Capturing animation by webcam and by video: what each one demands of the file
Studio turns video into keyframes in two ways: a webcam for the face, up to 60 seconds, and full-body video for an R15 rig, under 15 seconds and in a single take.
Produced with AI from the sources cited below, and checked by automated verification against those same sources. No person read this text before publication. Editorial Policy
Animating by hand is expensive. A decent walk cycle takes hours; a convincing facial expression takes longer. That is why Studio's two capture tools deserve attention even from people with no intention of abandoning manual animation.
They turn video into keyframes — and that word changes everything, because the result is not a video pasted onto the character.
The problem
The official proposition: you can record or upload video to the animation editor to capture movement and expression as keyframes. And the promise comes with the caveat that makes it useful: the tools create realistic animations in minutes, and still give access to the keyframes for further adjustment.
That second half is what separates capture from recording. If the result were a closed block, it would be an effect. Being editable keyframes, it is a starting point — you can move, delete, smooth and adjust everything afterwards, in the same editor where you would animate by hand.
There are two tools, with distinct purposes:
- Face: record your face with the webcam to animate characters with an animation-compatible head.
- Body: upload full-body video to generate realistic animation of the whole character.
And both share one barrier to entry: they are beta features, which have to be switched on by hand. Anyone looking for the button without enabling the feature concludes the tool does not exist — and that is probably the reason so few people use them.
Step by step
Face capture.
1. Enable the beta feature through the File menu, under Beta Features, in the facial capture option.
2. Prepare the environment. This step is not decorative, and the documentation explains why: be in a well lit environment and close enough to the camera, with your face in the centre of the frame. That improves the camera's ability to distinguish you from the environment, which is what produces higher quality animation.
Translated: the animation's quality is decided before you press record. Backlighting, distance and a face in the corner of the frame produce poor keyframes that no later adjustment recovers well.
3. Open the editor. In the Avatar tab, click Clip Editor. The animation editor window appears.
4. Select the rig in the viewport or the Explorer, give the animation a name and confirm.
5. In the track list, click Face Capture. A notice appears asking you to acknowledge consent to the collection of your facial movements, and camera access. With more than one camera, you can choose which to use.
6. Record. The red circle in the controls widget starts it; the red square or the timer ends it. Face capture records animations of up to 60 seconds.
7. Adjust and export. The keyframes corresponding to the facial movement appear on the timeline, and you can adjust them, save and export them for use in all your games.
Body capture.
The flow is similar, with one fundamental difference: instead of recording, you upload a file. It generates full-body animation for R15 rigs, and the documentation is direct about the gain — it considerably speeds up the manual process of creating animation.
The beta feature is a different one, also under File. And the file has to be .mp4 or .mov and meet specific requirements, which are the subject of the artifact.
The route: Clip Editor, select the R15 rig, name the animation, go to the playback controls, open the menu and choose to import through the live creation tool. In the dialog, the button is Choose Video.
After about a minute, the movement's keyframes appear on the timeline. With the import successful, you can save and export.
One difference between the two that changes planning. Face capture is immediate: you record and the keyframes appear. Body capture has a wait of about a minute per video, which sounds small and changes the workflow — you cannot iterate on framing in real time, so it is worth recording several takes before importing the first.
And there is a difference in commitment. Face capture asks for explicit consent to the collection of your facial movements, through a notice that has to be accepted. Body capture works on a file you choose to upload. They are two different relationships with your own image, and that is worth knowing before asking somebody on the team to record their face.
The artifact
The document below puts the two tools side by side — input, what they animate, length, where to enable them and waiting time — followed by each one's script.
The most useful part is the body video requirements list, because it is a list of elimination: a video that fails any item does not work, and most videos people already have fail at least one.
It is worth saying what the requirements list reveals about how it works: almost all of them exist so the tool can separate the person from everything else. One person only, well lit, visible the whole time, a stable camera, no cuts. Each item on the list removes an ambiguity that would make the tracking lose the body — and that is why breaking one of them does not degrade the result slightly, it produces an animation that cannot be used.
Common mistakes
The video that is too long. The limit is less than 15 seconds. That is far less than intuition suggests — fifteen seconds is a gesture, not a scene. Anyone filming a sequence of movements to use it all at once has to cut, and cutting runs into the next requirement.
A cut in the middle. The video has to be a single, continuous take. Any edit, any cut, any transition invalidates it. Which means body capture does not accept edited video — it accepts raw footage.
