> ## Documentation Index > Fetch the complete documentation index at: https://docs.senseaudio.cn/llms.txt > Use this file to discover all available pages before exploring further. # 语音合成 WebSocket (TTS) > 基于 WebSocket 的实时 TTS 合成协议参考 ## 说明 基于 WebSocket 的实时文本到语音合成接口，支持增量式文本输入与流式音频输出，适用于实时对话、在线客服、边生成边播报等低延迟场景。单次连接支持多次合成任务，最大文本长度 10000 字符。 * **接入地址**：`wss://api.senseaudio.cn/ws/v1/t2a_v2` * **鉴权方式**：Bearer Token，格式 `Authorization: Bearer SENSEAUDIO_API_KEY` * **消息格式**：`application/json`，音频仅返回 hex 编码字符串 * **计费单位**：按合成字符数计费，详见 [计费说明](/guides/account/billing) * **音色入参**：`voice_id` 必须为当前账号可用音色，参考 [音色列表](/guides/voice/catalog) * **标准 HTTP 合成**：参考 [语音合成 HTTP](/api-reference/endpoint/tts/synthesize) * **SSE 合成**：参考 [语音合成 HTTP 流式](/api-reference/endpoint/tts/synthesize-stream) - WebSocket 接口只支持返回 **hex 格式** 的音频数据。 - 当最后一次收到服务端返回结果后超过 **120 秒** 没有发送新事件时，WebSocket 连接自动断开。 - 音色模型名称：**senseaudio-tts-1.5-260319**。 ## 请求头 (Request Headers) | 参数名 | 必填 | 说明 | 示例 | | :---------------- | :- | :-------------------------------------- | :------------------ | | **Authorization** | 是 | 鉴权 Token。格式：Bearer SENSEAUDIO\_API\_KEY | Bearer sk-123456... | | **Content-Type** | 是 | 内容类型。固定为 application/json | application/json | *** ## 通信流程 ``` 1. 客户端建立 WebSocket 连接 ↓ 2. 服务端返回 connected_success 事件 ↓ 3. 客户端发送 task_start 事件（包含音色、音频格式等配置） ↓ 4. 服务端返回 task_started 事件 ↓ 5. 客户端发送 task_continue 事件（发送待合成文本） ↓ 6. 服务端返回音频数据（可多次发送 task_continue） ↓ 7. 客户端发送 task_finish 事件 ↓ 8. 服务端返回 task_finished 事件并关闭连接 ``` ```mermaid theme={null} sequenceDiagram participant Client as 客户端 participant Server as 服务端 Note over Client, Server: 1. 建立连接 Client->>Server: WebSocket Connect (带 Auth Header) Server-->>Client: JSON: {"event": "connected_success", ...} Note over Client, Server: 2. 开启任务 Client->>Server: JSON: {"event": "task_start", "model": "...", "voice_setting": {...}} Server-->>Client: JSON: {"event": "task_started", ...} Note over Client, Server: 3. 文本合成 (循环) loop 分段合成 Client->>Server: JSON: {"event": "task_continue", "text": "..."} Server-->>Client: JSON: {"event": "task_continued", "data": {"audio": "..."}} end Note over Client, Server: 4. 结束任务 Client->>Server: JSON: {"event": "task_finish"} Server-->>Client: JSON: {"event": "task_finished", ...} Server-->>Client: Close Connection ``` *** ## 客户端事件 ### 1. task\_start - 开始任务 发送此事件正式开始合成任务。当服务端返回 **task\_started** 事件时，标志着任务已成功开始。只有在接收到该事件后，才能向服务器发送 **task\_continue** 或 **task\_finish** 事件。 **请求参数** | 参数名 | 类型 | 必填 | 默认值 | 说明 | | ------------------------------- | ------- | -- | ------ | ----------------------------------------------------- | | **event** | string | 是 | 无 | 固定值：**task\_start** | | **model** | string | 是 | 无 | 模型名称，示例值：**senseaudio-tts-1.5-260319** | | **voice\_setting** | object | 是 | 无 | 声音设置 | | **voice\_setting.voice\_id** | string | 是 | 无 | 主音色名称 | | **voice\_setting.speed** | float | 否 | 1.0 | 语速，取值范围 \[0.5, 2.0] | | **voice\_setting.vol** | float | 否 | 1.0 | 音量，取值范围 \[0.01, 10.0] | | **voice\_setting.pitch** | int | 否 | 0 | 音调，取值范围 \[-12, 12]，0 表示保持原始音调 | | **voice\_setting.latex\_read** | boolean | 否 | false | 数学公式朗读，支持 LaTeX、MathML、Unicode 数学符号等格式。