Stuntvis Amps Up the Action

By IAN FAILES

Previs and its often-related disciplines of techvis and postvis are steps in the filmmaking process that greatly help in planning complex visual effects-heavy sequences. But there’s another step used regularly in such films that does not always get talked about: stuntvis.

Stuntvis – or stunt visualization – is usually some kind of video reference undertaken as part of the design of a project’s stunts by a stunt coordinator or action choreographer prior to anything being filmed on set. It’s a blueprint for a major action set piece or a fight sequence, and one that often segues into previs itself or becomes part of the ongoing edit.

To learn more about where stuntvis sits in the filmmaking pipeline, VFX Voice spoke to some stuntvis artists and visual effects supervisors about their recent experiences on major films and television shows with this nuanced discipline.

Chris Clements

“It’s not uncommon for the final sequence to be nearly identical to the stuntvis, from the camera angles down to little dust impacts that need to be digitally added.”

—Chris Clements, Stuntvis Artist

Yung Lee

“I prefer to call stuntvis ‘action design,’ as it involves a lot more than stunts and requires the involvement of many key departments. The entire process includes the development of scene blocking, choreography, stunt testing and rigging, set design, camera placement and movement, lighting setup, VFX breakdown, editing, and prop and weapon conceptualization.”

—Yung Lee, Stuntvis Artist

WHAT IS STUNTVIS?

Imagine a script calls for a major superhero fistfight between a hero and villain to take place on a freeway overpass. Quite often the stunts team will not have access to the final shooting location, so instead they might imagine the sequence in a rehearsal space and then edit the footage together. That edit, especially for a superhero showdown, will typically include some key placeholder visual effects elements, with the final product a visualization of how the stunt coordinator or action choreographer sees the fight as playing out. The sets and costumes are not important, just the main beats.

“Stuntvis has multiple purposes,” outlines Chris Clements, a stuntvis artist with credits on Neflix’s Daredevil and The Punisher television series and Pacific Rim Uprising. “It’s used to sell the director and producers on the vision, to work out certain beats before a shoot to make sure they play correctly to camera and, most importantly, to act as a souped-up storyboard that can be referenced by multiple departments.”

“It also gives the post visual effects team a great reference for what the stunt department is thinking the final product should look like,” adds Clements. “It’s not uncommon for the final sequence to be nearly identical to the stuntvis, from the camera angles down to little dust impacts that need to be digitally added.”

Clements recently worked with action consultant Philip J. Silvera on two large action sequences in Pacific Rim Uprising, helping to imagine how the frenetic fighting styles of the robot Jaegers would look. “Those sequences required a little bit of everything,” says Clements. “We added in miniature crowds, mech blades composited onto the stunt doubles, missiles, helicopters and a ton of destruction!”

Silvera is a strong proponent of stuntvis. He also used it to present his fight ideas on Deadpool and its sequel Deadpool 2, and for the Daredevil series. On this Netflix show, Clements was provided with edits of the planned action sequences to add effects elements. “Typically, this would involve gunshots, blood hits and dust hits to help sell the action,” he explains. “The stunt coordinators I work with are very story-driven, so it becomes important to highlight certain weapons or injuries or portions of a location in the stuntvis. The fun really starts for me when I need to figure out how to achieve a difficult effect in a believable, time-efficient manner. It becomes a tug of war between time and quality.”

“On season 2 of Daredevil, we worked on a huge stairwell fight that was sort of a sequel to the famous hallway fight in season 1,” continues Clements. “Daredevil fights with a chain as he makes his way down a stairwell and through about 30 thugs. It was extremely challenging figuring out how to make that chain look semi-realistic and hit all of the story beats in a day or two. I ended up creating several chain models in various positions, and I would transition from model to model depending on what Daredevil was doing. It’s not uncommon for me to have to find or create assets that we may not have access to. Sometimes, we get lucky and I can get some plates of the props or stuntmen on wires that I can composite into the stuntvis.”

A scene from season 2 of Daredevil sees the titular character face off against The Punisher. Fight scenes were choreographed by Philip J. Silvera and regularly involved stuntvis augmentations. (Image copyright © 2016 Netflix)

A LARGE PART OF THE PROCESS

“I prefer to call stuntvis ‘action design,’ as it involves a lot more than stunts and requires the involvement of many key departments,” says stuntvis artist Yung Lee, who has worked on films including Kingsman: The Golden Circle and Solo: A Star Wars Story. “The entire process includes the development of scene blocking, choreography, stunt testing and rigging, set design, camera placement and movement, lighting setup, VFX breakdown, editing, and prop and weapon conceptualization.”

