{"componentChunkName":"component---src-templates-acg-portal-new-template-tsx","path":"/smtihp4uo","result":{"data":{"markdownRemark":{"html":"

语音合成

\n

非流式在线合成（HTTP）

\n

接口描述

\n

接口描述：音色创建成功后，通过创建得到的音色ID进行文本的合成

\n

接口名称：https://aip.baidubce.com/rest/2.0/speech/publiccloudspeech/v1/voice/clone/tts

\n

请求方法：POST（Content-Type: application/json）

\n

输入参数

\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n
属性
参数名称类型是否必填说明
access_tokenstringquery二选一鉴权令牌
Authorizationstringheader二选一鉴权API Key
音色IDvoice_idint必填训练后的音色 ID
合成文本textstring必填总字数不超过500个字符，1个中文字、英文字母、数字或符号均算作1个字符
合成语言langstring可选待合成的语种类型，支持以下语种：
中英语：zh;
日语：ja;
若合成语音阶段未填写lang参数，默认为创建音色选择的语种。
合成方言dialectstring可选待合成的方言类型，支持以下方言：
上海话：wuu-CN-shanghai；
河南话：zh-CN-henan；
四川话：zh-CN-sichuan；
湖南话：zh-CN-hunan；
贵州话：zh-CN-guizhou；
dialect参数与emotion参数请勿同时上传。
合成情感emotionstring可选待合成语音的情感，支持以下情感：
高兴：happy;
悲伤：down；
惊讶：surprise;
愤怒：angry;
恐惧：fear；
厌恶：disgust;
不上传emotion参数时（例如text_ctrl={\"emo\": \"\"}），默认按照注册音频效果合成。
dialect参数与emotion参数请勿同时上传。
音调pitchfloat可选音调，取值范围[0, 15]，默认为5
音量volumefloat可选音量，取值范围[0, 15]，默认为5
语速speedfloat可选语速，取值范围[0, 15]，默认为5
输出格式media_typestring可选输出文件格式，支持wav、mp3，默认值为wav
采样率sample_rateint可选采样率，仅支持将采样率降采，支持8000、16000、24000
元数据标识metadatastring可选元数据标识，字段与取值：
ContentPropagator：内容传播服务提供者
PropagateID：内容传播编号
ContentPropagator、PropagateID请传入string类型的数据。
若不传入两个字段，默认不添加元数据标识；传入字段后音频header将添加Label、ContentProducer、ProduceID、ContentPropagator、PropagateID总计5个字段。
\n
\n

输出参数

\n

需要根据 Content-Type的头部来确定是否服务端合成成功。\n如果合成成功，返回的Content-Type以“audio”开头

\n\n

如果合成出现错误，则会返回json文本，具体header信息为：Content-Type: application/json。

\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n
属性参数名称参数类型描述
错误状态statusint0代表成功，其他为异常
错误信息messagestring错误消息
\n
\n

示例

\n\n
\n
\n
\n Json\n \n
\n
\n 
1curl --location --request POST 'https://aip.baidubce.com/rest/2.0/speech/publiccloudspeech/v1/voice/clone/tts?access_token=xxxx' \\\n2{\n3    \"text\": \"当春风拂过，大地渐渐回暖，万物复苏的季节到来了。花儿在这个时候竞相开放，桃花、樱花、杏花，还有那细雨中的紫藤，都以最绚烂的姿态展示自己的美丽。\",\n4    \"voice_id\": 100001,\n5    \"metadata\":{\n6           \"ContentPropagator\":XXXXX,\n7           \"PropagateID\":XXXXX,\n8    }\n9}
\n
\n
\n
\n \n
\n

流式在线合成（Websocket）

\n

接口描述

\n

接口描述：音色创建成功后，通过创建得到的音色ID进行文本的合成

\n

接口名称：wss://aip.baidubce.com/ws/2.0/speech/publiccloudspeech/v1/voice/clone/tts

\n

建立连接

\n

请求路径：/ws/2.0/speech/publiccloudspeech/v1/voice/clone/tts

\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n
参数名
类型参数位置是否必填说明
access_tokenstringquery二选一鉴权令牌
Authorizationstringheader二选一鉴权API Key
voice_idintquery必填音色ID
idle_timeoutintquery可选空闲超时时间，单位秒，默认值60，范围[5, 600]
\n
\n

