When Bird Box was released over the Christmas holidays, the film became an instant sensation. According to Netflix, over 45 million accounts tuned in to watch Sandra Bullock in a post-apocalyptic world trying to save herself and her children from death.
Deprived of sight – whatever it is, if you see it, you will die – a mother has to embark on a terrifying journey down the river to find a safe haven. VFX Supervisor Marcus Taormina nominated for a VES Award talks about how there were 520 visual effects shots in the final cut of the film and discusses the challenges of that dangerous river journey. He also talks about the creature and the decision not to let us see it. (Thank you!)
Bird Box became a social media/viral sensation when the film dropped on Netflix, What was that like for you?
It was quite a welcome surprise. I knew the film scored really well in the test screenings we had, but I didn’t think it would become such a phenomenon that it has.
At a press conference, Sandra talked about the decision to not have the audience see the creature because it made her laugh. Talk about the design and the original unseen creature.
It was interesting to see that article make the rounds. The specifics of that scene were solely based on what was written as Malorie’s nightmare; her vision of her unborn, half human, half otherworldly baby that Charlie eludes to in the beginning of the movie. From a VFX standpoint, we knew that her serpentine-like child would become a full CG creation, but we had hoped that we would be able to harvest some or all of the facial and body performance by actor Dirk Rogers. We were moving out of the conceptual and modeling phase of that scene when they pulled the plug on that scene.
Some of the procedural technicalities of visual effects often lead to oddities on the day, so I understand from Sandra’s point of view how ridiculous this all must have felt, but I have to say the design we were working towards that she was never privy to was rather disturbing and unique. In the end, I do agree with the filmmakers’ decision to pull that scene as it would have been the only true visual form of a creature, and it just didn’t fit within the final version of the film.
Did you create a backstory for the creature at all?
We did. We had many conversations and created many documents of what the creature could be, the backstory, the origins, why it was here and why it only influenced certain individuals. In the end, the amount of exposition to bring these ideas forward was just too much to fold into the plotline and story arch.
One of the most terrifying aspects of Bird Box is the river journey. Talk about working with water and the effects process?
In the early stages of prep, there was a suggestion to shoot the majority of this action on a section of the Smith River in Northern California during the summer months in which the river was extremely calm. While there were contingency plans in place for special effects to provide practical water agitation and sub-surface bubbles for visual reference and interaction against the boat for VFX, the technical difficulties, environmental restrictions, and cost ramifications for a partial or full CG water surface approach was a deterrent for production and the studio.
Ultimately, production swung the first half of the sequence shoot to later in the winter when the water would swell on this free flow river creating the rapids the audience was expecting. This approach, of course, presented additional concerns, most of which focused on the safety of both the shooting crew and the stunt team headed up by Malosi Leonard. Once safety was addressed, the technicalities of the shoot were discussed.
For the majority of the in-camera footage on location for the first half of the sequence, VFX augmented the stunt performers body shapes utilizing 2D compositing techniques that became known as “nip-tucking” to fall in line with the shape and look of Sandra Bullock and two six-year-old children. 2D blindfold patches added to the work, along with the reshaping and removal the safety helmets of the stunt performers. On occasion, the medium and close-up coverage along with the high-resolution deliverable presented the need for digital head replacements. Sandra was digitally cyber scanned and photographed on location in Los Angeles. ILM London led by supervisor Mark Bakowski rigged and rendered a full CG head complete with digital hair & cloth simulations for a handful of shots in the final cut. Finally, a blanket of 2D and 3D atmospherics were composited over these shots to give the sequence more of a menacing tone it deserved.
The apex of the sequence became the most challenging part of the sequence. The tank shoot was slated for the last two days of principal photography. A large water tank was painted blue, surrounded by over 200’ of blue screen and, with the help of the Special Effects crew led by Mike Meinardus filled with 300,000 gallons of heated, tinted water to match the practical footage in the first half of the sequence. The special effects team also provided a pair of air compressions at 750 cubic feet per minute provide omnidirectional water flow, subsurface agitation, and aeration along with the ebbs and flows from secondary waves for interaction against the practical rocks, boat, and oars that are seen in the final composites. Without this support, the footage would have lacked the frenetic nature it needed and deserved, and the performances from both Sandra and the two children may have suffered.
The background support for these blue screen composites was shot on location in Crescent City. Both static and spydercam array plates were utilized for the 2.5D and 3D approaches. The array run consisted of three digital camera bodies positioned in portrait mode to achieve the most coverage on the vertical axis due to the sheer size of the Redwoods on the Smith River. Multiple runs captured an extensive 360-degree cyclorama used by multiple VFX disciplines in post-production. Ground and aerial-based lidar scans were used for reprojection, layout, and match move purposes. Digital stills along with static and boat propelled motion picture footage was captured along the journey in multiple areas of high action for supplemental support purposes in post-production.
What were the challenges in creating the water effects?
In post-production, we hoped we had all the materials and data needed to accomplish the task at hand. Going in we knew we had some rather complicated shots, but we had hoped that we would be able to treat the majority of the tank footage as composites. As the cut developed, we knew we would have to approach some of these shots as 3D builds with 3D water simulations, so we got that process going as soon as we could in the dev phase.
The majority of the 3D water and environment builds that are in the final cut were approached in this manner due to their continuity in the final cut of this sequence. There was so much great stunt footage to work shot on location to support the coverage of Sandra in the tank, but ultimately some of those lacked the raging rapids we needed, and we knew we would need to add the 3D environmental elements to meld the action together in continuity.
With both the 3D builds and 3D simulations, we tried to split into any practical elements we could that we shot on location. For example, some of the 3D simulations of the water work included a handful of practical elements of raging water that was used to mask the line between CG and practical in certain shots. The environmental builds were based on the ground and aerial based lidar that was then textured and lit utilizing all of the DMP capture on location from both units. When we were able, we stole hero bits and pieces and kissed them into the shots. We tried to be very strategic on utilizing these elements and studying the footage to find other pieces that stood out to us from the location work that would also work to our advantage when viewing the sequence as a whole. When the environment and water simulations were complete, we added atmospherics and water droplets among other elements to marry the final composites together.