VAD Tier 1 Edge

5-Level Acoustic Telemetry & Binary Engine Verification Suite

Mic: Off
VAD: Idle
ASR: Standby
Voiceprint: Unenrolled
Liveness: Clean
PTT: [Space]
Tier 1: Compute Engines & Security Health (Binary Yes/No)
WebGPU Engine: ✓ PASS (Metal M3)
WASM SIMD Vector: ✓ PASS (128-bit)
Vector Store (DB): ✓ PASS (Connected)
16kHz Resampler: ✓ PASS (Active Guard)
Air-Gap Privacy: ✓ PASS (0.00 KB Egress)
Tier 2: Native Hardware DSP Bypass Verification (Binary Yes/No) Pure Acoustic Passthrough
Auto Gain (AGC): ✓ BYPASS (Raw Range)
Noise Suppression (NC): ✓ BYPASS (Raw Phase)
Echo Cancellation (AEC): ✓ BYPASS (Direct Line)
Audio Quantization: ✓ PASS (32-bit Float)
Tier 3: Transducer Profile & Live Gain Staging (5-Level Live Snapshots) Dynamic Audio Sizing
Transducer Hardware: L1: Laptop Array
Signal RMS Level: -18.4 dBFS L1: Optimal
Dynamic Headroom: 14.2 dB L1: Clean
Pre-Gain Staging: 2.0x L2: Auto-Boost
Tier 4: 5 Core Ambient Room Measurements (5-Level Physical Snapshots) ISO/IEC 30107 PA Tested
1. SNR / Floor: 16.8 dB L2: Moderate
2. Reverb (T60): 240 ms L1: Optimal
3. Direct Path (DRR): +9.4 dB L2: Moderate
4. Clarity (C50): +4.2 dB L2: Moderate
5. Modal Resonance: ±1.4 dB L2: Moderate
FSM: IDLE HEAP: 148 MB SINR: 0.0 dB RT60: 240 ms BIO MATCH: 0.00 SYNTHETIC PROB: 0.00 INFER LATENCY: 0 ms ACTIVE TRIGGER: 0.40
Signal Presence Probability SILERO VAD ONNX
Confidence Score:
0.00
Neural Thresholds & Biometric Gating
ParameterBaselineLive ValueAdjust Range
Positive Speech Trigger 0.40 0.40
Biometric Match Threshold (θ_user) 0.78 0.78
Synthetic Liveness Cutoff (θ_synth) 0.75 0.75
Soft-Knee Pre-Gain 2.0x 2.0x
Whisper Model Tier tiny.en
Dual-Track Intelligence Stream Click card to play audio snippet
Real-Time Telemetry Stream 0 events