Silero VAD v10 Enterprise Dashboard

ADVANCED ACOUSTIC & SPEECH TRANSMISSION DIAGNOSTICS ● ANSI S3.5 & STI MTF Live Engine Active

Speech Intelligibility (SII)

0.880

Good Intelligibility

Transmission Index (STI)

0.740

Good Transmission

Speech Clarity (C₅₀)

+5.2 dB

Clear Early Energy

Reverberation Time (RT₆₀)

0.45 s

Optimal Room Decay

Vocal Strain (Jitter/Shimmer)

0.65% / 1.8%

Healthy Voice

Formants (F₁ / F₂)

520 / 1650 Hz

Normal Articulation

Perceptual Audio MOS

4.2 / 5.0

High Perceptual Quality

Subsystems & Real-Time Health Diagnostics ● Diagnostics Active (60 FPS)

1. ONNX WASM EngineHEALTHY (100%)

Silero ONNX session ready; warnings suppressed.

2. Signal DSP & HPF FilterHEALTHY (100%)

80Hz Biquad filter & multi-band spectral estimator active.

3. Boundary HysteresisHEALTHY (100%)

400ms lookback ring buffer & hangover armed.

4. YIN Pitch (F₀) TrackerHEALTHY (100%)

Autocorrelation fundamental frequency lock active.

5. Speaker BiometricsSTANDBY

128-dim acoustic vector gate ready for enrollment.

6. Multimodal Vision FusionIDLE

Webcam lip & facial motion stream standby.

7. Local Speech STT HookREADY (100%)

Local speech-to-text inference dispatch active.

8. IndexedDB StorageHEALTHY (100%)

Persistent database connected (SileroV10EnterpriseDB).

Live Microphone Hardware & AGC Health Inspector

Select Microphone Hardware Device:

Device Name / Label

Default Microphone

Sample Rate

48.0 kHz

Channels & Latency

1 (Mono) | 10.0 ms

Auto Gain Control (AGC)

ENABLED ✓ (Suitable)

Hardware Noise Suppression

ENABLED ✓

Echo Cancellation (AEC)

ENABLED ✓

VAD Hardware Suitability Grade

EXCELLENT FOR VAD (100%)

Control & Dual Gain Hub

Dual-Stream Gain Multiplier (Path B)

Parallel Path B Gain Boost +6.0 dB (2.0x)

Adaptive Thresholds (Auto-Floor)

YIN Vocal Pitch Filter (F₀)

Target Speaker Isolation (≥ 0.70)

Visual VAD Sensor Fusion

Hysteresis & Boundary Controls

Speech Start Threshold 0.55

Speech End Threshold 0.35

Pre-Speech Lookback Pad (ms) 400

Redemption Silence Hangover (ms) 1000

Real-Time Acoustic & Dual Gain Stream State: Silent

Acoustic Prob

0.0 %

Path A RMS (0dB)

0.00

Path B RMS (+6dB)

0.00

Est. SNR

0 dB

Pitch (F₀)

0 Hz

Whisper Detect

Normal

Speech Probability Gauge0%

— Probability — Path A RMS (0dB) — Path B RMS (+X dB Boost) ... Noise Floor --- Start Threshold

Detections: 0 Misfires: 0 Total Speech: 0.0s Session: 00:00

Detected Voice Segments & Transcriptions

All Speech Misfires

#	Timestamp	Event	Duration	Pitch	SII / STI / MOS	Audio Clip	STT Transcription
No voice activity recorded yet. Click "Start Detection Engine" to begin.

⚠️ Engine Initialization Failed