speech

README

speech

Simple wrapper over text-to-speech service, previously Bing, currently AWS Polly.

Documentation

Overview

Package speech provides an interface, Speaker, that allows an application to access the basic function of a text-to-speech service, hiding the underlying provider, allowing migration (eg, from Bing to Polly). Call the provider-specific New function (eg, polly.NewPollyV1, polly.NewPollyV2) and assign the result to a Speaker variable.

Index

• type Credentials
• type Speaker
• type Spoken

Types

type Credentials

type Credentials struct {

	// ClientID is whatever is conventional as the ID finally presented to the service.
	ClientID string

	// Keys contains one or more service access keys.
	Keys []string
}

Credentials represents a common form of service credentials: a user ID string and then one or more keys to identify and authorise the service.

type Speaker

type Speaker interface {
	// Speak says something at a given speaking rate ("medium or "slow") within a locale, using a specified voice.
	Speak(text, speakingRate, locale, voice string) (*Spoken, error)
}

Speaker represents a session on a text-to-speech engine. The language, voice and speaking rate can be different on each call.

type Spoken

type Spoken struct {

	// Audio produces the stream of audio, sadly unseekable.
	Audio io.ReadCloser

	// TextLen is the length of the spoken text's recording format in bytes.
	TextLen int

	// Format names the recording format ("mp3", "ogg_vorbis", "pcm", "json").
	Format string

	// ContentType is the HTTP "Content-Type" header's value for the recording format.
	ContentType string
}

Spoken represents a stream of bytes in Audio, in the given Format.