voice

README

Installation

go get github.com/chenqinghe/baidu-ai-go-sdk/voice

Usage

• 创建VoiceClient

    	var apiKey,apiSecret string 
	apiKey = "XXXXXXX"
	apiSecret = "XXXXXXX"
	client := voice.NewVoiceClient(apiKey,apiSecret)

• 语音合成

	bts,err:=client.TextToSpeech("你好")
	if err!=nil{
		log.Fatal(err)
	}
	if err:=writeBytesToFile(bts,filename);err!=nil{//writeBytesToFile需要自己实现
		log.Fatal(err)
	}

• 语音识别

	var ap ASRParams = ASRParams{
		Format:  "wav",
		Rate:    16000,
		Channel: 1,
		Cuid:    "mac address",
		Token:   client.AccessToken,
		Lan:     "zh",
		Speech:  afterBase64Str,
		Len:     fiLen,
	}
	strs,err:=client.SpeechToText(ap)
	if err!=nil{
		log.Fatal(err)
	}	
	fmt.Println(strs)//[]string，翻译候选结果，5个。

Documentation

Overview

语音处理 利用百度RESTFul API 进行语音及文字的相互转换

Index

• Constants
• Variables
• type ASRParam
  • func Channel(c int) ASRParam
  • func Format(fmt string) ASRParam
  • func Language(lang string) ASRParam
  • func Rate(rate int) ASRParam
• type ASRParams
• type ASRResponse
• type TTSParam
  • func Person(per int) TTSParam
  • func Pitch(pit int) TTSParam
  • func Speed(spd int) TTSParam
  • func Volume(vol int) TTSParam
• type TTSParams
• type VoiceClient
  • func NewVoiceClient(apiKey, apiSecret string) *VoiceClient
  • func (vc *VoiceClient) SpeechToText(reader io.Reader, params ...ASRParam) ([]string, error)
  • func (vc *VoiceClient) TextToSpeech(txt string, params ...TTSParam) ([]byte, error)

Constants

const (
	B = 1 << (10 * iota)
	KB
	MB
	GB
	TB
	PB
)

const ASR_URL = "http://vop.baidu.com/server_api"

const TTS_URL = "http://tsn.baidu.com/text2audio"

Variables

var (
	ErrTextTooLong = errors.New("The input string is too long")
)

Types

type ASRParam

type ASRParam func(params *ASRParams)

func Channel

func Channel(c int) ASRParam

func Format

func Format(fmt string) ASRParam

func Language

func Language(lang string) ASRParam

func Rate

func Rate(rate int) ASRParam

type ASRParams

type ASRParams struct {
	Format   string `json:"format"`  //语音的格式，pcm 或者 wav 或者 amr。不区分大小写
	Rate     int    `json:"rate"`    //采样率，支持 8000 或者 16000
	Channel  int    `json:"channel"` //声道数，仅支持单声道，请填写固定值 1
	Cuid     string `json:"cuid"`    //用户唯一标识，用来区分用户，计算UV值。建议填写能区分用户的机器 MAC 地址或 IMEI 码，长度为60字符以内
	Token    string `json:"token"`   //开放平台获取到的开发者access_token
	Language string `json:"lan"`     //语种选择，默认中文（zh）。 中文=zh、粤语=ct、英文=en，不区分大小写
	Speech   string `json:"speech"`  //真实的语音数据 ，需要进行base64 编码。与len参数连一起使用
	Length   int    `json:"len"`     //原始语音长度，单位字节
}

语音识别参数

type ASRResponse

type ASRResponse struct {
	CorpusNo string   `json:"corpus_no"`
	ERRMSG   string   `json:"err_msg"`
	ERRNO    int      `json:"err_no"`
	Result   []string `json:"result"`
	SN       string   `json:"sn"`
}

语音识别响应信息

type TTSParam

type TTSParam func(params *TTSParams)

func Person

func Person(per int) TTSParam

func Pitch

func Pitch(pit int) TTSParam

func Speed

func Speed(spd int) TTSParam

func Volume

func Volume(vol int) TTSParam

type TTSParams

type TTSParams struct {
	Text       string `json:"tex"`
	Token      string `json:"tok"`
	Cuid       string `json:"cuid"`
	ClientType int    `json:"ctp"`
	Language   string `json:"lan"`
	Speed      int    `json:"spd"`
	Pitch      int    `json:"pit"`
	Volume     int    `json:"vol"`
	Person     int    `json:"per"`
}

type VoiceClient

type VoiceClient struct {
	*gosdk.Client
}

func NewVoiceClient

func NewVoiceClient(apiKey, apiSecret string) *VoiceClient

func (*VoiceClient) SpeechToText

func (vc *VoiceClient) SpeechToText(reader io.Reader, params ...ASRParam) ([]string, error)

//SpeechToText 语音识别，将语音翻译成文字

func (*VoiceClient) TextToSpeech

func (vc *VoiceClient) TextToSpeech(txt string, params ...TTSParam) ([]byte, error)

TextToSpeech 语音合成，将文字转换为语音