Lee collaborates regularly with Stunt Coordinator Brad Allan’s Action Design Inc. team, which specializes in designing high-end fight and stunt choreography. “My role involves editing the action sequences from the start of storyboard animatics to the stuntvis edits and finally to on-set editing during the production shoot, which I then pass onto editorial,” says Lee. “One of my key roles is to design/shoot/direct concept action scenes involving a lot of VFX elements. So if we have a piece in a fight scene involving a heavy VFX action beat, I will usually take charge in helping to flesh that concept out. Probably one of the most fun parts of my role is doing the stunt-camera tests – things like getting rigged up on a wire, and being dropped 70 feet down to the ground while falling with the performer and trying to keep him in frame the whole time!”

“The ability to quickly create 3D animation with real-time rendering is extremely useful on set. … I believe it is going to play a big role in the future of filmmaking.”

—Yung Lee, Stuntvis Artist

A final shot from the opening taxi cab chase in Kingsman: The Golden Circle. (Image copyright © 2017 20th Century Fox)

Stuntvis helped make parts of the conveyex raid in Solo possible. (Image copyright © 2018 Lucasfilm Ltd.)

Stuntvis incorporating stunt performers was generated for some of the robot Jaeger battles in Pacific Rim Uprising. (Image copyright © 2018 Universal Pictures)

On Kingsman: The Golden Circle, Lee worked closely with fight coordinator Guillermo Grispo to help imagine the diner showdown sequence, a long fight between several characters that plays out in stitched ‘oner’ shots, something the team called ‘hyper-cam’ that was first tested using an iPhone (Imageworks crafted the final visual effects work). “I was responsible for making all the hidden cuts in that scene work,” notes Lee, “while Guillermo would adjust choreography accordingly to help the stitch work better. Together we created the first pass of that sequence as Brad oversaw the performances. Then we had our on-set camera operator, Chris Cowan, come on board as an additional action designer to help spice up some of the camera moves for the sequence in the later passes.”

The Golden Circle also features an elaborate taxi cab chase through the streets of London, full of fast hand-to-hand combat and some seemingly impossible camera moves as the cars race through the streets. (The Third Floor also provided previs of the action, while Framestore delivered the final VFX.) The Action Design Inc. team imagined the taxi chase with a real taxi mock-up and some partial, even cardboard, set pieces and stunt performers in a rehearsal space, which stuntvis then used to help stitch shots, add weapon effects and even sparks to the shots.

A final shot from the convoy sequence in Deadpool 2, in which Deadpool and Cable fight on a moving prisoner transport truck. (Image copyright © 2018 20th Century Fox)

A scene from the opening speeder chase on Corellia in Solo. (Image copyright © 2018 Lucasfilm Ltd.)

Director Matthew Vaughan and actor Taron Egerton discuss a scene from the taxi cab chase in Kingsman: The Golden Circle. (Image copyright © 2017 20th Century Fox)

TOOLS OF THE TRADE

Stuntvis is regularly delivered on tight deadlines, and often requires quick iterations and constant collaboration between stuntvis artists and editors. For that reason, the usual tool of choice is Adobe After Effects. Clements says he also utilizes Video Copilot’s Element 3D to drop in CG models – typically weapons – to shots. “I also have a ridiculous library of blood hits and muzzleflash assets that I’ve compiled over the years,” he adds. “One of my favorite stuntvis pieces was in season 2 of Daredevil when the Punisher was locked up and fought virtually every inmate in the prison. I think I went through every blood asset that I had for that one. It was also fun coming up with a way for the blood to soak into the Punisher’s white jumpsuit to create the Punisher skull logo there.”

Lee is also an advocate for Adobe After Effects, which he uses hand in hand with Adobe Premiere for editing, oftentimes delivering quick on-set composites. A recent addition to his arsenal, however, has been game engines. “I decided to use some Grand Theft Auto V footage I shot in-game to use in our taxi stuntvis for The Golden Circle,” says Lee. “The game’s graphics are great. I was playing it at the time, and used the editor in the game to quickly and easily block out a shot, which I then comped our taxi into in After Effects. Then in Solo, working with Brad Allan again, I used Unreal Engine to add speeders, skiffs, and the AT-hauler to a scene. It never took very long, generally 20 minutes to an hour for each shot once I got used to it, and the renders came out so clean and quick in about only a minute or two.

“The ability to quickly create 3D animation with real-time rendering is just extremely useful on set,” continues Lee. “I had used it for moments in the speeder chase [in Solo] as I would animate and render out Moloch’s speeder into some shots very quickly and easily for Brad to see the rough idea of the final shot. Props to ILM for supplying me with the assets on demand. I believe it is going to play a big role in the future of filmmaking.”