状态码说明

\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n
状态码状态信息错误码错误消息说明
400Bad Request216100Invalid 'foo' value: bar.参数错误
400Bad Request216100Invalid 'foo' value: bar, range: [min, max].参数错误
400Bad Request216101Missing required parameter: 'foo'.参数缺失
400Bad Request400websocket: the client is not using the websocket protocol: ...WebSocket协议错误
401Unauthorized216100Invalid access_token parameter.access_token参数格式不正确
401Unauthorized216100Invalid Authorization header.Authorization请求头格式不正确
401Unauthorized216100dialect and emotion cannot be set at the same time.dialect和emotion参数无法同时上传
401Unauthorized216100Invalid 'emotion' value.上传的emotion参数为非法
401Unauthorized216101Missing access_token parameter or Authorization header.缺少鉴权参数
401Unauthorized110Access token invalid or no longer validaccess_token 校验不通过。
401Unauthorized111Access token expiredaccess_token 过期，建议更新获取Access token 重新请求。
401Unauthorized217IAM auth errorAPI Key校验不通过
403Forbidden6No permission to access dataappid没有相应接口权限，可以检查console页面app管理中勾选的AI能力情况。
404Not Found输入的url错误
429Too Many Requests15Open api concurrency limit reached触发并发限流
429Too Many Requests17Open api usage limit reached无剩余可用额度
400Bad Request216404Voice id not exists.voice_id不存在
500Internal Server Error282000Internal error.服务器内部错误
400Bad Request282101The voice_id service has not been started.服务未启动
400Bad Request216403Not allowed to use this voice.对该voice_id无权限
\n
\n

初始化

\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n
参数名称类型是否必填说明
typestring必填开始帧的类型，固定值 system.start
payloadobject可选合成参数，详见下表
\n

payload字段说明

\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n
参数名称类型是否必填说明
langstring可选待合成的语种类型，支持以下语种：
中英语：zh;
日语：ja;
若合成语音阶段未填写lang参数，默认为创建音色选择的语种。
dialectstring可选待合成的方言类型，支持以下方言：
上海话：wuu-CN-shanghai；
河南话：zh-CN-henan；
四川话：zh-CN-sichuan；
湖南话：zh-CN-hunan；
贵州话：zh-CN-guizhou；
dialect参数与emotion参数请勿同时上传。
emotionstring可选待合成语音的情感，支持以下情感：
高兴：happy;
悲伤：down；
惊讶：surprise;
愤怒：angry;
恐惧：fear；
厌恶：disgust;
不上传emotion参数时（例如text_ctrl={\"emo\": \"\"}），默认按照注册音频效果合成。
dialect参数与emotion参数请勿同时上传。
pitchfloat可选音调，取值 0-15，默认为 5
volumefloat可选音量，基础音库取值0-9，其他音库取值 0-15，默认为 5
speedfloat可选语速，取值 0-15，默认为 5
media_typestring可选音频格式，支持wav、mp3、pcm，默认值为wav
若希望使用postman测试流式合成效果，建议使用mp3格式
sample_rateint可选采样率，支持8000、16000、24000
初始化过程中，可以通过audio_ctrl参数指定采样率，sample_rate和audio_ctrl不可同时使用
metadatastring可选元数据标识，字段与取值：
ContentPropagator：内容传播服务提供者
PropagateID：内容传播编号
ContentPropagator、PropagateID请传入string类型的数据。
若不传入两个字段，默认不添加元数据标识；传入字段后音频header将添加Label、ContentProducer、ProduceID、ContentPropagator、PropagateID总计5个字段。
spd（兼容字段）int可选语速，取值 0-15，默认为 5
pit（兼容字段）int可选音调，取值 0-15，默认为 5
vol（兼容字段）int可选音量，基础音库取值0-9，其他音库取值 0-15，默认为 5
aue（兼容字段）int可选音频格式，3=mp3-16k/24k，4=pcm-16k/24k，5=pcm-8k，6=wav-16k/24k，和media_type参数不可同时设置
\n

示例代码如下：

\n\n
\n
\n
\n Json\n \n
\n
\n 
1{\n2    \"type\": \"system.start\",\n3    \"payload\": {\n4        \"lang\": \"zh\",\n5        \"media_type\": \"mp3\",\n6        \"pitch\": 5,\n7        \"volume\": 5,\n8        \"speed\": 5,\n9        \"dialect\":\"zh-CN-henan\",\n10         \"metadata\":{\n11            \"ContentPropagator\":XXXXX,\n12            \"PropagateID\":XXXXX,\n13        }\n14    }\n15}
\n
\n
\n
\n \n

正常响应将返回：

\n\n
\n
\n
\n Json\n \n
\n
\n 
1{\n2  \"type\": \"system.started\",\n3  \"code\": 0,\n4  \"message\": \"success\",\n5  \"headers\": {\n6    \"session_id\": \"57944e1a-470f-4c30-9cf6-36e0aa645e58\"\n7  }\n8}
\n
\n
\n
\n \n

返回码说明

\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n
typecodemessage说明
system.started216100Invalid system.start payload.参数错误
system.started216100Invalid 'foo' value: bar.参数错误
system.started216100Invalid 'foo' value: bar, range: [min, max].参数错误
\n
\n

发送文本

\n

示例代码如下：

\n\n
\n
\n
\n Json\n \n
\n
\n 
1{\n2    \"type\": \"text\",\n3    \"payload\": {\n4        \"text\": \"需要进行语音合成的文字\"\n5    }\n6}
\n
\n
\n
\n \n

返回码说明

\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n
typecodemessage说明
system.error216100Invalid text payload.参数错误
system.error216101Missing required parameter: 'text'.缺少text参数
system.error216103Text exceeded 1000 characters limit.单次文本不能超过1000字
system.error216429The current pending text is too long to be processed.发送频率过快，待处理文本过长
\n

备注：

\n
    \n
  1. 客户端发送的文本信息无内容时，返回system.error，code=216101，message=Missing required parameter: 'text'.
    2. \n
\n
\n

结束合成

\n

客户端请求

\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n
参数名称类型是否必填说明
typestring必填结束帧的类型，固定值 system.finish
\n\n
\n
\n
\n Json\n \n
\n
\n 
1{\n2    \"type\":\"system.finish\",\n3}
\n
\n
\n
\n \n

正常响应

\n\n
\n
\n
\n Json\n \n
\n
\n 
1{\n2  \"type\": \"system.finished\",\n3  \"code\": 0,\n4  \"message\": \"success\",\n5  \"headers\": {\n6    \"session_id\": \"58b18cef-c588-4f04-a4a4-4a2977072e58\"\n7  }\n8}
\n
\n
\n
\n \n
\n

服务端主动发送

\n

返回码说明

\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n
typecodemessage说明
system.error216604Open api usage limit reached额度已用完，无剩余可用额度
","fields":{"slug":"smtihp4uo","title":"语音合成","date":"2026-09-01","extractedHeadings":[]},"headings":[{"value":"语音合成","depth":1},{"value":"非流式在线合成（HTTP）","depth":2},{"value":"接口描述","depth":3},{"value":"输入参数","depth":3},{"value":"输出参数","depth":3},{"value":"示例","depth":3},{"value":"流式在线合成（Websocket）","depth":2},{"value":"接口描述","depth":3},{"value":"建立连接","depth":3},{"value":"初始化","depth":3},{"value":"发送文本","depth":3},{"value":"结束合成","depth":3},{"value":"服务端主动发送","depth":3}]}},"pageContext":{"isCreatedByStatefulCreatePages":false,"slug":"smtihp4uo","prev":{"id":"cmtihnk12","name":"音色创建与管理","path":"cmtihnk12","filePath":"模型/AI开放能力/大模型声音复刻/音色创建与管理.md","seo":null,"parentIds":["Hm9b3tx1a","um9b473u2","5mtihkkkt"],"parents":[{"id":"Hm9b3tx1a","documentId":"d53594ee-6426-4489-ab0f-d46881463d5d","name":"模型","repoName":"qianfan-api","filePath":"模型","disabled":false,"path":"Hm9b3tx1a","lastMergeTime":null,"isApiDoc":null,"httpMethod":null,"seo":null,"sourceOrgName":null,"sourceRepoName":null,"sourceDocumentId":null,"solutionTag":null,"solutionColor":null},{"id":"um9b473u2","documentId":"e658cedb-8c8f-48a7-b3a3-4d8a29ad87af","name":"AI开放能力","repoName":"qianfan-api","filePath":"模型/AI开放能力","disabled":false,"path":"um9b473u2","lastMergeTime":null,"isApiDoc":null,"httpMethod":null,"seo":null,"sourceOrgName":null,"sourceRepoName":null,"sourceDocumentId":null,"solutionTag":null,"solutionColor":null},{"id":"5mtihkkkt","documentId":"e0e33348-b829-4100-9923-c16a0f93b07b","name":"大模型声音复刻","repoName":"qianfan-api","filePath":"模型/AI开放能力/大模型声音复刻","disabled":false,"path":"5mtihkkkt","lastMergeTime":null,"isApiDoc":null,"httpMethod":null,"seo":null,"sourceOrgName":null,"sourceRepoName":null,"sourceDocumentId":"","solutionTag":null,"solutionColor":null}]},"next":{"id":"Ymio9i9i4","name":"图像生成","path":"Ymio9i9i4","filePath":"模型/图像生成/蒸汽机 Air-Image.md","seo":null,"parentIds":["Hm9b3tx1a","0mio99s5b"],"parents":[{"id":"Hm9b3tx1a","documentId":"d53594ee-6426-4489-ab0f-d46881463d5d","name":"模型","repoName":"qianfan-api","filePath":"模型","disabled":false,"path":"Hm9b3tx1a","lastMergeTime":null,"isApiDoc":null,"httpMethod":null,"seo":null,"sourceOrgName":null,"sourceRepoName":null,"sourceDocumentId":null,"solutionTag":null,"solutionColor":null},{"id":"0mio99s5b","documentId":"3e4d5ea8-af7b-4edd-af82-a953de19321d","name":"图像生成","repoName":"qianfan-api","filePath":"模型/图像生成","disabled":false,"path":"0mio99s5b","lastMergeTime":null,"isApiDoc":null,"httpMethod":null,"seo":null,"sourceOrgName":null,"sourceRepoName":null,"sourceDocumentId":"","solutionTag":null,"solutionColor":null}]},"parents":[{"id":"Hm9b3tx1a","documentId":"d53594ee-6426-4489-ab0f-d46881463d5d","name":"模型","repoName":"qianfan-api","filePath":"模型","disabled":false,"path":"Hm9b3tx1a","lastMergeTime":null,"isApiDoc":null,"httpMethod":null,"seo":null,"sourceOrgName":null,"sourceRepoName":null,"sourceDocumentId":null,"solutionTag":null,"solutionColor":null},{"id":"um9b473u2","documentId":"e658cedb-8c8f-48a7-b3a3-4d8a29ad87af","name":"AI开放能力","repoName":"qianfan-api","filePath":"模型/AI开放能力","disabled":false,"path":"um9b473u2","lastMergeTime":null,"isApiDoc":null,"httpMethod":null,"seo":null,"sourceOrgName":null,"sourceRepoName":null,"sourceDocumentId":null,"solutionTag":null,"solutionColor":null},{"id":"5mtihkkkt","documentId":"e0e33348-b829-4100-9923-c16a0f93b07b","name":"大模型声音复刻","repoName":"qianfan-api","filePath":"模型/AI开放能力/大模型声音复刻","disabled":false,"path":"5mtihkkkt","lastMergeTime":null,"isApiDoc":null,"httpMethod":null,"seo":null,"sourceOrgName":null,"sourceRepoName":null,"sourceDocumentId":"","solutionTag":null,"solutionColor":null}],"specificSeo":null}}}