# coding=utf-8 import pyaudio import os import sys import requests import base64 import pathlib import threading import time import dashscope from dashscope.audio.qwen_tts_realtime import QwenTtsRealtime, QwenTtsRealtimeCallback, AudioFormat # 添加项目根目录到 sys.path 以便导入 Config root_dir = os.path.dirname(os.path.dirname(os.path.abspath(__file__))) if root_dir not in sys.path: sys.path.append(root_dir) try: from Config.Config import ALY_LLM_API_KEY except ImportError: ALY_LLM_API_KEY = os.getenv("DASHSCOPE_API_KEY") # ======= 常量配置 ======= DEFAULT_TARGET_MODEL = "qwen3-tts-vc-realtime-2026-01-15" DEFAULT_PREFERRED_NAME = "guanyu" DEFAULT_AUDIO_MIME_TYPE = "audio/mpeg" class MyCallback(QwenTtsRealtimeCallback): """ 自定义 TTS 流式回调，增加了播放缓冲区以解决语音断续问题 """ def __init__(self, buffer_seconds=0.5): self.complete_event = threading.Event() self.playback_start_event = threading.Event() # 新增：播放开始事件 self._player = pyaudio.PyAudio() self._stream = self._player.open( format=pyaudio.paInt16, channels=1, rate=24000, output=True, frames_per_buffer=1024 ) self.audio_queue = [] # 缓冲区 self.buffer_threshold = int(24000 * 2 * buffer_seconds) # 字节数 (16bit=2bytes) self.is_playing = False self.playback_thread = None self.all_data_received = False def on_open(self) -> None: print('[TTS] 连接已建立') def on_close(self, close_status_code, close_msg) -> None: self.all_data_received = True if self.playback_thread: self.playback_thread.join() if self._stream: self._stream.stop_stream() self._stream.close() if self._player: self._player.terminate() print(f'[TTS] 连接关闭 code={close_status_code}, msg={close_msg}') def _playback_worker(self): """播放线程，从缓冲区读取并播放""" print("[TTS] 缓冲区达到阈值，开始播放...") self.playback_start_event.set() # 触发播放开始事件 chunk_count = 0 while not self.all_data_received or self.audio_queue: if self.audio_queue: data = self.audio_queue.pop(0) self._stream.write(data) chunk_count += 1 else: if self.all_data_received: break time.sleep(0.01) # 等待数据 print(f"[TTS] 播放线程结束，共播放 {chunk_count} 个数据块") def on_event(self, response: dict) -> None: try: event_type = response.get('type', '') if event_type == 'session.created': print(f'[TTS] 会话开始: {response["session"]["id"]}') elif event_type == 'response.audio.delta': audio_data = base64.b64decode(response['delta']) self.audio_queue.append(audio_data) # 计算当前已缓冲的总字节数 current_buffer_size = sum(len(d) for d in self.audio_queue) # 如果还没开始播放且达到阈值，启动播放线程 if not self.is_playing and current_buffer_size >= self.buffer_threshold: self.is_playing = True self.playback_thread = threading.Thread(target=self._playback_worker) self.playback_thread.start() elif event_type == 'response.done': total_size = sum(len(d) for d in self.audio_queue) print(f'[TTS] 响应完成，总接收音频大小: {total_size} 字节') elif event_type == 'session.finished': print('[TTS] 会话结束') self.all_data_received = True self.complete_event.set() # [Fix] 如果直到会话结束都还没达到缓冲阈值（短文本情况），强制启动播放 if not self.is_playing: print("[TTS] 会话已结束但未达缓冲阈值，强制启动播放...") self.is_playing = True self.playback_thread = threading.Thread(target=self._playback_worker) self.playback_thread.start() except Exception as e: print(f'[Error] 处理回调事件异常: {e}') def wait_for_finished(self): self.complete_event.wait() # 确保播放线程也执行完毕 if self.playback_thread and self.playback_thread.is_alive(): self.playback_thread.join() class QwenTTSManager: """ 通义千问实时语音合成管理类 """ def __init__(self, api_key=ALY_LLM_API_KEY, model=DEFAULT_TARGET_MODEL): self.api_key = api_key self.model = model dashscope.api_key = self.api_key self.callback = None self.qwen_tts_realtime = None def create_voice_enrollment(self, file_path, preferred_name=DEFAULT_PREFERRED_NAME): """ 创建声音复刻音色 """ file_path_obj = pathlib.Path(file_path) if not file_path_obj.exists(): raise FileNotFoundError(f"音频文件不存在: {file_path}") base64_str = base64.b64encode(file_path_obj.read_bytes()).decode() data_uri = f"data:{DEFAULT_AUDIO_MIME_TYPE};base64,{base64_str}" url = "https://dashscope.aliyuncs.com/api/v1/services/audio/tts/customization" payload = { "model": "qwen-voice-enrollment", "input": { "action": "create", "target_model": self.model, "preferred_name": preferred_name, "audio": {"data": data_uri} } } headers = { "Authorization": f"Bearer {self.api_key}", "Content-Type": "application/json" } resp = requests.post(url, json=payload, headers=headers) if resp.status_code != 200: raise RuntimeError(f"创建 voice 失败: {resp.status_code}, {resp.text}") return resp.json()["output"]["voice"] def start_synthesis(self, voice_id, text_list, wait_finished=True, buffer_seconds=0.5): """ 开始实时语音合成并播放 """ self.callback = MyCallback(buffer_seconds=buffer_seconds) self.qwen_tts_realtime = QwenTtsRealtime( model=self.model, callback=self.callback, url='wss://dashscope.aliyuncs.com/api-ws/v1/realtime' ) self.qwen_tts_realtime.connect() self.qwen_tts_realtime.update_session( voice=voice_id, response_format=AudioFormat.PCM_24000HZ_MONO_16BIT, mode='server_commit' ) for text in text_list: print(f'[发送文本]: {text}') self.qwen_tts_realtime.append_text(text) time.sleep(0.1) self.qwen_tts_realtime.finish() if wait_finished: self.callback.wait_for_finished() return self.callback def wait_for_playback_start(self): """等待播放实际开始""" if self.callback: self.callback.playback_start_event.wait()