Python speech_recognition missing PocketSphinx language data directory zh-cn

•        Python : 3.8.13
•          OS : Windows 21H1
•       Conda : 4.12.0
•     PyCharm : 2022.1 (Community Edition)

 

代码及报错

import speech_recognition as sr

audio_file = 'test.wav'
r = sr.Recognizer()

# 打开语音文件
with sr.AudioFile(audio_file) as source:
    audio = r.record(source)

print('文本内容: ', r.recognize_sphinx(audio, language='zh_CN'))  # 汉语

D:\Develop\Anaconda3\envs\pyqtenv\python.exe C:/Users/admin/Desktop/pythonProject/main.py
Traceback (most recent call last):
  File "C:/Users/admin/Desktop/pythonProject/main.py", line 10, in <module>
    print('文本内容: ', r.recognize_sphinx(audio, language='zh_CN'))  # 汉语
  File "D:\Develop\Anaconda3\envs\pyqtenv\lib\site-packages\speech_recognition\__init__.py", line 741, in recognize_sphinx
    raise RequestError("missing PocketSphinx language data directory: \"{}\"".format(language_directory))
speech_recognition.RequestError: missing PocketSphinx language data directory: "D:\Develop\Anaconda3\envs\pyqtenv\Lib\site-packages\speech_recognition\pocketsphinx-data\zh_CN"

Process finished with exit code 1

 

解决方法

• 报错的核心内容

speech_recognition.RequestError: missing PocketSphinx language data directory: "D:\Develop\Anaconda3\envs\pyqtenv\Lib\site-packages\speech_recognition\pocketsphinx-data\zh_CN"

• 检查步骤

1. 核对路径信息，确认该路径存在

"D:\Develop\Anaconda3\envs\pyqtenv\Lib\site-packages\speech_recognition\pocketsphinx-data\zh_CN"

2. 确认language的值与文件夹名称相同

r.recognize_sphinx(audio, language='zh_CN')
"D:\Develop\Anaconda3\envs\pyqtenv\Lib\site-packages\speech_recognition\pocketsphinx-data\zh_CN"

3. 确认zh_CN文件夹中的内容齐全(我用的是大佬提供的中文模型包)

D:\Develop\Anaconda3\envs\pyqtenv\Lib\site-packages\speech_recognition\pocketsphinx-data>tree /F
卷 新加卷 的文件夹 PATH 列表
卷序列号为 944F-CFF3
D:.
├─en-US
│  │  language-model.lm.bin
│  │  LICENSE.txt
│  │  pronounciation-dictionary.dict
│  │
│  └─acoustic-model
│          feat.params
│          mdef
│          means
│          noisedict
│          README
│          sendump
│          transition_matrices
│          variances
│
└─zh_CN
    │  language-model.lm.bin
    │  pronounciation-dictionary.dict
    │  zh_broadcastnews_16k_ptm256_8000.tar.bz2
    │  zh_broadcastnews_64000_utf8.DMP
    │
    └─acoustic-model
            feat.params
            mdef
            means
            mixture_weights
            noisedict
            sendump
            transition_matrices
            variances

4. 前三步完成后，我的程序就可以正常运行了。

 

学习资源

• PocketSpinx添加中文语音识别包
• python使用Sphinx实现中文普通话语音识别
• WINDOW下使用CMUSPHINX实现中文识别

 

学习推荐

• Python文档 - English
• Python文档 - 中文
• Python规范 PEP
• Python规范 google版
• Python 源码
• Python PEP
• 掘金平台
• gitee平台

 

---

Python具有开源、跨平台、解释型、交互式等特性，值得学习。
Python的设计哲学：优雅，明确，简单。提倡用一种方法，最好是只有一种方法来做一件事。
代码的书写要遵守规范，这样有助于沟通和理解。
每种语言都有独特的思想，初学者需要转变思维、踏实践行、坚持积累。