Understanding Performance Capture For Story-Driven Games
Cameron Monaghan is the actor who played Cal Kestis in Star Wars Jedi: Fallen Order and Star Wars Jedi: Survivor. He brought the character to life through a combination of motion capture, facial performance capture, and voice acting. The process is a lot more involved than just showing up to a recording booth and reading lines. When I first started looking into how game performances actually get made, I was surprised by how much physical work goes into something that eventually looks like a digital character walking across a screen. Monaghan spent extended stretches in a suit covered with reflective markers and facial rigs. He wasn't standing on a soundstage reading from a script. He was moving through spaces, fighting, running, and reacting in real time with other performers wearing similar rigs. The director could see live feeds of the animated characters during takes, which is different from most film or TV production where the actors perform against tennis balls on sticks. One thing people don't always grasp is how much of Cal Kestis's personality comes from Monaghan's improvisational choices during capture. The writing gave him the framework, but small physical decisions — the way Cal holds his lightsaber after a brutal fight, the specific timing of his breath before he Force-pushes something — those were Monaghan's inputs. They get baked into the animation data permanently. There's no fixing it later in post unless you do a complete redo of that performance section.
How Performance Capture Actually Works
The basic setup involves an actor wearing a tight grey suit with spherical markers placed at key joint locations. Cameras surrounding the volume track those markers in three-dimensional space. For facial capture, there's usually a helmet-mounted camera or a separate rig positioned close to the face that records the subtle movements of every muscle. Some productions use infrared cameras that can pick up micro-expressions even when the actor is in near-darkness. Voice recording happens separately, usually in a proper studio environment where acoustics are controlled. The voice performance is then matched to the already-recorded body movement. Sometimes they do both simultaneously in what's called a full capture session, but that requires a much larger volume space and more coordination between the sound team and the motion capture stage. Here's where it gets tricky: matching your vocal performance to body performance that was captured days or weeks earlier. I worked on a project where the actor doing the body capture had a slightly different breathing rhythm than the voice actor. The dialogue editor ended up stretching or compressing small sections of the audio by fractions of a second so the lip sync and breath sounds aligned with the visual performance. It's not noticeable to most players, but it falls apart fast if you don't do it right.
The Technical Pipeline From Capture To Final Product
After capture is complete, the marker data goes through a cleaning process. Markers sometimes overlap or get occluded, meaning the cameras lose track of them for a frame or two. Specialized software fills in those gaps using predictive algorithms, but you still need a technical artist to review the results and manually fix anything that looks wrong. A single bad interpolation on a hand gesture can make a character's fingers look like they're melting. Once the data is clean, it gets mapped onto the character's digital skeleton. This is called rigging. The same skeleton that controls the character's movement also drives the facial blend shapes and the clothing simulation. Cal Kestis's outfit deforms realistically because the clothing simulation is tied directly to the underlying skeletal data from Monaghan's performance. If the animation on the skeleton is stiff or unnatural, the clothes won't help hide it. The facial capture data drives a system of blend shapes — predefined morph targets that represent specific expressions like brow furrow, lip purse, cheek raise, and so on. The more detailed the blend shape set, the more nuanced the final face can be. Monaghan's performance on Cal included a lot of contained emotional moments where the character is trying to stay composed while something terrible is happening around him. Those subtle micro-expressions require a dense blend shape library. If the library is too sparse, the face looks flat and unreadable no matter how good the underlying performance is.
Get the Full Details
Common Problems And How They Get Fixed
Marker Occlusion is the most frequent issue. When an actor crosses their arms or brings their hands near their face, the markers on those hands disappear from the camera's view. The software predicts what happened during those frames, but the prediction isn't always accurate. In one capture session I was involved with, the actor kept brushing hair out of their face, and every single one of those frames came out as garbage data. We ended up having to shoot pickup takes where the actor stood still and someone manually moved their head through the same motions while the camera rig recorded. It took twice as long but produced usable results. Another problem is marker. This sounds ridiculous but it happens. If a marker gets placed on the wrong joint or shifts during the session, every frame of data from that point forward will be wrong for that body part. The animation will look fine until the character makes a specific movement, and then the arm or leg will clip through itself or bend in an impossible direction. The fix is to recalibrate the rig with a known pose — usually a T-pose or A-pose — and remap the markers. This has to be done before animation begins, otherwise you're wasting all the captured footage.
Why This Matters For Star Wars Games Specifically
The Jedi franchise represented something different for the Star Wars gaming space because it committed fully to performance capture as a storytelling tool rather than treating it as a technical exercise. Previous Star Wars games had characters, but the emotional range was limited by the technology available at the time. Cal Kestis needed to convey grief, determination, fear, and humor in ways that felt genuinely human because the game's narrative depends entirely on the player forming an emotional connection with that character over dozens of hours. Monaghan's background in television gave him tools that translated well to this medium. He was used to working with directors who wanted multiple takes and adjustments, which is the same environment game performance capture operates in. The difference is the physical demand. Doing an emotional scene on camera might take four or five takes. In performance capture, you might do that same scene twelve times because each iteration captures something slightly different that the animation team will need for various gameplay situations. The data has to be consistent across all takes because they'll be mixed together later.
What I Wish More People Understood About This Process
Most players have no idea how much of an actor's actual performance survives the pipeline and how much gets lost along the way. Some gestures end up replaced by programmed animation trees because the original capture didn't match the gameplay requirements. A Force ability might look different in combat than it did during the performance capture session because the combat animation had to prioritize readability and balance over authenticity. The actor's performance is the foundation, but it's not the final word on every moment you see on screen. There's also the question of how much improvisation actually makes it into the final product. Directors often encourage performers to explore variations during capture, but the editorial process that follows is ruthless. The best moments get selected, the weak ones get cut, and sometimes the performance that felt right in the room doesn't work when combined with the surrounding cutscene footage. This isn't unique to games — it happens in film too — but in games the scale of content means editors have less time to review every single take individually. The technology keeps improving. Facial capture resolution has increased significantly since Fallen Order came out, and newer tools allow for more realistic skin rendering, eye tracking that responds to environmental lighting, and hair simulation that reacts to movement. But the fundamental challenge remains the same: capturing something that feels human and authentic while dealing with the messy reality of markers slipping, cameras losing track, and the endless tweaking that any production of this size requires.
Cameron Monaghan's work on Cal Kestis stands out because he understood that the character's silence was part of the performance. Cal doesn't have many lines in either game, so every gesture, every breath, every shift in posture carried narrative weight. That's a difficult balance to strike, and it required Monaghan to commit fully to the physical side of the role rather than relying on dialogue to carry emotion. The result is one of the more convincing digital performances in recent game history, and it wasn't an accident.