（会产生额外的性能损耗） | | **audio\_setting** | object | 否 | 无 | 音频格式设置 | | **audio\_setting.sample\_rate** | int | 否 | 32000 | 音频采样率，取值范围 \[8000, 16000, 22050, 24000, 32000, 44100] | | **audio\_setting.bitrate** | int | 否 | 128000 | 音频码率，取值范围 \[32000, 64000, 128000, 256000] | | **audio\_setting.format** | string | 否 | mp3 | 输出格式：mp3、wav、pcm、flac | | **audio\_setting.channel** | int | 否 | 2 | 音频声道，1：单声道，2：双声道 | **请求示例** ```json theme={null} { "event": "task_start", "model": "senseaudio-tts-1.5-260319", "voice_setting": { "voice_id": "male_0004_a", "speed": 1, "vol": 1, "pitch": 0, "latex_read": false }, "audio_setting": { "sample_rate": 32000, "bitrate": 128000, "format": "mp3", "channel": 1 } } ``` **响应参数** | 参数名 | 类型 | 说明 | | --------------------------- | ------- | ------- | | **session\_id** | string | 会话 ID | | **event** | string | 事件类型 | | **trace\_id** | string | 链路追踪 ID | | **base\_resp** | object | 请求状态信息 | | **base\_resp.status\_code** | integer | 状态码 | | **base\_resp.status\_msg** | string | 状态详情 | **响应示例** ```json theme={null} { "session_id": "69c20e38c8761996a85d57881fe4d817", "event": "task_started", "trace_id": "69c20e38c8761996a85d57881fe4d817", "base_resp": { "status_code": 0, "status_msg": "success" } } ``` ### 2. task\_continue - 任务继续 当收到服务端返回的 **task\_started** 事件后，任务正式开始，可通过发送 **task\_continue** 事件发送要合成的文本。支持顺序发送多个 **task\_continue** 事件，实现分段文本合成。 **请求参数** | 参数名 | 类型 | 必填 | 说明 | 示例 | | -------------- | ------ | -- | -------------------------------------------- | ------------------------------------------------- | | **event** | string | 是 | 固定值：**task\_continue** | | | **text** | string | 是 | 合成文本内容，支持中英文；支持 `` 停顿符，详见下方停顿符说明 | | | **dictionary** | array | 否 | 多音字配置列表（**模型必须为 senseaudio-tts-1.5-260319**） | `[{"original": "好干净","replacement": "[hao4]干净"}]` | #### LaTeX 公式说明 控制是否朗读 latex 公式，默认为 false。 * 请求中的公式需要在公式的首尾加上 `$$`。 * 请求中公式若有 `\`，需转义成 `\\`。 示例：动能公式 动能公式 表示为：`$$E_k = \\frac{1}{2} m v^2$$` ```json theme={null} { "event": "task_continue", "text": "$$E_k = \\frac{1}{2} m v^2$$" } ``` #### `` 停顿符说明 **``** 用于在语音合成中插入停顿。 ```xml theme={null} ``` * **time** 单位为毫秒（ms） * **500** 表示停顿 500 毫秒 * **最小值为 100 毫秒，最大值无限制** 示例： ```json theme={null} { "event": "task_continue", "text": "你好，这是来自 SenseAudio 的第一条语音" } ``` #### dictionary (多音字纠正) | 参数名 | 类型 | 必填 | 描述 | 默认值 | 示例 | | :-------------- | :----- | :- | :----- | :-- | :-------- | | **original** | string | 是 | 原始文本。 | 无 | 好干净 | | **replacement** | string | 是 | 多音字配置。 | 无 | \[hao4]干净 | **请求示例** ```json theme={null} { "event": "task_continue", "text": "好干净", "dictionary": [ { "original": "好干净", "replacement": "[hao4]干净" } ] } ``` **响应参数** | 参数名 | 类型 | 说明 | | ----------------------------------- | ------- | ------------------------------------------- | | **session\_id** | string | 会话 ID | | **event** | string | 事件类型 | | **trace\_id** | string | 链路追踪 ID | | **is\_final** | boolean | 当前 `task_continue` 语音合成任务是否结束 | | **base\_resp** | object | 请求状态信息 | | **base\_resp.status\_code** | integer | 状态码 | | **base\_resp.status\_msg** | string | 状态详情 | | **data** | object | 返回的合成数据对象（可能为 null） | | **data.audio** | string | 合成后的音频数据（hex 编码） | | **data.status** | integer | 音频流状态：1 表示合成中，2 表示合成结束 | | **extra\_info** | object | 音频附加信息（流式返回时只有最后一个 chunk 会返回） | | **extra\_info.audio\_length** | integer | 音频时长（毫秒） | | **extra\_info.audio\_sample\_rate** | integer | 音频采样率 | | **extra\_info.audio\_size** | integer | 音频文件大小（字节） | | **extra\_info.bitrate** | integer | 音频比特率 | | **extra\_info.audio\_format** | string | 音频格式，取值范围 \[mp3, pcm, flac, wav] | | **extra\_info.audio\_channel** | integer | 音频声道数，1：单声道，2：双声道 | | **extra\_info.word\_count** | integer | 字数统计（按字素簇 grapheme cluster 统计，排除纯空白/标点/控制符） | | **extra\_info.usage\_characters** | integer | 字符数统计（按 Unicode 码点统计） | **响应示例** ```json theme={null} { "session_id": "69c20e38c8761996a85d57881fe4d817", "event": "task_continued", "trace_id": "69c20e38c8761996a85d57881fe4d817", "is_final": false, "data": { "audio": "hex编码的音频数据...", "status": 1 }, "base_resp": { "status_code": 0, "status_msg": "success" } } ``` ### 3. task\_finish - 结束任务 服务端收到此事件后，会等待当前队列中所有合成任务完成后，关闭 WebSocket 连接并结束任务。 **请求参数** | 参数名 | 类型 | 必填 | 说明 | | --------- | ------ | -- | -------------------- | | **event** | string | 是 | 固定值：**task\_finish** | **请求示例** ```json theme={null} { "event": "task_finish" } ``` **响应示例** ```json theme={null} { "session_id": "69c20e38c8761996a85d57881fe4d817", "event": "task_finished", "trace_id": "69c20e38c8761996a85d57881fe4d817", "base_resp": { "status_code": 0, "status_msg": "success" } } ``` *** ## 服务端事件 ### connected\_success - 连接建立成功 初次请求接口时，表示 WebSocket 连接建立成功。 ```json theme={null} { "session_id": "xxxx", "event": "connected_success", "trace_id": "xxx", "base_resp": { "status_code": 0, "status_msg": "success" } } ``` ### task\_started - 任务已开始 标志任务已成功开始，客户端可以开始发送 **task\_continue** 事件。 ```json theme={null} { "session_id": "xxxx", "event": "task_started", "trace_id": "xxxxx", "base_resp": { "status_code": 0, "status_msg": "success" } } ``` ### task\_finished - 任务已结束 标志任务已结束，WebSocket 连接即将关闭。 ```json theme={null} { "session_id": "xxxx", "event": "task_finished", "trace_id": "xxxx", "base_resp": { "status_code": 0, "status_msg": "success" } } ``` ### task\_failed - 任务失败 标志任务失败，`base_resp.status_msg` 中包含错误信息。 ```json theme={null} { "session_id": "xxxx", "event": "task_failed", "trace_id": "xxxxx", "base_resp": { "status_code": 1004, "status_msg": "具体错误信息" } } ``` *** ## 使用示例 依赖：**npm install ws** ```javascript theme={null} const fs = require('fs') const WebSocket = require('ws') const WS_URL = 'wss://api.senseaudio.cn/ws/v1/t2a_v2' const API_KEY = process.env.SENSEAUDIO_API_KEY if (!API_KEY) throw new Error('Missing env: SENSEAUDIO_API_KEY') const output = fs.createWriteStream('output.mp3') const ws = new WebSocket(WS_URL, { headers: { Authorization: `Bearer ${API_KEY}`, 'Content-Type': 'application/json', }, }) ws.on('open', () => console.log('WebSocket 连接已建立')) ws.on('message', (data) => { const response = JSON.parse(data.toString()) console.log('收到服务端事件:', response.event) if (response.event === 'connected_success') { ws.send(JSON.stringify({ event: 'task_start', model: 'senseaudio-tts-1.5-260319', voice_setting: { voice_id: 'male_0004_a', speed: 1.0, vol: 1.0, pitch: 0, latex_read: false }, audio_setting: { sample_rate: 32000, bitrate: 128000, format: 'mp3', channel: 1 }, })) return } if (response.event === 'task_started') { ws.send(JSON.stringify({ event: 'task_continue', text: '道可道，非常道。名可名，非常名。无名天地之始，有名万物之母。' })) ws.send(JSON.stringify({ event: 'task_continue', text: '故常无欲，以观其妙；常有欲，以观其徼。' })) ws.send(JSON.stringify({ event: 'task_finish' })) return } if (response?.data?.audio) { output.write(Buffer.from(response.data.audio, 'hex')) } if (response.event === 'task_finished') { output.end(); ws.close() } if (response.event === 'task_failed') { console.error('任务失败:', response?.base_resp?.status_msg) output.end(); ws.close() } }) ws.on('error', (err) => console.error('WebSocket 错误:', err)) ws.on('close', () => console.log('WebSocket 连接已关闭')) ``` 依赖：**pip install websocket-client** ```python theme={null} import json import os import websocket WS_URL = "wss://api.senseaudio.cn/ws/v1/t2a_v2" API_KEY = os.getenv("SENSEAUDIO_API_KEY") if not API_KEY: raise RuntimeError("Missing env: SENSEAUDIO_API_KEY") output_file = open("output.mp3", "wb") def on_open(ws): print("WebSocket 连接已建立") def on_message(ws, message: str): resp = json.loads(message) event = resp.get("event") print("收到服务端事件:", event) if event == "connected_success": ws.send(json.dumps({ "event": "task_start", "model": "senseaudio-tts-1.5-260319", "voice_setting": {"voice_id": "male_0004_a", "speed": 1.0, "vol": 1.0, "pitch": 0, "latex_read": False}, "audio_setting": {"sample_rate": 32000, "bitrate": 128000, "format": "mp3", "channel": 1} })) return if event == "task_started": ws.send(json.dumps({"event": "task_continue", "text": "道可道，非常道。