Lucky-183
微软 Azure ASR 和 TTS 问题
我注册了 Azure 语音服务，注册完成后给我了两个密钥，我应该用哪个？还有节点地址发现和代码中调用的地址不一样，是我注册错误呢还是这块需要更改代码重新编写？
微软Azure 位置区域：eastasia 微软Azure 节点：https://eastasia.api.cognitive.microsoft.com/
作者大大，我注册的SpeechServices区域和节点是 koreacentral https://koreacentral.api.cognitive.microsoft.com/
然后测试的时候，deepseek api返回了文字，但是无法播放语音。log显示
2025-03-14 14:16:00 | INFO | chat.py:421 - The command is '你好'
2025-03-14 14:16:03 | INFO | chat.py:137 - start new conversation
2025-03-14 14:16:06 | INFO | chat.py:322 -
嗯，你好呀！😊 其实最近我在忙着准备考试呢！你最近怎么样？有什么想聊的吗？
2025-03-14 14:16:06 | INFO | chat.py:358 - A conversation end
2025-03-14 14:16:08 | INFO | chat.py:137 - start new conversation
2025-03-14 14:16:09 | INFO | chat.py:191 - prepare to start record
2025-03-14 14:16:10 | INFO | speechpoint.py:30 - 正在获取声音中...
2025-03-14 14:16:15 | WARNING | chat.py:200 - listening timed out while waiting for phrase to start
请问该如何解决呢？谢谢
节点地址不需要更改，按照代码，填入密钥应该就可以运行
@swansun95 将 tts.py azure_reco.py tts_stream.py 三个文件中的 url 地址由 eastasia 替换为 koreacentral , 试一下
koreacentral
@Lucky-183 报错了
2025-03-14 15:00:02 | INFO |
你好呀！有什么我可以帮你的吗？最近你好像特别开心呢！有什么特别值得分享的吗？哈哈～
Exception ignored on calling ctypes callback function: <function PushAudioOutputStream.__write_callback at 0x7f96e51760>
Traceback (most recent call last):
File "/home/swansun/pi-ass/lib/python3.11/site-packages/azure/cognitiveservices/speech/audio.py", line 309, in __write_callback
return obj.write(view)
^^^^^^^^^^^^^^^
File "/home/swansun/PI-Assistant/tts_stream.py", line 56, in write
self.stream.write(buffer.tobytes())
File "/home/swansun/pi-ass/lib/python3.11/site-packages/pyaudio.py", line 585, in write
pa.write_stream(self._stream, frames, num_frames,
SystemError: PY_SSIZE_T_CLEAN macro must be defined for '#' formats
Exception ignored on calling ctypes callback function: <function PushAudioOutputStream.__write_callback at 0x7f96e51760>
Traceback (most recent call last):
File "/home/swansun/pi-ass/lib/python3.11/site-packages/azure/cognitiveservices/speech/audio.py", line 309, in __write_callback
return obj.write(view)
^^^^^^^^^^^^^^^
File "/home/swansun/PI-Assistant/tts_stream.py", line 56, in write
self.stream.write(buffer.tobytes())
File "/home/swansun/pi-ass/lib/python3.11/site-packages/pyaudio.py", line 585, in write
pa.write_stream(self._stream, frames, num_frames,
SystemError: PY_SSIZE_T_CLEAN macro must be defined for '#' formats
Exception ignored on calling ctypes callback function: <function PushAudioOutputStream.__write_callback at 0x7f96e51760>
Traceback (most recent call last):
File "/home/swansun/pi-ass/lib/python3.11/site-packages/azure/cognitiveservices/speech/audio.py", line 309, in __write_callback
return obj.write(view)
^^^^^^^^^^^^^^^
File "/home/swansun/PI-Assistant/tts_stream.py", line 56, in write
self.stream.write(buffer.tobytes())
File "/home/swansun/pi-ass/lib/python3.11/site-packages/pyaudio.py", line 585, in write
pa.write_stream(self._stream, frames, num_frames,
SystemError: PY_SSIZE_T_CLEAN macro must be defined for '#' formats
Exception ignored on calling ctypes callback function: <function PushAudioOutputStream.__write_callback at 0x7f96e51760>
Traceback (most recent call last):
File "/home/swansun/pi-ass/lib/python3.11/site-packages/azure/cognitiveservices/speech/audio.py", line 309, in __write_callback
return obj.write(view)
^^^^^^^^^^^^^^^
File "/home/swansun/PI-Assistant/tts_stream.py", line 56, in write
self.stream.write(buffer.tobytes())
@Lucky-183 解决了，根据chatGPT提示，卸载了 PyAudio-0.2.11，安装最新的 pyaudio-0.2.14就可以了。我是 Raspberry Pi OS（64bit）。默认带了python 3.11.2
谢谢