接入必读

请先查看接入必读了解具体接入方式，再参考此文档完成接入。

接口说明

• 当前支持通过 WebSocket 协议在线调用
• 请求内容包括：payload字段为将请求参数序列化后的json文本

公共参数

参考详细说明功能调用-通用协议-WebSocket。

配置参数

payload配置参数为json字符串格式

示例：

{
    "text": "欢迎使用文本转语音服务。",
    "speaker": "zh_female_qingxin",
    "audio_config":
    {
        "format": "wav",
        "sample_rate": 16000
    }
}

响应格式

响应中不同消息类型的接收：

• Text message文本消息类型，包含控制事件和响应数据
• Binary message二进制消息类型，接收 只包含二进制数据 的响应。当enable_timestamp=false时，服务端返回二进制消息类型。

文本消息类型响应的定义如下：

• 响应结果payload为json字符串格式，json内容格式如下：

  字段	描述	类型
  duration	音频时长，单位秒	number
  words	字的时间戳，单位秒。需要请求参数audio_config.enable_timestamp=true	array
  words.word	字内容	string
  words.start_time	当前字开始时间	number
  words.end_time	当前字结束时间	number
  phonemes	音素的时间戳，单位秒。需要请求参数audio_config.enable_timestamp=true	array
  phonemes.phone	音素内容	string
  phonemes.start_time	当前音素开始时间	number
  phonemes.end_time	当前音素结束时间	number

  payload示例：

  {
      "duration": 3.0,
      "words":
      [
          {
              "word": "你",
              "start_time": "0",
              "end_time": "0.05"
          },
          ...
      ],
      "phonemes":
      [
          {
              "phone": "C0n",
              "start_time": "0",
              "end_time": "0.025"
          },
          ...
      ]
  }
  

参考示例

流式调用方式参考公共WebSocket流式协议

Golang

// Code sample:
// use websocket client to invoke SAMI Streaming Service

package main

import (
	"bytes"
	"encoding/json"
	"flag"
	"fmt"
	"io/ioutil"
	"log"
	"net/url"
	"os"
	"os/signal"
	"time"

	"github.com/gorilla/websocket"
)

type WebSocketRequest struct {
	Token     string  `header:"SAMI-Token,required" json:"token,required" query:"token,required"`
	Appkey    string  `json:"appkey,required" query:"appkey,required" vd:"$!=''"`
	Namespace string  `json:"namespace,required" query:"namespace,required" vd:"$!=''"`
	Version   string  `json:"version,omitempty" query:"version"`
	Event     string  `json:"event,omitempty" query:"event"`
	Payload   *string `form:"payload" json:"payload,omitempty"`
	Data      []byte  `form:"data" json:"data,omitempty"`
	TaskId    string  `json:"task_id,omitempty" query:"task_id"`
}
type WebSocketResponse struct {
	TaskId     string  `form:"task_id,required" json:"task_id,required" query:"task_id,required"`
	MessageId  string  `form:"message_id,required" json:"message_id,required" query:"message_id,required"`
	Namespace  string  `form:"namespace,required" json:"namespace,required" query:"namespace,required"`
	Event      string  `form:"event,required" json:"event,required" query:"event,required"`
	StatusCode int32   `form:"status_code,required" json:"status_code,required" query:"status_code,required"`
	StatusText string  `form:"status_text,required" json:"status_text,required" query:"status_text,required"`
	Payload    *string `form:"payload,omitempty" json:"payload,omitempty" query:"payload,omitempty"`
	Data       []byte  `form:"data,omitempty" json:"data,omitempty" query:"data,omitempty"`
}

const (
	EventStartTask    = "StartTask"
	EventTaskStarted  = "TaskStarted"
	EventFinishTask   = "FinishTask"
	EventTaskFinished = "TaskFinished"
)

var (
	// websocket domain
	addr = flag.String("addr", "sami.bytedance.com", "http service address")
	// user auth
	token  = "your_token"
	appkey = "your_appkey"

	u url.URL
	c *websocket.Conn

	interrupt chan os.Signal
	done      chan struct{}
	err       error

	inputFile  = "input.pcm"
	outputFile = "output.wav"
)

func main() {
	flag.Parse()
	log.SetFlags(0)

	interrupt = make(chan os.Signal, 1)
	signal.Notify(interrupt, os.Interrupt)
	done = make(chan struct{})

	u = url.URL{Scheme: "wss", Host: *addr, Path: "/api/v1/ws"}
	log.Printf("connecting to %s", u.String())

