update:优化

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

@@ -11,7 +11,6 @@ from core.handle.reportHandle import enqueue_tts_report
 from abc import ABC, abstractmethod
 from core.utils.tts import MarkdownCleaner
 from core.utils.util import audio_to_data
-from core.utils import opus_encoder_utils
 
 TAG = __name__
 logger = setup_logging()
@@ -33,9 +32,6 @@ class TTSProviderBase(ABC):
         self.output_file = config.get("output_dir")
         self.tts_queue = queue.Queue()
         self.audio_play_queue = queue.Queue()
-        self.opus_encoder = opus_encoder_utils.OpusEncoderUtils(
-            sample_rate=16000, channels=1, frame_size_ms=60
-        )
         # 添加实现类型属性，默认为非流式
         self.interface_type = TTSImplementationType.NON_STREAMING
 
@@ -43,9 +39,14 @@ class TTSProviderBase(ABC):
     def generate_filename(self):
         pass
 
-    def to_tts(self, text):
+    def to_tts(self, text, index):
         """如果是流式实现，一般没有文件生成，我们返回枚举值"""
-        if self.interface_type != TTSImplementationType.SINGLE_STREAMING:
+        if self.interface_type != TTSImplementationType.NON_STREAMING:
+            if index == 1:
+                future = asyncio.run_coroutine_threadsafe(
+                    self.start_session(self.conn.session_id), loop=self.conn.loop
+                )
+                future.result()
             asyncio.run(self.text_to_speak(text, None))
             return self.interface_type.value
 
@@ -221,10 +222,6 @@ class TTSProviderBase(ABC):
                     f"audio_play_priority priority_thread: {text} {e}"
                 )
 
-    def wav_to_opus_data_audio_raw(self, raw_data_var, is_end=False):
-        opus_datas = self.opus_encoder.encode_pcm_to_opus(raw_data_var, is_end)
-        return opus_datas
-
     async def start_session(self, session_id):
         pass
 

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

@@ -4,12 +4,11 @@ import threading
 import traceback
 import uuid
 import json
-from datetime import datetime
-
 import websockets
-
+from datetime import datetime
 from config.logger import setup_logging
 from core.providers.tts.base import TTSProviderBase, TTSImplementationType
+from core.utils.util import pcm_to_data
 
 TAG = __name__
 logger = setup_logging()
@@ -167,7 +166,6 @@ class TTSProvider(TTSProviderBase):
             target=self._start_monitor_tts_response_thread(), daemon=True
         )
         tts_priority.start()
-        await self.start_session(conn.session_id)
 
     def generate_filename(self, extension=".wav"):
         return os.path.join(
@@ -381,7 +379,7 @@ class TTSProvider(TTSProviderBase):
                     and res.header.message_type == AUDIO_ONLY_RESPONSE
                 ):
                     logger.bind(tag=TAG).info(f"推送数据到队列里面～～")
-                    opus_datas = self.wav_to_opus_data_audio_raw(res.payload)
+                    opus_datas = pcm_to_data(res.payload)
                     logger.bind(tag=TAG).info(
                         f"推送数据到队列里面帧数～～{len(opus_datas)}"
                     )
@@ -390,15 +388,15 @@ class TTSProvider(TTSProviderBase):
                     json_data = json.loads(res.payload.decode("utf-8"))
                     self.tts_text = json_data.get("text", "")
                     logger.bind(tag=TAG).info(f"句子开始～～{self.tts_text}")
-                    self.audio_play_queue.put((None, self.tts_text, 0))
+                    self.audio_play_queue.put(([], self.tts_text, 0))
 
                 elif res.optional.event == EVENT_TTSSentenceEnd:
                     logger.bind(tag=TAG).info(f"句子结束～～{self.tts_text}")
-                    self.audio_play_queue.put((None, self.tts_text, 0))
+                    self.audio_play_queue.put(([], self.tts_text, 0))
                 elif res.optional.event == EVENT_SessionFinished:
                     logger.bind(tag=TAG).info(f"会话结束～～,最后一句补零")
-                    # opus_datas = self.wav_to_opus_data_audio_raw(b"", is_end=True)
-                    self.audio_play_queue.put((None, self.tts_text, 0))
+                    opus_datas = pcm_to_data(b"")
+                    self.audio_play_queue.put((opus_datas, self.tts_text, 0))
                 else:
                     continue
             except websockets.ConnectionClosed: