speech

Documentation

Index

• func SetLogLevel(level LogLevel)
• func SetLogrusLogLevel(level LogLevel)
• type Detector
  • func NewDetector(cfg DetectorConfig) (*Detector, error)
  • func (sd *Detector) ChangeLogLevel(logLevel LogLevel)
  • func (sd *Detector) Destroy() error
  • func (sd *Detector) Detect(pcm []float32) ([]Segment, error)
  • func (sd *Detector) Reset() error
• type DetectorConfig
  • func (c DetectorConfig) IsValid() error
• type LogLevel
• type Segment

Functions

func SetLogLevel

func SetLogLevel(level LogLevel)

func SetLogrusLogLevel

func SetLogrusLogLevel(level LogLevel)

Types

type Detector

type Detector struct {
	// contains filtered or unexported fields
}

func NewDetector

func NewDetector(cfg DetectorConfig) (*Detector, error)

func (*Detector) ChangeLogLevel

func (sd *Detector) ChangeLogLevel(logLevel LogLevel)

func (*Detector) Destroy

func (sd *Detector) Destroy() error

func (*Detector) Detect

func (sd *Detector) Detect(pcm []float32) ([]Segment, error)

func (*Detector) Reset

func (sd *Detector) Reset() error

type DetectorConfig

type DetectorConfig struct {
	// The path to the ONNX Silero VAD model file to load.
	ModelPath string
	// The sampling rate of the input audio samples. Supported values are 8000 and 16000.
	SampleRate int
	// The number of samples to process at each infer.
	WindowSize int
	// The probability threshold above which we detect speech. A good default is 0.5.
	Threshold float32
	// The duration of silence to wait for each speech segment before separating it.
	MinSilenceDurationMs int
	// The padding to add to speech segments to avoid aggressive cutting.
	SpeechPadMs int
}

func (DetectorConfig) IsValid

func (c DetectorConfig) IsValid() error

type LogLevel

type LogLevel int

const (
	LogLevelFatal LogLevel = iota
	LogLevelError
	LogLevelWarning
	LogLevelInfo
	LogLevelVerbose
)

type Segment

type Segment struct {
	// The relative timestamp in seconds of when a speech segment begins.
	SpeechStartAt float64
	// The relative timestamp in seconds of when a speech segment ends.
	SpeechEndAt float64
}

Segment contains timing information of a speech segment.