名可名，非常名。无名天地之始，有名万物之母。"})) ws.send(json.dumps({"event": "task_continue", "text": "故常无欲，以观其妙；常有欲，以观其徼。"})) ws.send(json.dumps({"event": "task_finish"})) return data = resp.get("data") or {} audio_hex = data.get("audio") if audio_hex: output_file.write(bytes.fromhex(audio_hex)) if event in ("task_finished", "task_failed"): if event == "task_failed": print("任务失败:", (resp.get("base_resp") or {}).get("status_msg")) output_file.close(); ws.close() def on_error(ws, err): print("WebSocket 错误:", err) def on_close(ws, status_code, msg): print("WebSocket 连接已关闭:", status_code, msg) ws = websocket.WebSocketApp( WS_URL, header=[f"Authorization: Bearer {API_KEY}", "Content-Type: application/json"], on_open=on_open, on_message=on_message, on_error=on_error, on_close=on_close, ) ws.run_forever() ``` 依赖：**go get github.com/gorilla/websocket** ```go theme={null} package main import ( "encoding/hex" "encoding/json" "log" "net/http" "os" "github.com/gorilla/websocket" ) const wsURL = "wss://api.senseaudio.cn/ws/v1/t2a_v2" func mustMarshal(v any) []byte { b, err := json.Marshal(v); if err != nil { panic(err) }; return b } func main() { apiKey := os.Getenv("SENSEAUDIO_API_KEY") if apiKey == "" { log.Fatal("Missing env: SENSEAUDIO_API_KEY") } header := http.Header{} header.Set("Authorization", "Bearer "+apiKey) header.Set("Content-Type", "application/json") c, _, err := websocket.DefaultDialer.Dial(wsURL, header) if err != nil { log.Fatal("dial:", err) } defer c.Close() out, err := os.Create("output.mp3") if err != nil { log.Fatal(err) } defer out.Close() for { _, msg, err := c.ReadMessage() if err != nil { log.Fatal("read:", err) } var resp map[string]any if err := json.Unmarshal(msg, &resp); err != nil { log.Fatal(err) } event, _ := resp["event"].(string) log.Println("收到服务端事件:", event) switch event { case "connected_success": _ = c.WriteMessage(websocket.TextMessage, mustMarshal(map[string]any{ "event": "task_start", "model": "senseaudio-tts-1.5-260319", "voice_setting": map[string]any{"voice_id": "male_0004_a", "speed": 1.0, "vol": 1.0, "pitch": 0, "latex_read": false}, "audio_setting": map[string]any{"sample_rate": 32000, "bitrate": 128000, "format": "mp3", "channel": 1}, })) case "task_started": _ = c.WriteMessage(websocket.TextMessage, mustMarshal(map[string]any{"event": "task_continue", "text": "道可道，非常道。名可名，非常名。无名天地之始，有名万物之母。"})) _ = c.WriteMessage(websocket.TextMessage, mustMarshal(map[string]any{"event": "task_continue", "text": "故常无欲，以观其妙；常有欲，以观其徼。"})) _ = c.WriteMessage(websocket.TextMessage, mustMarshal(map[string]any{"event": "task_finish"})) case "task_failed", "task_finished": if event == "task_failed" { if baseResp, ok := resp["base_resp"].(map[string]any); ok { if msg, ok := baseResp["status_msg"].(string); ok { log.Println("任务失败:", msg) } } } return } if data, ok := resp["data"].(map[string]any); ok { if audioHex, ok := data["audio"].(string); ok && audioHex != "" { b, err := hex.DecodeString(audioHex) if err != nil { log.Fatal(err) } if _, err := out.Write(b); err != nil { log.Fatal(err) } } } } } ``` 依赖：OkHttp + Jackson ```java theme={null} import com.fasterxml.jackson.databind.JsonNode; import com.fasterxml.jackson.databind.ObjectMapper; import okhttp3.