Speech-and-Text
语音转文字(支持实时麦克风输入和从音频文件读入):
- 百度 API
- 科大讯飞 API
- SpeechRecognition (CMU PocketSphinx)
文字转语音:
- pyttsx3
Environment
- Python 3.6.7
- MacOS(以下环境配置方式均基于Mac系统,其他系统的配置方式可能会有一些不同)
Speech to Text
百度
在 https://cloud.baidu.com/product/speech 申请API。
文档:http://ai.baidu.com/docs#/ASR-API
Configuration
安装:
pip install baidu-aip
在 speech_to_text_baidu()
中填入APPID、API_KEY、SECRET_KEY:
APP_ID = ""
API_KEY = ""
SECRET_KEY = ""
(也可以直接使用REST API:Demo
Usage
from Speech_and_Text import speech_to_text_baidu
# 从文件读入
speech_to_text_baidu(audio_path = "path_of_audio", if_microphone = False)
# 从麦克风读入
speech_to_text_baidu(if_microphone = True)
科大讯飞
在 https://www.xfyun.cn/services/voicedictation 申请API。
文档:https://doc.xfyun.cn/rest_api/index.html
Configuration
在 speech_to_text_ifly()
填入 APPID、API_KEY:
URL = "http://api.xfyun.cn/v1/service/v1/iat"
APPID = ""
API_KEY = ""
要在讯飞管理面板中添加调用方api,否则会报错。
Usage
from Speech_and_Text import speech_to_text_ifly
# 从文件读入
speech_to_text_ifly(audio_path = "path_of_audio", if_microphone = False)
# 从麦克风读入
speech_to_text_ifly(if_microphone = True)
SpeechRecognition
使用了Python的语音识别库 SpeechRecognition
源码:https://github.com/Uberi/speech_recognition
Configuration
SpeechRecognition
安装:
pip install SpeechRecognition
PyAudio
使用麦克风进行输入
主页:http://people.csail.mit.edu/hubert/pyaudio/
# Mac上的安装方式
xcode-select --install # 安装xcode, 已经装好的的话,执行的时候会提示
# 先用homebrew安装portaudio(pyaudio需要的库),否则会提示:'portaudio.h' file not found
brew remove portaudio # 先用homebrew卸载
brew install portaudio # 重新安装
sudo pip install pyaudio # 安装pyaudio
Reference: https://stackoverflow.com/questions/33851379/pyaudio-installation-on-mac-python-3
PocketSphinx
CMU Sphinx 是卡内基梅隆大学开发的开源语音识别引擎,可以离线工作,支持多种语言(包括中文)。
源码:https://github.com/cmusphinx
PocketSphinx 是 CMU Sphinx 的 Python 封装接口。
源码:https://github.com/cmusphinx/pocketsphinx-python
安装:
pip install PocketSphinx
添加中文语言包:
查看 SpeechRecognition
包的安装路径('/path'
):
python -c "import speech_recognition as sr, os.path as p; print(p.dirname(sr.__file__))"
然后下载并解压 Mandarin Chinese 语言包,把 zh-CN
文件夹放入 '/path/pocketsphinx-data'
中
Usage
from Speech_and_Text import speech_to_text_cmu
# 从文件读入
speech_to_text_cmu(audio_path = "path_of_audio", if_microphone = False)
# 从麦克风读入
speech_to_text_cmu(if_microphone = True)
Text to Speech
使用了Python的文字转语音库 pyttsx3
源码:https://github.com/nateshmbhat/pyttsx3
文档:https://pyttsx3.readthedocs.io
Configuration
pip install pyttsx3
pip install pyobjc # 依赖模块
Usage
from Speech_and_Text import text_to_speech
# Example
text_to_speech(sentence = "人类的本质是复读机")