my first time working with Stef Ceretti, he has a history with Method from previous projects and also as former Method Supervisor.
How was this collaboration with director Scott Derrickson and VFX Supervisor Stephane Ceretti?
It was great. I’m a bit of a horror buff and was a fan of Scotts previous films so I was curious what influence his background would have on DOCTOR STRANGE. Scott brought a very clear creative and aesthetic vision to the film which definitely made it unique and both Scott and Stef, as well as the Marvel team, were always very open and receptive to new and crazy ideas, which allowed us to explore things without worrying about trying ideas that might be considered too far out there…. in fact at times it was highly encouraged. This was my first time working with Stef, but over the course of the shoot we were able to develop a good communication flow which carried forward through the rest of the production schedule which made it a really enjoyable experience.
What was their approaches and expectations about the visual effects?
Well, I think with DOCTOR STRANGE being a new character introduced into the Marvel universe, and with so many other characters already existing (each with their own unique powers), they really wanted to make sure what we created for Strange was unique and fitting to his character and to the history of the character as depicted in the comic books. At the same time there was a specific directive to make sure the effects felt like something relatable and physical, as opposed to something overly magical that the audience wouldn’t be able to connect with. Everything that we created needed to have a physical or textural quality to it that was based in the real world, even for those effects that seemed completely out of this world.
What are the sequences made by Method Studios?
Method worked on the Car Crash sequence, the Kathmandu Arrival and Rooftop Training sequences in the Kamar Taj, the Tea Room mirror dimension and portal sequence, the Apple Moment sequence (with the time shifting apple and time shards), the Narthex sequences, and the Magical Mystery Tour.
How did you organize the work amongst the different offices of Method Studios?
The Majority of the work was completed at Method Vancouver, however since the Magical Mystery Tour was somewhat of a sequence unto itself, it was split between Method LA and Method Vancouver, where Vancouver did all shots involving Strange’s journey through earth and space, and Method LA doing all shots within the Multiverse environments under the supervision of Doug Bloom, with Olivier Dumont serving as visual effects designer.
Doctor Strange have an impressive car crash. How did you approach this sequence?
The beginning and end shots of the car crash were shot in-camera on location at Greenwich and on the Longcross Studios backlot. Everything in between was digital including the environment, the car and the majority of Strange for the interior car shots during the crash. For each location, knowing we would need to re-create and extend each environment to remain consistent and look like the New York mountain roads near Bear Bridge, we had teams collecting reference photos, tiles, and pano’s to ensure we had more than enough reference material to create a unified set of digital environments for each section during the cars descent to the river side.
We also had various stages that the Lamborghini had to go through. We started and ended with the sequence with the practical cars shot using special effects rigs, so we had to match some of the damage on the practical cars while at the same time, trying to ensure we didn’t limit the amount of damage we saw along the way. So we had a bit of a hybrid approach where much of the car destruction was achieved using RBD’s, but also had some model based shapes that we needed to blend to for the hook ups.
Did you received specific references for this sequence?
We were actually able to gather very thorough reference material while shooting at the various locations. Since they damaged numerous cars during filming, we were able to photograph and scan the cars in various stages of damage and took detailed photographic reference of the undamaged car to ensure we could recreate an exact digital replica. We also had teams at each location gathering reference for the environments in varying lighting conditions, so we really had a thorough library to utilize while building out our digital assets.
Can you explain in details about the creation of the inside shots?
The interior car crash shots were critical in conveying the story point of Strange’s hands taking the most impact in the crash, leading him on his quest to find someone who could heal him. Story wise it was a pivotal moment in the character development, so there was much emphasis put on the importance of these shots showing the moment of impact. Knowing we would need to replace or destroy the majority of the Lamborghini interior, as well as Strange, we went through an in-depth process of modeling and sculpting every detail of the car interior. Although we received LIDAR scans of the interior, due to the limitations of getting scanning equipment inside the car as well as the many reflective surfaces not being ideal for scanning, we relied more on photo reference to ensure an exact match, using the LIDAR more for general placement. For Strange, we had initially intended to use as much of Benedict shot in-camera as possible, however, once Scott started to art direct the shots, we needed to replace most of his body, arms and hands, keeping only his shoulders and head. We had already built an asset for Strange in his Kathmandu costume, but the car crash sequence required a new costume and a more hero version of the hands for an extreme close up where they crash through the dash of the car. For the destruction, we ran a number of RBD simulations for the car interior, emphasizing the front of the car collapsing in and pushing the dash towards Strange as his hands are pushed through the glass display, all at high speed so we really had great attention focused on directing the speed and amount of debris and destruction happening within the car.
Can you tell us more about the creation of the huge environment?
The environment was created by referencing numerous plates and photos taken at various locations during the shoot in London and New York. The backgrounds were predominantly DMP. For the mid-ground and foreground, we modeled out the roads and hillside, including the switchback area which was fully digital. We used photogrammetry for a general mockup of some of the locations, and detailed things out from there, sculpting the rock side in ZBrush with a high level of detail since we would be relying mostly on highlights in a dark environment to give the most perceivable detail. The trees were a combination of instanced static and animated tree’s and dynamic tree’s simulated in Houdini for interacting with the car.