*; import okio.ByteString; import java.io.FileOutputStream; import java.util.HashMap; import java.util.Map; public class SenseAudioTtsWsExample { private static final String WS_URL = "wss://api.senseaudio.cn/ws/v1/t2a_v2"; public static void main(String[] args) throws Exception { String apiKey = System.getenv("SENSEAUDIO_API_KEY"); if (apiKey == null || apiKey.isEmpty()) throw new RuntimeException("Missing env: SENSEAUDIO_API_KEY"); OkHttpClient client = new OkHttpClient(); ObjectMapper mapper = new ObjectMapper(); Request request = new Request.Builder() .url(WS_URL) .addHeader("Authorization", "Bearer " + apiKey) .addHeader("Content-Type", "application/json") .build(); FileOutputStream out = new FileOutputStream("output.mp3"); client.newWebSocket(request, new WebSocketListener() { @Override public void onOpen(WebSocket webSocket, Response response) { System.out.println("WebSocket 连接已建立"); } @Override public void onMessage(WebSocket webSocket, String text) { try { JsonNode resp = mapper.readTree(text); String event = resp.path("event").asText(); System.out.println("收到服务端事件: " + event); if ("connected_success".equals(event)) { Map payload = new HashMap<>(); payload.put("event", "task_start"); payload.put("model", "senseaudio-tts-1.5-260319"); Map voice = new HashMap<>(); voice.put("voice_id", "male_0004_a"); voice.put("speed", 1.0); voice.put("vol", 1.0); voice.put("pitch", 0); voice.put("latex_read", false); payload.put("voice_setting", voice); Map audio = new HashMap<>(); audio.put("sample_rate", 32000); audio.put("bitrate", 128000); audio.put("format", "mp3"); audio.put("channel", 1); payload.put("audio_setting", audio); webSocket.send(mapper.writeValueAsString(payload)); return; } if ("task_started".equals(event)) { webSocket.send(mapper.writeValueAsString(Map.of("event", "task_continue", "text", "道可道，非常道。名可名，非常名。无名天地之始，有名万物之母。"))); webSocket.send(mapper.writeValueAsString(Map.of("event", "task_continue", "text", "故常无欲，以观其妙；常有欲，以观其徼。"))); webSocket.send(mapper.writeValueAsString(Map.of("event", "task_finish"))); return; } JsonNode audioHex = resp.path("data").path("audio"); if (!audioHex.isMissingNode() && !audioHex.isNull()) { byte[] bytes = ByteString.decodeHex(audioHex.asText()).toByteArray(); out.write(bytes); } if ("task_finished".equals(event) || "task_failed".equals(event)) { if ("task_failed".equals(event)) System.err.println("任务失败: " + resp.path("base_resp").path("status_msg").asText()); out.close(); webSocket.close(1000, "done"); } } catch (Exception e) { e.printStackTrace(); webSocket.close(1001, "error"); } } @Override public void onFailure(WebSocket webSocket, Throwable t, Response response) { t.printStackTrace(); try { out.close(); } catch (Exception ignored) {} } }); Thread.sleep(60_000); client.dispatcher().executorService().shutdown(); } } ``` *** ## 技术规格 * **模型名称**：**senseaudio-tts-1.5-260319** * **最大文本长度**：10000 字符 * **连接超时**：最后一次收到服务端返回后 120 秒无新事件时自动断开 ### 音频参数范围 | 参数 | 取值范围 | 默认值 | 说明 | | ------- | -------------------------------------------- | ------ | --------- | | **语速** | \[0.5, 2.0] | 1.0 | 数值越大语速越快 | | **音量** | \[0.01, 10.0] | 1.0 | 数值越大音量越大 | | **音调** | \[-12, 12] | 0 | 正值提高，负值降低 | | **采样率** | 8000, 16000, 22050, 24000, 32000, 44100 (Hz) | 32000 | 推荐 32000 | | **码率** | 32000, 64000, 128000, 256000 (bps) | 128000 | 仅 MP3 格式 | | **声道** | 1 (单声道), 2 (双声道) | 2 | - | ### 支持的音频格式 * **MP3**：推荐，压缩率高，兼容性好。 * **WAV**：无损音质，文件较大。 * **PCM**：原始音频数据。 * **FLAC**：无损压缩。 *** ## 错误码说明 | 状态码 | 说明 | 解决方案 | | ---- | ---- | ------------- | | 0 | 成功 | - | | 1001 | 参数错误 | 检查请求参数格式和取值范围 | *** ## 注意事项 1. **音频数据格式**：WebSocket 接口只支持返回 hex 编码的音频数据，需要在客户端将 hex 字符串转换为二进制数据；音频格式由 `audio_setting.format` 决定。 2. **连接超时机制**：最后一次收到服务端返回后 120 秒内无新事件发送时连接自动断开；建议在任务完成后主动发送 `task_finish`，长时间无操作时可发送心跳保持连接。 3. **事件发送顺序**：必须按照 `task_start → task_continue → task_finish` 顺序发送；只有在收到 `task_started` 后才能发送 `task_continue`；可以发送多个 `task_continue` 事件。 ## 相关资源 标准 HTTP 合成接口参数详解。 流式 HTTP 合成接口参数详解。 TTS 核心特性与音色生态。 系统音色清单与 `voice_id` 规则。 ## AsyncAPI ````yaml api-reference/endpoint/tts/websocket.asyncapi.json tts id: tts title: 语音合成 WebSocket description: 实时 TTS 合成通道，客户端与服务端通过 JSON 消息交换事件。 