The two requirements together define the method: record several short, separate takes, and import one at a time. Trying to fit three movements into fifteen seconds produces a rushed animation that is no use for anything.
More than one person in the frame. The video has to have only one person, well lit and visible the whole time. "Visible the whole time" rules out more videos than it seems: a hand leaving the frame, a turn that hides half the body, somebody walking past in the background.
An unstable camera. The requirement asks for a stable camera. Handheld phone video introduces movement the tool has no way of separating from the person's movement — and the result is a character who travels without having taken a step. A tripod, or the phone propped up, solves it.
Bad lighting in face capture. It is the same mistake on the webcam side, and the documentation warns about it before the record button exists. Backlighting is the classic case: a window behind you turns your face into a silhouette, and a silhouette has no expression to capture.
Expecting the result to be finished. Both tools deliver keyframes, not finished animation. Face capture picks up micro-expressions you did not intend; body capture picks up balance wobbles. Later adjustment is part of the process — and that is precisely why the output is editable.
It is also worth combining the two. A body animation captured from video and a face animation recorded from a webcam are independent tracks, in the same editor — so an NPC who gestures and talks can have the body from one take and the expression from another. It is the kind of result that, done by hand, would be too expensive for a secondary character.
And keep in mind the real limitation of both: they capture what happened in front of the camera. Exaggerated cartoon movement, an impossible leap, a stylised reaction — none of that comes from capture, because nobody can perform it in front of a webcam. Capture solves the natural; the extraordinary is still an animator's work.
Next step
The use with the biggest return is not replacing manual animation. It is filling the gap — those animations the game needs and that nobody is going to spend a day making.
An NPC who gestures while talking, a character reacting to something, a victory expression, a wave, a shrug. They are animations that rarely justify hours of work, and that capture delivers in minutes with a naturalness that is hard to imitate by hand.
The practical exercise: make a list of ten small animations your game would have if they were cheap. Record all ten in one afternoon, in short, separate takes, with the phone propped up and good light. Import them one by one, adjust whatever bothers you, export.
And keep the original video of each one. Since the output is editable but the capture is not repeatable without re-recording, the source file is what lets you try again with a different framing — and it is the thing people throw away without thinking.
Sources
Ready-to-use artifact
ANIMATION CAPTURE — the TWO tools. Checked on 23/09/2026.
Both are BETA features and have to be switched on by hand before they
appear. Anyone looking for the button without enabling the feature
concludes the tool does not exist.
| Criterion | FACE (webcam) | BODY (video) |
|-----------------|------------------------------|------------------------------|
| input | your webcam, live | an .mp4 or .mov file |
| what it animates| an animation-COMPATIBLE | a whole R15 rig |
| | head | |
| length | up to 60 seconds | LESS than 15 seconds |
| enable at | File > Beta Features > | File > Beta Features > |
| | Face Capture | Live Animation Creator |
| wait | immediate | about one minute |
== FACE: the script ==
[ ] enable the beta feature (Face Capture)
[ ] a WELL LIT environment
[ ] close enough to the camera, face in the CENTRE of the frame
^ this improves the camera's ability to distinguish you from the
environment, and it is what decides the quality of the result
[ ] Avatar tab > Clip Editor
[ ] select the rig in the viewport or the Explorer
[ ] name the animation > Create
[ ] track list > Face Capture button
[ ] accept the consent notice about collecting your facial movements
[ ] (optional) choose the camera, if there is more than one
[ ] RED circle to record
[ ] RED square (or the timer) to stop
[ ] adjust the keyframes, save, export
== BODY: FILE requirements ==
The video has to meet all THREE, not two:
[ ] only ONE person, well lit and visible the WHOLE time
[ ] a SINGLE, CONTINUOUS take -- no cuts
[ ] LESS than 15 seconds, a STABLE camera
[ ] it meets the community standards and the terms of use
== BODY: the script ==
[ ] enable the beta feature (Live Animation Creator)
[ ] Avatar tab > Clip Editor
[ ] select the R15 rig
[ ] name the animation > Create
[ ] playback controls > menu > Import > Live Animation Creator
[ ] Choose Video > pick the file > Open
[ ] wait about ONE MINUTE
[ ] save and export
== WHAT BOTH HAVE IN COMMON, AND IT IS WHAT MATTERS ==
Both return EDITABLE KEYFRAMES, not a video pasted onto the rig.
That means the capture's result is a STARTING POINT: you can move,
delete, smooth and adjust everything afterwards, in the same editor.
