Select an avatar and open Tracking to configure how movement and microphone input affect that avatar.
App Tracking vs. Avatar Tracking
The app’s Tracking page manages sources and avatar assignments. The selected avatar’s Tracking section contains its movement controls and Microphone Lip Sync setting. Use Face Tracking, Pose Tracking, or Webcam Tracking to assign sources.
Controller Movement selects a controller profile and configures movement speed and response. VRMs also support walking clips and facing the direction of travel; see Controllers. These controls change the avatar instance’s placement and are saved with the scene.
Microphone Lip Sync chooses whether microphone mouth movement is off, used only when face tracking is unavailable, or used all the time. Select and calibrate the microphone on the app’s Lip Sync page. This preference belongs to the avatar instance.
Live2D additionally provides Whole-Model Movement, described below. Unlike the instance controls above, it is saved with the model and shared by its loaded instances.
Whole-Model Movement
Applies toLive2DLocated in the avatar's Tracking section, Whole-Model Movement corresponds to VTube Studio's Movement Config: the avatar translates horizontally (X) and vertically (Y) with tracked head positions, and scales (Z) as you lean toward or away from the camera.
Whole-Model Movement adjusts the avatar's placement on the stage canvas rather than driving model parameters. It stacks additively with any FacePosition* parameter mappings configured in Parameter Bindings—both read the same tracking stream without scaling each other. If a rig already binds its own Z-axis parameter, set depth to 0 or disable the master switch.
Controls include:
| Option | Description |
|---|---|
| Toggle | Enables or disables whole-model movement (enabled by default) |
| Horizontal | Distance the model slides as your head moves sideways. Range is -50 to 50; negative inverts, default is 6 |
| Vertical | Distance the model slides as your head moves up and down. Range is -50 to 50; negative inverts, default 8 |
| Depth | Scale increase as you lean toward the camera. Range is -50 to 50; negative inverts, default 11 |
| Depth Pivot | Sets the point the model grows around as you lean in or out. Defaults to model center, or can be chosen from the list or picked on stage by clicking an ArtMesh |
Movement configuration is stored per model rather than per instance:
- If
.vtube.jsonexists alongside the model, settings are written to itsModelPositionMovementsection, which keeps them compatible with VTube Studio in both directions. - Otherwise, they are saved to
movementin.persona.json. - The one exception is Depth Pivot, which VTube Studio has no field for. It is stored in
.persona.jsoneither way, and evaluated at the point's position in the model's rest pose, so head motion never moves the model.
Sliders range from -50 to 50, while VTube Studio's own sliders stop at ±20. Larger amounts load as written in both apps, but releasing any movement slider in VTube Studio writes all three axes back clamped to ±20.
Changes are saved when you release a slider and apply immediately to all loaded, unpinned instances of the model. Smoothing levels are read once from the configuration file (default 10).
Last updated on October 21, 2026