The sequence is seen by night. How did you handle the lighting challenges?
Well, that was one of the bigger challenges – lighting a highly reflective car in a predominantly moonlit forest environment. CG cars can be surprisingly tricky to make photo real in the best of lighting conditions, so having a dark car travel through the forest in a moonlit environment proved challenging. We found that we had to make a few shader changes in order to inject a bit more diffuse contribution in the car shader to avoid areas becoming too dark due to really only reflecting dark surfaces. So we tended to over light our scenes slightly which allowed our comp team more control to balance the levels in a more creatively driven manner.
Strange is then moving to Kathmandu. Can you tell us more about your work on it?
Strange’s Kathmandu arrival was the first sequence to be shot in Nepal during a 2-week shoot which involved shooting exteriors and aerials at various locations around Kathmandu, shootings extensive reference photos throughout the city, as well as traveling to Pikes Peak and Pashupati to gather tiles and panos of the Himalaya mountain range.
How did you recreated and extend this city?
We were fortunate to be able to gain access to about 10 different residential rooftop locations over a two-day period near Patan Durbar Square, where we had 2 teams simultaneously taking photos in the morning and afternoon to ensure we had every angle you might possibly see from the Kamara Taj, shot in multiple lighting conditions. This was a tremendous experience not only for the amazing reference we were able to gather but also the experience of having local residents invite us into their homes and allow us access to their rooftops throughout the two-day shoot. Additionally, Todd Perry (CG Supervisor) and I spent another 4-5 days exploring different areas of the city and literally wandering through different neighborhoods gathering texture and architectural reference. This reference allowed us to create a full 360 degree 2.5d version of the city and Himalayan mountain panorama.
Can you explain in details about the creation of Kamara Taj?
There were 2 interior sets built at Longcross Studios in London, one for the rooftop training area (doubling as the courtyard), and one for the interior Sanctuary (Tea Room). The sets were built to an incredible degree of authenticity, so our reference and texture photos from Kathmandu were extremely useful when creating the digital extension for the Kamar Taj. Originally the intent was to only build out the areas of the Kamar Taj as seen from the perspective of the physical sets, however, part way through production Scott wanted to add a few fully digital establishing shots of the Kamar Taj which weren’t shot during the Nepal shoot, nor did the sets build in London exist anymore. So we ended up building the entire Kamar Taj to an extremely high level of detail, to allow for complete flexibility when setting up new shots and not being bound by the assets limitations. Ultimately this proved to come in very handy due to re-shoots which also required a fully digital Kamar Taj for a new sparring session sequence between Strange and Mordo.
How did you design and created the Mirror Dimension?
Since the mirror dimension was a look that was to be shared between vendors, production supplied us with the general concept for the mirror dimension and we created a set up that would allow us to either generate 3d geometry which comp could use to dial up or down the refractive levels, or in some simpler shots, comp would create their own shapes that they could control completely. The look itself was based in subtleties, and really the idea was to create the illusion of space being sliced up into sections or slices, which only affected those not within the mirror dimension. So you would get these reflective, refractive slices going through an environment, pushing depth planes to unnatural levels causing the viewer to not quite understand what they’re seeing. Overall it was a sublet but very cool optical illusion.
Can you tell us more about the Narthex Globes?
The Narthex globe had a couple of purposes story wise. It was used by Mordo, Wong and the Ancient One to monitor the 3 sanctums (London, Hong Kong, and New York), and it was also used to help tell the backstory of how the sanctums formed a protective shield from Dormamu. The idea behind the effects on the globe was to show the connection between the sanctums, depicted using ley lines and light points throughout the globe. When exploring ideas for this effect, one key piece of reference stood out, which was the Earths City Lights image from NASA, which is an image created using satellite captured tiles of the earth stitched together to show the entirety of the planet lit at night. This instantly resonated and we incorporated it into some early tests that eventually found their way back into the final result after a few different conceptual changes.
Lots of magic happens in Kamar Taj. Can you explain in details about its design and its creation?
After Strange arrives at Kamar Taj, the story focuses on his training and challenges as he learns about the mystic arts, and how to use them. The Ancient Ones takes him through the creation of mandala’s and shields, and Mordo and Wong teach him how to create portals, weapons and train him to fight. Many of these weapons and shields are created using sparks and fire based elements, so they needed to have proper physics and gravity to avoid looking too magical or like pixie dust. The key directive from Scott and Stef was to make sure they always felt like something that was physical, that the audience could relate to and understand. That was a common theme in nearly all of the magical effects developed for Strange.
How did you handle the animation of the magic?
