Crate pipecrab_vad 

pipecrab-vad: the voice-activity-detection interface, in two tiers.

Voice activity has two natural shapes, and this crate names both:

• VoiceActivityDetector — the stage-facing capability: audio in, speech edges out (VadEvent::SpeechStarted / VadEvent::SpeechStopped). Segmenter-class engines that already speak in segments — sherpa’s VAD, platform VADs — implement this directly. It is what VadStage drives.
• SpeechScorer — the raw-model tier: a per-window speech probability. This is what a bare silero build (native ort, browser onnxruntime-web) exposes. A scorer becomes a detector through the Debounced adapter, which owns the windowing, threshold, and hangover.

Probabilities used to live on the one VAD trait; they moved down to SpeechScorer. A segmenter never has a probability to hand back, and anything downstream that wants confidence (a turn manager, prosody) composes its own SpeechScorer rather than leaning on the VAD to surface one.

The edge contract

Across the lifetime of a VoiceActivityDetector, events alternate, starting with SpeechStarted: started, stopped, started, stopped, … This is a documented invariant that VadStage and everything downstream of it trust.

VadStage turns those edges into a gate: it owns a pre-roll ring and emits speech-only audio, bracketed by the edges. Note the contract this inverts — downstream of the gate, SpeechStarted precedes an utterance’s audio (pre-roll included) and SpeechStopped follows its last chunk. (The older lane-discipline design emitted the edge after the chunk that triggered it; that wording is gone.) See VadStage for the full gate algorithm.

No runtime format detection

Both traits take an Arc<[f32]>, which carries no sample rate. Samples are interpreted as input_format(); no runtime detection is possible. The stage enforces the format fatally before any audio reaches an engine, so an engine never sees nonconforming samples through the pipeline — which is why VadError carries no format-mismatch variant.

Platform-neutral and wasm32-checkable: the concrete engines live elsewhere, each behind these traits, so the interface itself carries no backend dependency and compiles for the host and wasm32-unknown-unknown.

Structs

DebounceConfig

Threshold and hangover for Debounced.

Debounced

Lifts a SpeechScorer into a VoiceActivityDetector by windowing its input, thresholding its probabilities, and debouncing the result into edges. The sibling of pipecrab-stt’s Buffered; see the module docs.

GateConfig

Tuning for VadStage’s pre-roll ring.

VadEffect

One step the gate asks perform to carry out: VadStage’s Processor::Effect. Its contents are private — only the stage constructs one — so the effect vocabulary stays opaque to callers.

VadStage

Adapts any VoiceActivityDetector into a pipeline Stage — as a gate, not a tap.

Enums

VadError

Why a VoiceActivityDetector::process or SpeechScorer::score call failed.

VadEvent

A speech-state transition, emitted by a VoiceActivityDetector.

Traits

SpeechScorer

The raw-model tier: per-window speech probability.

VoiceActivityDetector

The stage-facing voice-activity capability: audio in, speech edges out.