HuoshanDoubleStreamTTS增加自定义配置

main/xiaozhi-server/config.yaml

@@ -715,13 +715,31 @@ TTS:
     speaker: zh_female_wanwanxiaohe_moon_bigtts
     # 开启WebSocket连接复用，默认复用（注意：复用后设备处于聆听状态时空闲链接会占并发数）
     enable_ws_reuse: True
-    speech_rate: 0
-    loudness_rate: 0
-    pitch: 0
-    # 多情感音色参数，注意：当前仅部分音色支持设置情感。
-    # 相关音色列表：https://www.volcengine.com/docs/6561/1257544
-    emotion: "neutral"  # 情感类型，可选值为：neutral、happy、sad、angry、fearful、disgusted、surprised
-    emotion_scale: 4  # 情感强度，可选值为：1~5，默认值为4
+    # 相关参数文档：https://www.volcengine.com/docs/6561/1329505
+    # 音频输出配置（audio_params）- 用户可自定义添加火山引擎支持的任何音频参数
+    audio_params:
+      speech_rate: 0  # 语速(-50~100)
+      loudness_rate: 0  # 音量(-50~100)
+      # 多情感音色参数，注意：当前仅部分音色支持设置情感。
+      # 相关音色列表：https://www.volcengine.com/docs/6561/1257544
+      # emotion: "neutral"  # 情感类型（仅部分音色支持）：neutral、happy、sad、angry、fearful、disgusted、surprised
+      # emotion_scale: 4  # 情感强度(1~5)
+    # 高级文本处理配置（additions）- 用户可自定义添加火山引擎支持的任何高级参数
+    additions:
+      post_process:
+        pitch: 0  # 音高(-12~12)
+      # aigc_metadata: {}  # AIGC元数据配置
+      # cache_config: {}  # 缓存配置
+    # 混音控制配置（mix_speaker）- 多音色混合（仅 TTS 1.0）
+    # 混音功能主要适用于豆包语音合成模型1.0的音色，使用时需要将req_params.speaker设置为custom_mix_bigtts
+    # mix_speaker:
+    #   speakers:
+    #     - source_speaker: zh_male_bvlazysheep
+    #       mix_factor: 0.3
+    #     - source_speaker: BV120_streaming
+    #       mix_factor: 0.3
+    #     - source_speaker: zh_male_ahu_conversation_wvae_bigtts
+    #       mix_factor: 0.4
   CosyVoiceSiliconflow:
     type: siliconflow
     # 硅基流动TTS

main/xiaozhi-server/core/providers/tts/huoshan_double_stream.py

@@ -154,17 +154,30 @@ class TTSProvider(TTSProviderBase):
             self.voice = config.get("private_voice")
         else:
             self.voice = config.get("speaker")
-        speech_rate = config.get("speech_rate", "0")
-        loudness_rate = config.get("loudness_rate", "0")
-        pitch = config.get("pitch", "0")
-        self.speech_rate = int(speech_rate) if speech_rate else 0
-        self.loudness_rate = int(loudness_rate) if loudness_rate else 0
-        self.pitch = int(pitch) if pitch else 0
-        # 多情感音色参数
-        self.emotion = config.get("emotion", "neutral")  
-        emotion_scale = config.get("emotion_scale", "4")
-        self.emotion_scale = int(emotion_scale) if emotion_scale else 4
 
+        # 默认 audio_params 配置
+        default_audio_params = {
+            "speech_rate": 0,
+            "loudness_rate": 0
+        }
+
+        # 默认 additions 配置
+        default_additions = {
+            "aigc_metadata": {},
+            "cache_config": {},
+            "post_process": {
+                "pitch": 0
+            }
+        }
+
+        # 默认 mix_speaker 配置
+        default_mix_speaker = {}
+
+        # 合并用户配置
+        self.audio_params = {**default_audio_params, **config.get("audio_params", {})}
+        self.additions = {**default_additions, **config.get("additions", {})}
+        self.mix_speaker = {**default_mix_speaker, **config.get("mix_speaker", {})}
+        print(self.mix_speaker)
         self.ws_url = config.get("ws_url")
         self.authorization = config.get("authorization")
         self.header = {"Authorization": f"{self.authorization}{self.access_token}"}
@@ -179,6 +192,8 @@ class TTSProvider(TTSProviderBase):
     async def open_audio_channels(self, conn):
         try:
             await super().open_audio_channels(conn)
+            # 更新 audio_params 中的采样率为实际的 conn.sample_rate
+            self.audio_params["sample_rate"] = conn.sample_rate
         except Exception as e:
             logger.bind(tag=TAG).error(f"Failed to open audio channels: {str(e)}")
             self.ws = None
@@ -645,18 +660,17 @@ class TTSProvider(TTSProviderBase):
         speaker="",
         audio_format="pcm",
     ):
-        audio_params = {
-            "format": audio_format,
-            "sample_rate": self.conn.sample_rate,
-            "speech_rate": self.speech_rate,
-            "loudness_rate": self.loudness_rate
+        # 构建 req_params
+        req_params = {
+            "text": text,
+            "speaker": speaker,
+            "audio_params": {**self.audio_params, "format": audio_format},
+            "additions": json.dumps(self.additions)
         }
-
-        # 如果是多情感音色,添加情感参数
-        if '_emo_' in self.voice:
-            if self.emotion:
-                audio_params["emotion"] = self.emotion
-                audio_params["emotion_scale"] = self.emotion_scale
+        
+        # 如果有 mix_speaker 配置，添加到 req_params
+        if self.mix_speaker:
+            req_params["mix_speaker"] = self.mix_speaker
 
         return str.encode(
             json.dumps(
@@ -664,17 +678,7 @@ class TTSProvider(TTSProviderBase):
                     "user": {"uid": uid},
                     "event": event,
                     "namespace": "BidirectionalTTS",
-                    "req_params": {
-                        "text": text,
-                        "speaker": speaker,
-                        "audio_params": audio_params,
-                        "additions": json.dumps({
-                            "post_process": {
-                                "pitch": self.pitch
-                            }
-                        })
-                    },
-
+                    "req_params": req_params
                 }
             )
         )