One example that was developed which was the basis for most of the shields and weapons were the mandalas, which the ancient one teaches Strange to create while training in the Kamar Taj. The FX set up was created by our FX lead Ryan Costar and the look driven heavily in comp, developed by our comp supervisor Sandra Balej. We were provided designs as starting points from the production side art department which we then revised to allow more functionality as a 3-dimensional object. We converted the designs into spline based shapes, which we would then run particles systems through to essentially « draw on » the shapes matching the Ancient Ones hand gestures (or whoever was creating them), as well as emitting particles and sparks along the way. The splines also had additional repeating fractal patterns run through them that allowed much more complexity and detail within each shape, allowing for unique patterns to be applied to each individual shape. Once all the elements were generated, a complex set of AOV’s were generated which allowed our comp team full control of how light and intensity contributed to the look of the mandalas, which was a huge part in the overall look. Since the effect is very light based, Scott was very specific that he wanted to make sure it always looked photographic.
Doctor Strange is put into space and then the Multiverse. Can you explain in details about this crazy sequence?
In an attempt to prove and explain the multiverse to a skeptical Strange, the Ancient One sends him on The Magical Mystery Tour, which is a psychedelic trip through space and into the realms of the multiverse. It starts with her telling Strange « You think you know? You don’t know » and bam, she sends him on an incredible journey. It starts with the tea room expanding back into an impossible perspective, and Strange sling shouting through the room, breaking through the wall, and is hurled impossibly into outer space. As he deals with the losing his grasp on reality, things get even weirder as he finally comes to a brief pause floating above the Earth, only to find a monarch butterfly flying casually by him….in space. That’s when the real trip begins. He is once again thrust back at incredible speeds, skimming the stratosphere of earth, as a giant wormhole forms and he is sucked into it and directly into the multiverse, which consists of 5 unique and complex dimensions.
How did you manage the design and animation for this sequence?
The MMT sequence was a very ambitious endeavor. It started out as a 7-minute sequence with 12 dimensions in addition to the space sections where he was floating around outer space. That was the working length for a significant amount of time, before eventually being cut down to a 2-minute sequence with 12 dimensions being trimmed to 5. Since it started as such a large sequence, we relied on the team at Method LA, supervised by Doug Bloom to take on the majority of the sequence. Specifically, the multiverse design and execution with Olivier Dumont severing as the Visual Effects Designer, who worked closely with Scott and Stef to ensure this massive sequence got full attention as a stand-alone sequence. Vancouver’s work on MMT focused on the space shots and all that was not part of the multiverse dimensions, which also had some significant trimming over the course of the editorial process.
How did you handle so many elements during this sequence?
Since the sequence was split between offices, Vancouver dealt with the earth and space environments and LA handled all the multiverse environments. However, when it came to our digital Strange asset which was shared between locations, there was some juggling required. It had many components specifically designed or re-designed to handle the level of detail needed for extreme close ups which in many shots were completely digital. It required a collaborative effort between teams in Vancouver and LA to develop a completely new hair pipeline, a new skin shade system, a newly designed facial rig system, and building on a pre-existing cloth system which had been developed by our LA team for ANT MAN and pushed further to facilitate the needs of STRANGE. So with developers collaborating in both locations, we were able to utilize the talents in both locations to focus on the many aspects needed to push Strange to a new level.
What was the main challenge on the show and how did you achieve it?
I think that like any show where you have multiple vendors working on and having to share key effects across multiple sequences, there is always a bit of back and forth and ever changing goal posts in terms of matching looks or sharing effects set ups. With things like the mirror dimension, portals, eldritch shields and whips, and mandlebroting being used my multiple vendors, keeping consistency and continuity can always be challenging, especially when the design or look of the effects evolve over the course of the production.
Was there a shot or a sequence that prevented you from sleep?
There was a shot where in the tea room, the Ancient One is trying to explain to Strange the mirror dimension and to prove a point she shifts the entire room into a fully mandelbroting environment with all components shifting, morphing and repeating while in constant motion. This shot was a one off for Method, but was heavily featured in other non-Method sequences. Since the other sequences were being designed and developed by other vendors, trying to create a set up for a one off shot (versus developing a complex set of tools for multiple shots or sequences) proved extremely challenging. Typically, something like that you would spend a great deal of time writing tools and developing workflows to allow for mass production, but because we only had the one shot, and the overall look was being determined by other vendors, we treated it as a one off and really had to try to work in shot to match a look that was constantly evolving in the other non-Method sequences. In the end the shot turned out great, but there was definitely a moment of middle of the night panic wondering how we were going to pull it off. Fortunately, our FX lead Ryan Coster dedicated the last couple months to cracking the code of this shot and developed a workflow that allowed flexibility and a quick turnaround when we received feedback.
What do you keep from this experience?
That nothing is impossible in the Marvel Universe. Just when you think things can’t get more insane, they just keep pushing the envelope (in a good way).
How long have you worked on this film?
In total with the shoot, around 12 months.
How many shots have you done?
Just under 300 between both facilities.
What was the size of your team?
Approximately 140 artists and production staff.
What is your next project?
Something I can’t talk about currently, but will be revealed soon enough.
A big thanks for your time.
// WANT TO KNOW MORE?
– Method Studios: Dedicated page about DOCTOR STRANGE on Method Studios website.
© Vincent Frei – The Art of VFX – 2016