	c, _, err = websocket.DefaultDialer.Dial(u.String(), nil)
	if err != nil {
		log.Fatal("dial:", err)
	}
	defer c.Close()

	streamingTTSTest("zh_female_qingxin", "Hello, 欢迎使用文本转语音服务", "wav", 16000, true)
	// streamingTTSTest("zh_female_qingxin", "Hello, 欢迎使用文本转语音服务", "wav", 16000, false)
}

func readTaskStartedEvent() error {
	msgType, message, err := c.ReadMessage()
	if err != nil {
		log.Println("read TaskStarted event failed, ", err)
		return err
	}
	if msgType != websocket.TextMessage {
		log.Println("read TaskStarted event failed, message type not TextMessage: ", msgType)
		return fmt.Errorf("MessageTypeNotMatch")
	}
	taskStartedEvent := &WebSocketResponse{}
	err = json.Unmarshal(message, taskStartedEvent)
	if err != nil {
		return err
	}
	if taskStartedEvent.Event != EventTaskStarted {
		log.Printf("read TaskStarted event failed, event type not match: %+v", *taskStartedEvent)
		return fmt.Errorf("EventTypeNotMatch")
	}
	log.Printf("read TaskStarted event: %+v", *taskStartedEvent)
	return nil
}

func streamingTTSTest(speaker, text, format string, sampleRate int, enableTimestamp bool) {
	var buf bytes.Buffer
	defer func() {
		log.Println("save bytes into file:", buf.Len())
		if buf.Len() > 0 {
			_ = ioutil.WriteFile(outputFile, buf.Bytes(), 0644)
		}
	}()

	// send control message
	payloadStr := fmt.Sprintf(
		`{"audio_config":{"format":"%v","speech_rate":0, "sample_rate":%v, "enable_timestamp":%t},"speaker":"%v","text":"%v"}`,
		format, sampleRate, enableTimestamp, speaker, text,
	)
	controlReq := &WebSocketRequest{
		Token:     token,
		TaskId:    "test_mock",
		Appkey:    appkey,
		Namespace: "TTS",
		Event:     EventStartTask,
		Payload:   &payloadStr,
	}
	controlMsg, _ := json.Marshal(controlReq)
	err = c.WriteMessage(websocket.TextMessage, controlMsg)
	if err != nil {
		log.Println("write:", err)
		return
	}
	if err = readTaskStartedEvent(); err != nil {
		log.Println("read failed, ", err)
		return
	}
	controlReq.Event = EventFinishTask
	controlMsg, _ = json.Marshal(controlReq)
	err = c.WriteMessage(websocket.TextMessage, controlMsg)
	if err != nil {
		log.Println("write:", err)
		return
	}

	go func() {
		defer close(done)
		isFirst := true
		startTime := time.Now()
		for {
			mt, message, err := c.ReadMessage()
			if err != nil {
				log.Println("read:", err)
				return
			}
			if isFirst {
				startTime = time.Now()
				isFirst = false
			}
			if mt == websocket.BinaryMessage {
				log.Printf("recv: byte[%v]", len(message))
				buf.Write(message)
			} else {
				wsResp := WebSocketResponse{}
				ttsPayload := ""
				err := json.Unmarshal(message, &wsResp)
				if err != nil {
					log.Printf("recv text message, parse failed")
				}
				if wsResp.Event == EventTaskFinished {
					log.Printf(
						"recv TaskFinished event: %+v, cost_time=%v", wsResp, time.Since(startTime).Milliseconds(),
					)
					return
				}
				if wsResp.Payload != nil {
					ttsPayload = *wsResp.Payload
				}
				buf.Write(wsResp.Data)
				log.Printf("recv: data=byte[%d], payload=%v", len(wsResp.Data), ttsPayload)
			}
		}
	}()

	for {
		select {
		case <-done:
			return
		case <-interrupt:
			log.Println("interrupt")

			// Cleanly close the connection by sending a close message and then
			// waiting (with timeout) for the server to close the connection.
			err := c.WriteMessage(
				websocket.CloseMessage, websocket.FormatCloseMessage(websocket.CloseNormalClosure, ""),
			)
			if err != nil {
				log.Println("write close:", err)
				return
			}
			select {
			case <-done:
			case <-time.After(5 * time.Second):
			}
			return
		}
	}
}

常见问题

• 通用FAQ

• 通用状态码

• 特定错误码