Avatar Models
Streamoji supports two primary avatar model types to optimize for different use cases and performance requirements.
Full-Body
Complete character model from head to toe. Ideal for VR, third-person games, social hubs, and applications where the full character is visible.
Mesh: ~15k-25k PolysTextures: 2K-4K resolutionFile Size: ~3-5MBHalf-Body
Optimized model cropped at the waist. Best for VR meetings, first-person experiences, and applications where only the upper body is visible. Reduces memory usage and improves performance.
Mesh: ~8k-12k PolysTextures: 2K resolutionFile Size: ~2-3MBModel Specifications
| Property | Full-Body | Half-Body |
|---|---|---|
| Polygon Count | 15,000 - 25,000 | 8,000 - 12,000 |
| Texture Resolution | 2K - 4K | 2K |
| Rig Type | Humanoid (Unity/Unreal compatible) | |
| Bone Count | ~75 bones (standard humanoid skeleton) | |
| Blendshapes | 52 ARKit-compatible facial blendshapes | |
| File Format | GLB (GLTF Binary) | |
Texture Maps
All avatars include the following texture maps:
Morph Targets & Blendshapes
All Streamoji avatars include a comprehensive set of facial blendshapes for realistic animation, lip-sync, and expression control.
ARKit Standard (52 Blendshapes)
Full ARKit-compatible blendshape set for facial tracking and animation:
Oculus Visemes (15 Phonemes)
Phoneme-based blendshapes optimized for speech animation:
Custom Expressions
Pre-built expression combinations for common emotions:
Using Blendshapes
Blendshapes can be controlled programmatically in your application:
// Three.js example
const mesh = avatar.scene.getObjectByName('Head');
const blendshape = mesh.morphTargetDictionary['eyeBlinkLeft'];
mesh.morphTargetInfluences[blendshape] = 1.0; // Fully close left eye
// Unity example
SkinnedMeshRenderer renderer = avatar.GetComponent<SkinnedMeshRenderer>();
int blendShapeIndex = renderer.sharedMesh.GetBlendShapeIndex("eyeBlinkLeft");
renderer.SetBlendShapeWeight(blendShapeIndex, 100f); // 0-100 rangeAnimation Rig
Streamoji avatars use a standard humanoid rig compatible with most animation systems and motion capture solutions.
Bone Structure
The avatar skeleton follows a standard humanoid hierarchy:
- Root/Hips: Base of the skeleton, controls overall position
- Spine Chain: Spine, Spine1, Spine2 for torso movement
- Arms: Shoulder, UpperArm, LowerArm, Hand with finger bones
- Legs: UpperLeg, LowerLeg, Foot with toe bones
- Head: Neck, Head for head movement
Animation Compatibility
Included Animations
Base avatar models include standard T-pose and idle animations. Additional animations can be applied using your preferred animation system.