servers: [] address: /ws/v1/t2a_v2 parameters: [] bindings: [] operations: - &ref_3 id: sendTaskStart title: 开始合成任务 description: 开始合成任务 type: receive messages: - &ref_7 id: taskStart contentType: application/json payload: - name: task_start description: 开始合成任务 type: object properties: - name: event type: string description: 固定值：task_start required: true - name: model type: string description: 模型名称 required: true - name: voice_setting type: object description: 声音设置 required: true properties: - name: voice_id type: string description: 主音色名称 required: true - name: speed type: number description: 语速 required: false - name: vol type: number description: 音量 required: false - name: pitch type: integer description: 音调，0 表示原始音调 required: false - name: latex_read type: boolean description: 数学公式朗读（会产生额外性能损耗） required: false - name: audio_setting type: object description: 音频格式设置 required: false properties: - name: sample_rate type: integer description: 采样率 enumValues: - 8000 - 16000 - 22050 - 24000 - 32000 - 44100 required: false - name: bitrate type: integer description: 码率 enumValues: - 32000 - 64000 - 128000 - 256000 required: false - name: format type: string description: 输出格式 enumValues: - mp3 - wav - pcm - flac required: false - name: channel type: integer description: 声道数 enumValues: - 1 - 2 required: false headers: [] jsonPayloadSchema: type: object required: - event - model - voice_setting properties: event: type: string const: task_start description: 固定值：task_start x-parser-schema-id: model: type: string description: 模型名称 example: senseaudio-tts-1.5-260319 x-parser-schema-id: voice_setting: type: object description: 声音设置 required: - voice_id properties: voice_id: type: string description: 主音色名称 example: male_0004_a x-parser-schema-id: speed: type: number format: float minimum: 0.5 maximum: 2 default: 1 description: 语速 x-parser-schema-id: vol: type: number format: float minimum: 0.01 maximum: 10 default: 1 description: 音量 x-parser-schema-id: pitch: type: integer minimum: -12 maximum: 12 default: 0 description: 音调，0 表示原始音调 x-parser-schema-id: latex_read: type: boolean default: false description: 数学公式朗读（会产生额外性能损耗） x-parser-schema-id: x-parser-schema-id: audio_setting: type: object description: 音频格式设置 properties: sample_rate: type: integer enum: - 8000 - 16000 - 22050 - 24000 - 32000 - 44100 default: 32000 description: 采样率 x-parser-schema-id: bitrate: type: integer enum: - 32000 - 64000 - 128000 - 256000 default: 128000 description: 码率 x-parser-schema-id: format: type: string enum: - mp3 - wav - pcm - flac default: mp3 description: 输出格式 x-parser-schema-id: channel: type: integer enum: - 1 - 2 default: 2 description: 声道数 x-parser-schema-id: x-parser-schema-id: x-parser-schema-id: TaskStartPayload title: task_start description: 开始合成任务 example: |- { "event": "", "model": "", "voice_setting": { "voice_id": "", "speed": 123, "vol": 123, "pitch": 123, "latex_read": true }, "audio_setting": { "sample_rate": 123, "bitrate": 123, "format": "", "channel": 123 } } bindings: [] extensions: - id: x-parser-unique-object-id value: taskStart bindings: [] extensions: &ref_0 - id: x-parser-unique-object-id value: tts - &ref_4 id: sendTaskContinue title: 发送待合成文本 description: 发送待合成文本 type: receive messages: - &ref_8 id: taskContinue contentType: application/json payload: - name: task_continue description: 发送待合成文本 type: object properties: - name: event type: string description: 固定值：task_continue required: true - name: text type: string description: 合成文本内容，支持 停顿符 required: true - name: dictionary type: array description: 多音字配置列表 required: false properties: - name: original type: string description: 原始文本 required: true - name: replacement type: string description: 多音字配置 required: true headers: [] jsonPayloadSchema: type: object required: - event - text properties: event: type: string const: task_continue description: 固定值：task_continue x-parser-schema-id: text: type: string description: 合成文本内容，支持 停顿符 x-parser-schema-id: dictionary: type: array description: 多音字配置列表 items: type: object required: - original - replacement properties: original: type: string description: 原始文本 x-parser-schema-id: replacement: type: string description: 多音字配置 x-parser-schema-id: x-parser-schema-id: x-parser-schema-id: x-parser-schema-id: TaskContinuePayload title: task_continue description: 发送待合成文本 example: |- { "event": "", "text": "", "dictionary": { "original": "", "replacement": "" } } bindings: [] extensions: - id: x-parser-unique-object-id value: taskContinue bindings: [] extensions: *ref_0 - &ref_5 id: sendTaskFinish title: 结束合成任务 description: 结束合成任务 type: receive messages: - &ref_9 id: taskFinish contentType: application/json payload: - name: task_finish description: 结束合成任务 type: object properties: - name: event type: string description: 固定值：task_finish required: true headers: [] jsonPayloadSchema: type: object required: - event properties: event: type: string const: task_finish description: 固定值：task_finish x-parser-schema-id: x-parser-schema-id: TaskFinishPayload title: task_finish description: 结束合成任务 example: |- { "event": "" } bindings: [] extensions: - id: x-parser-unique-object-id value: taskFinish bindings: [] extensions: *ref_0 - &ref_6 id: receiveServerEvents title: 接收服务端事件 description: 接收服务端事件 type: send messages: - &ref_10 id: connectedSuccess contentType: application/json payload: - name: connected_success description: 连接建立成功 type: object properties: - name: session_id type: string description: 会话 ID required: false - name: event type: string description: 事件类型 required: false - name: trace_id type: string description: 链路追踪 ID required: false - name: base_resp type: object required: false properties: - name: status_code type: integer description: 状态码，0 表示成功 required: false - name: status_msg type: string description: 状态详情 required: false headers: [] jsonPayloadSchema: &ref_1 type: object properties: session_id: type: string description: 会话 ID x-parser-schema-id: event: type: string description: 事件类型 x-parser-schema-id: trace_id: type: string description: 链路追踪 ID x-parser-schema-id: base_resp: type: object properties: status_code: type: integer format: int64 description: 状态码，0 表示成功 x-parser-schema-id: status_msg: type: string description: 状态详情 x-parser-schema-id: x-parser-schema-id: BaseResp x-parser-schema-id: BaseServerEvent title: connected_success description: 连接建立成功 example: |- { "session_id": "", "event": "", "trace_id": "", "base_resp": { "status_code": 123, "status_msg": "" } } bindings: [] extensions: - id: x-parser-unique-object-id value: connectedSuccess - &ref_11 id: taskStarted contentType: application/json payload: - name: task_started description: 任务已开始 type: object properties: - name: session_id type: string description: 会话 ID required: false - name: event type: string description: 事件类型 required: false - name: trace_id type: string description: 链路追踪 ID required: false - name: base_resp type: object required: false properties: - name: status_code type: integer description: 状态码，0 表示成功 required: false - name: status_msg type: string description: 状态详情 required: false headers: [] jsonPayloadSchema: *ref_1 title: task_started description: 任务已开始 example: |- { "session_id": "", "event": "", "trace_id": "", "base_resp": { "status_code": 