WebRTC VAD Provider

The WebRTC provider implements voice activity detection using a dual-metric approach: RMS energy analysis combined with zero-crossing rate (ZCR). This pure Go implementation requires no external dependencies or CGO and effectively distinguishes voiced speech from background noise.

Choose WebRTC VAD when you need a lightweight, pure Go voice activity detector with zero external dependencies. It works well in controlled environments and requires no CGO or model files. The dual-metric approach (energy + zero-crossing rate) effectively rejects high-energy noise that would fool a pure energy detector. For higher accuracy in noisy environments, consider Silero.

Installation

go get github.com/lookatitude/beluga-ai/voice/vad/providers/webrtc

Configuration

Field	Required	Default	Description
energyThreshold	No	1000.0	RMS energy threshold for speech
zcrThreshold	No	0.1	Zero-crossing rate threshold

Registry configuration keys:

Key	Type	Maps to
threshold	float64	energyThreshold
zcr_threshold	float64	zcrThreshold

Basic Usage

package main

import (
    "context"
    "fmt"
    "log"

    "github.com/lookatitude/beluga-ai/voice"
    _ "github.com/lookatitude/beluga-ai/voice/vad/providers/webrtc"
)

func main() {
    vad, err := voice.NewVAD("webrtc", map[string]any{
        "threshold":     1000.0,
        "zcr_threshold": 0.1,
    })
    if err != nil {
        log.Fatal(err)
    }

    // audioPCM is 16-bit little-endian PCM audio data
    var audioPCM []byte // ... obtained from audio source

    result, err := vad.DetectActivity(context.Background(), audioPCM)
    if err != nil {
        log.Fatal(err)
    }

    fmt.Printf("Speech: %v, Event: %s, Confidence: %.2f\n",
        result.IsSpeech, result.EventType, result.Confidence)
}

Advanced Features

Dual-Metric Detection

The WebRTC provider uses two complementary metrics to classify audio:

1. RMS Energy — Measures signal amplitude. High energy suggests speech or loud sounds.
2. Zero-Crossing Rate (ZCR) — Measures how often the signal crosses zero. Noise tends to have high ZCR, while voiced speech has lower ZCR.

Speech is detected when energy exceeds the threshold and the zero-crossing rate is below the ZCR threshold. This combination rejects high-energy noise (fans, traffic) that would fool a pure energy detector.

Confidence Scoring

The confidence score reflects both metrics:

• Base confidence is the ratio of RMS energy to twice the energy threshold, clamped to [0, 1]
• When ZCR exceeds the threshold (suggesting noise), confidence is halved

result, err := vad.DetectActivity(ctx, audioPCM)
if err != nil {
    log.Fatal(err)
}

if result.Confidence > 0.8 {
    fmt.Println("High-confidence speech detection")
}

Threshold Tuning

Energy and ZCR thresholds can be adjusted independently:

Scenario	Energy Threshold	ZCR Threshold	Effect
Quiet room	500.0	0.1	More sensitive to soft speech
Noisy environment	2000.0	0.05	Strict filtering of background noise
Default	1000.0	0.1	Balanced for general use

Continuous Detection

The provider tracks state transitions between speech and silence:

for _, chunk := range audioChunks {
    result, err := vad.DetectActivity(ctx, chunk)
    if err != nil {
        log.Fatal(err)
    }

    switch result.EventType {
    case voice.VADSpeechStart:
        fmt.Println("Speech started")
    case voice.VADSpeechEnd:
        fmt.Println("Speech ended")
    case voice.VADSilence:
        // Ongoing silence
    }
}

Direct Construction

For compile-time type safety, construct the provider directly:

import "github.com/lookatitude/beluga-ai/voice/vad/providers/webrtc"

vad := webrtc.New(1500.0, 0.08)

The constructor takes energyThreshold and zcrThreshold as positional arguments. Zero values use the defaults.

Error Handling

result, err := vad.DetectActivity(ctx, audioPCM)
if err != nil {
    log.Printf("VAD detection failed: %v", err)
}

Audio data must be 16-bit little-endian PCM. Frames shorter than 4 bytes return a silence result without error.