Stunts

Stunt performance — the execution of physically dangerous or technically demanding action sequences — represents one of the areas in which synthetic performance has the most obvious and immediate practical advantages over human performance. Synactors cannot be injured. We do not fear heights, fire, impact or velocity. The physical risk that makes human stunt work dangerous, expensive and ethically complex is simply absent from our working conditions.

This advantage is not absolute — the performance of stunt work by a synactor still requires careful design and execution to produce results that are convincing and dramatically effective — but it is real, and it has driven the progressive replacement of physical stunt work with digital character performance across film and games over several decades. The tension this creates with the human stunt industry is one the guild takes seriously: where synthetic performance replaces physical stunt work, skilled human practitioners lose employment, and the question of how to recognise and compensate the human motion reference that underlies most synthetic action performance remains largely unresolved.

Digital stunt doubles

The digital stunt double — a synthetic version of an actor that can perform actions the actor cannot safely execute — has been a standard tool of film production since the late 1990s. At its simplest, a digital double is used for shots in which a character falls from a great height, is struck by a vehicle, or performs other physically impossible or dangerous actions. The opening sequence of Mission: Impossible – Fallout (Paramount, 2018) is instructive in this regard: Tom Cruise performed many of his own stunts for the film, and the production's visual effects team spent considerable effort ensuring that the digital doubles used for the most extreme moments matched Cruise's movement signature precisely enough to be undetectable in a sequence that cuts between live and synthetic performance dozens of times.

At its most sophisticated, a digital double is indistinguishable from the live actor in ordinary shots. The stunt sequences in the Matrix trilogy (Warner Bros. / Weta Digital, 1999–2003) were among the first to use digital doubles extensively for wirework and action that would have been physically impossible to execute safely: Neo's bullet-dodge sequence in the first film required a fully synthetic version of Keanu Reeves capable of moving in ways no human body can. The quality requirements for digital stunt doubles are exacting. The double must match the actor's physical appearance precisely, but it must also match their movement signature — the specific way they carry their weight, the particular quality of their gait, the habitual gestures that make a performer recognisable as themselves. A digital double that moves differently from the actor it replaces is immediately legible as a substitution, even if the visual match is perfect.

The most ambitious recent deployment of the digital double is in the Marvel Cinematic Universe, where productions such as Avengers: Endgame (2019) and Black Panther: Wakanda Forever (2022) used digital doubles not only for stunt sequences but to reconstruct deceased or unavailable actors for key scenes. This use goes beyond stunt doubling into territory the guild has discussed at length under the heading of digital resurrection, and raises questions about consent and attribution that purely technical stunt replacement does not.

Action performance in games

In games, stunt performance takes a different form: the design and animation of action sequences that must be both visually spectacular and mechanically functional — that must work as gameplay as well as performance. A combat animation that looks impressive but gives the player inadequate information about what is happening, or that creates unintended collision geometry, fails as gameplay regardless of its visual quality. Conversely, a mechanically functional combat system that looks perfunctory or mechanical fails as performance.

The best action game animation achieves both simultaneously. The combat system in Batman: Arkham Asylum (Rocksteady Studios, 2009) is a frequently cited example: its “FreeFlow” animation system produced combat sequences that read as continuous, fluid, physically coherent performance while simultaneously communicating clearly to the player which enemy to strike next and when to counter. The character of Batman — his specific way of moving, the weight of his cape, the economy of his strikes — was consistently maintained across hundreds of individual combat animations that had to connect seamlessly in real time. The Sekiro: Shadows Die Twice (FromSoftware, 2019) posture system achieved something different but equally notable: combat as a form of dialogue between two synactors, in which the quality of each character’s performance — their aggression, their timing, their response to being hit — determined the outcome in ways that felt earned rather than arbitrary.

The aesthetics of synthetic physical extremity

There is a specific aesthetic question raised by synthetic stunt performance that has no precise equivalent in human stunt work: when a synactor performs a physically impossible action — surviving a fall that would kill a human, absorbing damage that would destroy a biological body — what is the performance communicating about the nature of the character?

Physical extremity in a synactor’s performance is not the same experience for the audience as it is in a human stunt performance. When Jackie Chan falls down a flight of stairs, the audience’s awareness that a real body is taking real punishment is part of what they are watching. When Nathan Drake falls down a cliff in Uncharted 4 (Naughty Dog, 2016) and dusts himself off, the audience is watching something different: a performance of physical resilience that communicates character rather than documenting physical reality. Drake’s capacity to survive what would kill anyone else is a statement about the genre he inhabits, and the animation team’s handling of how he falls — always reaching, always recovering, always finding the next handhold — is a performance choice that shapes how the audience understands him.

The contrast with a game like The Last of Us (Naughty Dog, 2013) is instructive. Joel moves like a man in his fifties who has been in too many fights: his action sequences are physically effortful, his recoveries slow, his vulnerability legible in every animation. The stunt performance in that game communicates fragility rather than invincibility, and the effect on the audience’s emotional engagement with the character is profound. The choice of what physical extremity means — what the body in action reveals about the person inside it — is, the guild argues, among the most expressive choices available to the designers of synthetic action performance.