Professional Facial Mocap, Made Simple.

The AnimStarter Face Mocap ecosystem is a high-performance, frictionless pipeline engineered from the ground up for professional 3D animators, technical directors, and indie studios. Stop renting your workflow. Capture pristine, high-fidelity facial data on your iPhone, import it into Maya seamlessly, and retain absolute control over your production assets.

✨ Key Advantages & Production Benefits

• ⚡ One-Click Rig Profiles (.asface): Eliminate repetitive technical setups. Map the 53 captured blendshapes to your character’s specific facial rig controls once, save it as an .asface profile, and load it instantly for all future takes. Retargeting has never been this fast.

• 🎵 Automatic WAV Audio Extraction: Perfect lip-sync out of the box. When you finish a take, the iOS engine automatically extracts a clean, uncompressed .wav audio track (16-bit / 44.1kHz) alongside your motion data, ready to be imported directly onto your Maya timeline.

• 📊 Open Pipeline Architecture (.asfm): No proprietary black boxes. The .asfm format is a fully documented, standardized CSV structure containing 60 columns of dense data (Timecode, Blendshapes, and 6-axis Head transforms). While optimized for Maya, it is fully adaptable for Blender, Unreal Engine, or custom studio pipelines.

• 🔒 Standalone & Frictionless (DRM-Free): Built for real-world production environments. The Maya Python desktop tool uses a trust-based "soft-license" system—no internet dependency, no floating license managers to crash during late-night deadlines, and absolutely no subscriptions. The companion iOS app is 100% free.

📱 THE BACKEND: AnimStarter Face Mocap iOS App

Leveraging the iPhone’s advanced hardware, the app captures 53 discrete blendshape channels (including custom tracking like tongueOut) and complete 6-axis head orientation at a rock-solid 60 frames per second.

1. Capture Best Practices

To achieve performance-capture fidelity that eliminates hours of manual clean-up:

• Lighting: Use soft, diffused frontal lighting. Avoid harsh side-shadows or heavy backlighting that can distort the infrared TrueDepth sensor grid.

• Distance: Keep the device between 35 and 50 cm from the actor's face for optimal depth-map resolution.

• Neutral Anchor: Always begin your recording with 1–2 seconds of a completely relaxed, neutral facial expression to calibrate baseline values.

2. Adaptive Jitter Removal: The One Euro Filter

Raw performance data inherently contains high-frequency noise and micro-trembles. To solve this without adding processing latency, the app integrates the industry-standard One Euro Filter—an adaptive low-pass filter that scales dynamically based on velocity. It applies heavy smoothing when your face is relatively still, and instantly reduces filtering during fast, explosive expressions to preserve pure raw emotion.

• Standard: Perfectly balanced (minCutoff: 1.0, beta: 0.05). Recommended for 90% of facial acting and dialogue.

• Light: Minimal filtering (minCutoff: 2.0, beta: 0.10). Best for hyper-expressive, rapid martial arts or high-energy cartoon action.

• Heavy: Strong smoothing (minCutoff: 0.5, beta: 0.02). Designed for slow, subtle dramatic close-ups or stylized, heavy characters.

3. Desktop-Grade Math: The Vertex Solver

For elite-level accuracy, toggle the Vertex Solver. Instead of relying solely on standard ARKit approximations, the solver tracks the raw, dense geometry of your face (1,220 vertices) in real time using Apple's Accelerate hardware framework.

• The Math: It executes an on-device Non-Negative Least Squares (NNLS) ridge regression algorithm, solving for mathematically optimal blendshape weights native to your face shape. It then blends this data to generate an ultra-stable _solved.asfm file.

• The Calibration View: Requires a one-time, 40-second configuration per actor. Simply follow the on-screen emoji prompts, holding each key expression firmly. The resulting basis matrix is saved locally on your device for all future recording sessions.

​

​

💻 THE FRONTEND: ASFM Retargeter for Maya

Once you transfer your .asfm and .wav files to your workstation via AirDrop or local storage, the Python desktop plugin handles the heavy lifting inside Autodesk Maya.

Step 1: Automated Interface Referencing

When you hit "Import into Maya", the script automatically references a dedicated visual curve interface (AS_FACE_UI) into your scene. Your raw animation data is baked directly onto this modular panel, keeping your production scene clean and isolated from the character geometry.

Step 2: Smart Framerate Retiming

Facial capture runs natively at 60 FPS, but film and game timelines often sit at 24, 25, or 30 FPS. The ASFM Importer features an advanced Retime Strategy that downsamples and resamples keyframes linearly, preserving perfect timing and performance length without altering scene speed. Alternatively, you can force the Maya timeline to match the native capture rate in a single click.

Step 3: Advanced Channel Control & Mapping

The retargeting window gives you surgical control over how performance data drives your character:

• Coefficients & Offsets: Scale individual channels (e.g., boosting eye directions by a factor of 100.0 or dampening an over-expressive jaw) and add custom value offsets directly inside the UI.

• Head Space Conversion: The importer takes care of complex positional transformations, neutralizing the world-space camera offsets from the iPhone and converting them into clean, local-space values ready for your character's neck controllers.

• Duplicate Detection & Multi-Binding: The UI automatically flags duplicate targets and seamlessly allows multiple capture channels to combine and drive a single controller attribute smoothly.

Step 4: Exporting Your .asface Master Profile

Once your character rig is perfectly dialled in, save your setup. Next time you record, skip the technical node mapping entirely: Load Profile -> Import Mocap -> Animate.