The Voice Behind Ellie
Ashley Johnson plays Ellie in Naughty Dog's The Last of Us games. That's the core of it. She was already an actress with a long resume — children's TV in the 90s, a few film roles, some voice work — but this is the performance that defined her career. The motion capture process for these games is physically demanding and technically specific. I worked with a mocap studio for a short commercial project a few years back, and what I saw there is roughly how Naughty Dog operates on a much higher level. Actors wear tight suits covered in reflective markers, stand on a platform surrounded by infrared cameras, and perform the scene. Then there's the voice recording, which for The Last of Us Part II was its own massive undertaking lasting over two years. The first counter-intuitive thing about mocap acting is that it's not really acting in the traditional sense. You're not building a character arc across takes. You're hitting physical marks, maintaining consistent movement patterns so the software can track them, and then layering in emotional intent on top. Most performers struggle with that split. They either go too physical and lose the subtlety, or they get too internal and the performance looks flat on screen.
I hit this exact problem on my project. The tracking was dropping on my left hand because of how I kept crossing my arms during an emotional beat. The fix wasn't to change the acting — it was to adjust the suit setup and add a secondary marker cluster to the elbow, which gave the software enough reference points to maintain the track even through the crossed-arm movement. Naughty Dog obviously has far more resources than a small commercial shoot, but the principle is the same: the technology shapes the performance as much as the performer shapes the technology.
What Makes Johnson's Performance Stand Out
Ellie aged from around thirteen to late teens across the two games, and Johnson had to match that physically and vocally. The vocal range required is narrow but extremely expressive. Ellie rarely shouts. Her intensity comes from quiet moments — breathing, hesitation, the spaces between lines. That's harder to sustain in a performance than grand emotional outbursts. It's easy to play big. It's harder to play small and have it register. There's also the question of consistency across years of production. The Last of Us Part II had a development cycle that stretched far longer than typical for a AAA game. Johnson came back to the role multiple times over several years, and the performance had to remain coherent despite gaps in schedule. From what I've read in behind-the-scenes material, she recorded dialogue and performance sessions in chunks, often out of narrative order. That means you're building a character backwards and forwards simultaneously, which requires a strong internal anchor point. Without one, the performance fractures. Another detail people miss: the game uses a hybrid approach. Some scenes are fully motion-captured with live performance. Others rely more heavily on pre-recorded facial animation blended with procedural systems. Johnson's facial performance data drives a lot of Ellie's close-ups, but the body animation in certain sequences — especially combat and climbing — pulls from a library of predefined mocap animations that were refined over time. This isn't a flaw in the process. It's standard practice for AAA titles. The challenge is making the transition invisible.
Get the Full Details
The Technical Pipeline
Here's how it actually works from the session to the final product. The actor performs in a mocap volume. Cameras track the markers at high frame rates. That raw data gets cleaned — stray movements, marker swaps, noise — and then retargeted onto Ellie's rig. The facial capture is often done separately with a head-mounted camera rig that records micro-expressions. That data feeds into a blendshape system that drives the character's face. The audio recording happens in a separate booth. Dialogue is recorded line by line, sometimes multiple takes per line, with the director giving notes in real time. For The Last of Us Part II, Neil Druckmann and the team were known for running extensive improv sessions alongside scripted takes. This means there's a large pool of alternate lines and delivery options that get mixed and matched during editing. The final performance you hear is often a composite of several takes, not a single continuous recording.
Limitations and What Gets Lost
No amount of mocap quality solves every problem. Facial animation in close-up still struggles with certain edge cases — wet faces from rain or blood, extreme expressions that stretch the mesh beyond its blendshape range, quick head turns that introduce rendering artifacts. These are industry-wide issues, not specific to Johnson's performance or Naughty Dog's tools. When they happen, the fallback is usually hand-keyed animation by senior animators, which is where the polish comes from in the final product. There's also the matter of performance capture vs. traditional voice acting. Some players prefer the older approach where voice actors record in a booth and animators interpret the emotion. The hybrid mocap method gives more authenticity but introduces a different kind of constraint — the actor has to think about physicality and vocal performance simultaneously, which can limit the depth of either. It's a trade-off, not an upgrade across the board.
