PyAudio引发语音命令程序报错（Errno -9999/-9988）求解决方法

问题背景 #

已安装与Python 3.8.5兼容的PyAudio 0.2.13版本，运行语音识别代码时触发OSError: [Errno -9999] Unanticipated host error，后续伴随OSError: [Errno -9988] Stream closed错误。

原代码 #

import pyttsx3 
import speech_recognition as sr
import webbrowser
import datetime
import pyjokes
import smtplib


def sptext():
    recognizer = sr.Recognizer() 
    with sr.Microphone() as source:
        print("Listening...")
        # Noise cancellation
        recognizer.adjust_for_ambient_noise(source)
        audio= recognizer.listen(source)
        try:
            print("Jack is recognizing...")
            data = recognizer.recognize_google(audio)
            print(data)
        except sr.UnknownValueError:
            print("Jack is uncertain")
            
        
sptext()

错误回溯 #

Listening...
Traceback (most recent call last):
  File "c:\Users\91999\.vscode\Python\tempCodeRunnerFile.py", line 16, in sptext
    audio= recognizer.listen(source)
  File "C:\Users\91999\AppData\Local\Programs\Python\Python38\lib\site-packages\speech_recognition\__init__.py", line 523, in listen
    buffer = source.stream.read(source.CHUNK)
  File "C:\Users\91999\AppData\Local\Programs\Python\Python38\lib\site-packages\speech_recognition\__init__.py", line 199, in read
    return self.pyaudio_stream.read(size, exception_on_overflow=False)
  File "C:\Users\91999\AppData\Local\Programs\Python\Python38\lib\site-packages\pyaudio\__init__.py", line 570, in read
    return pa.read_stream(self._stream, num_frames,
OSError: [Errno -9999] Unanticipated host error

During handling of the above exception, another exception occurred:

Traceback (most recent call last):
  File "c:\Users\91999\.vscode\Python\tempCodeRunnerFile.py", line 27, in <module> 
    sptext()
  File "c:\Users\91999\.vscode\Python\tempCodeRunnerFile.py", line 23, in sptext   
    print(" is uncertain")
  File "C:\Users\91999\AppData\Local\Programs\Python\Python38\lib\site-packages\speech_recognition\__init__.py", line 189, in __exit__
    self.stream.close()
  File "C:\Users\91999\AppData\Local\Programs\Python\Python38\lib\site-packages\speech_recognition\__init__.py", line 204, in close
    if not self.pyaudio_stream.is_stopped():
  File "C:\Users\91999\AppData\Local\Programs\Python\Python38\lib\site-packages\pyaudio\__init__.py", line 515, in is_stopped
    return pa.is_stream_stopped(self._stream)
OSError: [Errno -9988] Stream closed
PS C:\Users\91999\.vscode\Python> 

修复步骤 #

• 检查麦克风权限与设备状态

  1. 关闭所有占用麦克风的程序（如视频会议、语音聊天软件）
  2. 进入Windows设置>隐私和安全性>麦克风，确认允许当前Python环境访问麦克风
  3. 使用系统自带录音工具测试麦克风硬件是否正常工作

• 指定音频设备参数

  1. 运行以下代码获取可用音频设备列表，找到麦克风对应的deviceIndex：

     import pyaudio
     p = pyaudio.PyAudio()
     for i in range(p.get_device_count()):
         print(f"设备索引{i}: {p.get_device_info_by_index(i)['name']}")
     

  2. 修改原代码中sr.Microphone()的初始化，指定设备索引和采样率：

     with sr.Microphone(device_index=你的设备索引, sample_rate=16000) as source:
     

• 优化异常处理逻辑
  扩展异常捕获范围，覆盖PyAudio设备错误、API请求错误等场景，避免流关闭时触发二次错误，修改后的完整代码：

  import pyttsx3 
  import speech_recognition as sr
  import webbrowser
  import datetime
  import pyjokes
  import smtplib
  
  def sptext():
      recognizer = sr.Recognizer() 
      # 替换为你的麦克风设备索引
      with sr.Microphone(device_index=0, sample_rate=16000) as source:
          print("Listening...")
          # 缩短降噪时长，避免设备超时
          recognizer.adjust_for_ambient_noise(source, duration=0.5)
          try:
              # 设置监听超时和短语时长，防止无限等待
              audio = recognizer.listen(source, timeout=5, phrase_time_limit=10)
              print("Jack is recognizing...")
              data = recognizer.recognize_google(audio)
              print(data)
          except sr.UnknownValueError:
              print("Jack is uncertain")
          except sr.RequestError:
              print("Google语音识别API请求失败")
          except OSError as e:
              print(f"音频设备错误: {str(e)}")
          except Exception as e:
              print(f"未知错误: {str(e)}")
            
  sptext()
  

• 重新安装PyAudio（若上述步骤无效）

  1. 卸载现有PyAudio：pip uninstall -y pyaudio
  2. 下载对应Python3.8的PyAudio预编译wheel文件（适配系统架构）
  3. 本地安装wheel文件：pip install PyAudio-0.2.13-cp38-cp38-win_amd64.whl（替换为实际下载的文件名）

内容的提问来源于stack exchange，提问作者Muskaan Sharma