hear

README

hear

A Go package that uses PortAudio and Google's Speech-to-Text API to provide live recognition of human mouth sounds.

Documentation

Index

• Constants
• func FreeAudio()
• func InitAudio()
• func ListenIntoBuffer(opts ListenOpts) (*bytes.Buffer, error)
• type GCPSpeechConv
  • func NewGCPSpeechConv(accountFile string) (*GCPSpeechConv, error)
  • func (gcp *GCPSpeechConv) Convert(data []byte) (string, error)
• type ListenOpts
• type State
• type VAD
  • func NewVAD(width int) *VAD
  • func (v *VAD) FFT() []complex128
  • func (v *VAD) Flux(samples []int16) float64

Constants

const DefaultQuietTime = time.Second

Functions

func FreeAudio

func FreeAudio()

func InitAudio

func InitAudio()

func ListenIntoBuffer

func ListenIntoBuffer(opts ListenOpts) (*bytes.Buffer, error)

Types

type GCPSpeechConv

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

func NewGCPSpeechConv

func NewGCPSpeechConv(accountFile string) (*GCPSpeechConv, error)

func (*GCPSpeechConv) Convert

func (gcp *GCPSpeechConv) Convert(data []byte) (string, error)

type ListenOpts

type ListenOpts struct {
	State            func(State)
	QuietDuration    time.Duration
	AlreadyListening bool
}

type State

type State int

const (
	Waiting State = iota
	Listening
	Asking
)

type VAD

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

func NewVAD

func NewVAD(width int) *VAD

func (*VAD) FFT

func (v *VAD) FFT() []complex128

func (*VAD) Flux

func (v *VAD) Flux(samples []int16) float64

Given the samples, return the spectral flux value as compared to the previous samples.