About VRC FaceTracking
This page explains how a VRChat avatar built with VRC FaceTracking (VRCFT) is treated on the LiveStudio side when brought in via VRChatAvatarTransfer.
What is VRC FaceTracking?
VRChat itself does not natively support facial tracking. To fill that gap, the community-made tool VRCFaceTracking (VRCFT) and a common blendshape naming convention called Unified Expressions are widely used.
- A wide variety of input devices is supported (Vive Pro Eye / Quest Pro / Pimax Crystal / iPhone ARKit, etc.)
- If the avatar is set up with blendshapes named after Unified Expressions, the face will move the same way regardless of which device is used
- Among VRChat avatars, those marked as "FT compatible" have these blendshapes built in
Differences from VirgoMotion's Face Tracking
VirgoMotion uses the 52 blendshapes (PerfectSync) output by the iPhone's ARKit as-is.
| Item | VirgoMotion (PerfectSync) | VRC FaceTracking (Unified Expressions) |
|---|---|---|
| Input device | iPhone ARKit only | Many devices unified |
| Blendshape naming | ARKit standard names (jawOpen, eyeBlinkLeft, etc.) | Unified Expressions names (JawOpen, EyeClosedLeft, etc.) |
| Number of shapes | 52 | 60+ (adds finer shapes for tongue, cheeks, mouth, etc.) |
The two are different, but the ARKit-derived shape group is largely shared. In practice, many VRCFT-compatible avatars also carry ARKit / PerfectSync names, so in many cases the face moves correctly in VirgoMotionStudio as well.
For a reference table mapping the 52 PerfectSync shape names to their corresponding facial regions, see the link below. Let's make a PerfectSync-compatible model! Production notes on 52 BlendShapes for iPhone face tracking - Hinukka Memo
Behavior When a VRChat Avatar is Brought In
When you import an avatar output by VRChatAvatarTransfer into VirgoMotionStudioTemplate,
how the face moves is determined by the avatar's blendshape configuration.
By attaching a VRCFTAvatar component to the avatar itself, ARKit parameters are converted into VRCFT-compatible Unified Expressions parameters and applied.
| Avatar Configuration | Component |
|---|---|
| VRCFT-compatible Avatar | VRCFTAvatar |
| PerfectSync-compatible VRM | VRM10Avatar (PerfectSync setting) |
VRM (only simple presets like Joy / Angry) | VRM10Avatar (Preset setting) |
Where to Configure?
- Expression adjustments (intensity, interference avoidance, etc.) are done in VirgoMotion's PerfectSync expression settings. See Expression Settings (PerfectSync) for details.
- Adding or renaming blendshapes on the avatar itself is a matter of VRM / FBX editing in Unity.
Sources
VRCFaceTracking (VRCFT) is an open-source project.
- Official documentation: docs.vrcft.io
- GitHub: benaclejames/VRCFaceTracking
VirgoMotion / LiveStudio itself does not bundle the VRCFaceTracking code. The Unified Expressions blendshape naming convention referenced here is maintained by the VRCFaceTracking project, and the LiveStudio SDK provides functionality to manipulate facial expression parameters on avatars compatible with that naming.
Related Links
- Use VRChat Avatars — Steps for bringing VRChat avatars into VirgoMotionStudio
- Expression Settings (PerfectSync) — PerfectSync-side expression adjustments
- VRCFaceTracking GitHub
- VRCFaceTracking official documentation — Unified Expressions specification