123, "status_msg": "" } } bindings: [] extensions: - id: x-parser-unique-object-id value: taskStarted - &ref_12 id: taskContinued contentType: application/json payload: - allOf: &ref_2 - *ref_1 - type: object properties: is_final: type: boolean description: 当前 task_continue 语音合成任务是否结束 x-parser-schema-id: data: type: object nullable: true description: 返回的合成数据对象 properties: audio: type: string description: 合成后的音频数据（hex 编码） x-parser-schema-id: status: type: integer format: int64 description: 音频流状态：1 合成中，2 合成结束 x-parser-schema-id: x-parser-schema-id: extra_info: type: object description: 音频附加信息，流式返回时仅最后一个 chunk 携带 properties: audio_length: type: integer format: int64 description: 音频时长（毫秒） x-parser-schema-id: audio_sample_rate: type: integer format: int64 description: 音频采样率 x-parser-schema-id: audio_size: type: integer format: int64 description: 音频文件大小（字节） x-parser-schema-id: bitrate: type: integer format: int64 description: 音频比特率 x-parser-schema-id: audio_format: type: string enum: - mp3 - pcm - flac - wav description: 音频格式 x-parser-schema-id: audio_channel: type: integer description: 音频声道数 x-parser-schema-id: word_count: type: integer format: int64 description: 字数统计（按 grapheme cluster 统计） x-parser-schema-id: usage_characters: type: integer format: int64 description: 字符数统计（按 Unicode 码点） x-parser-schema-id: x-parser-schema-id: x-parser-schema-id: x-parser-schema-id: TaskContinuedPayload name: task_continued description: 音频分片返回 headers: [] jsonPayloadSchema: allOf: *ref_2 x-parser-schema-id: TaskContinuedPayload title: task_continued description: 音频分片返回 example: '{}' bindings: [] extensions: - id: x-parser-unique-object-id value: taskContinued - &ref_13 id: taskFinished contentType: application/json payload: - name: task_finished description: 任务已结束 type: object properties: - name: session_id type: string description: 会话 ID required: false - name: event type: string description: 事件类型 required: false - name: trace_id type: string description: 链路追踪 ID required: false - name: base_resp type: object required: false properties: - name: status_code type: integer description: 状态码，0 表示成功 required: false - name: status_msg type: string description: 状态详情 required: false headers: [] jsonPayloadSchema: *ref_1 title: task_finished description: 任务已结束 example: |- { "session_id": "", "event": "", "trace_id": "", "base_resp": { "status_code": 123, "status_msg": "" } } bindings: [] extensions: - id: x-parser-unique-object-id value: taskFinished - &ref_14 id: taskFailed contentType: application/json payload: - name: task_failed description: 任务失败 type: object properties: - name: session_id type: string description: 会话 ID required: false - name: event type: string description: 事件类型 required: false - name: trace_id type: string description: 链路追踪 ID required: false - name: base_resp type: object required: false properties: - name: status_code type: integer description: 状态码，0 表示成功 required: false - name: status_msg type: string description: 状态详情 required: false headers: [] jsonPayloadSchema: *ref_1 title: task_failed description: 任务失败 example: |- { "session_id": "", "event": "", "trace_id": "", "base_resp": { "status_code": 123, "status_msg": "" } } bindings: [] extensions: - id: x-parser-unique-object-id value: taskFailed bindings: [] extensions: *ref_0 sendOperations: - *ref_3 - *ref_4 - *ref_5 receiveOperations: - *ref_6 sendMessages: - *ref_7 - *ref_8 - *ref_9 receiveMessages: - *ref_10 - *ref_11 - *ref_12 - *ref_13 - *ref_14 extensions: - id: x-parser-unique-object-id value: tts securitySchemes: - id: bearerAuth name: bearerAuth type: http description: 'Bearer Token 鉴权，格式 `Authorization: Bearer `' scheme: bearer extensions: [] ````