├── README.md
├── out.wav
└── videotoword.py


/README.md:
--------------------------------------------------------------------------------
1 | # video-to-word
2 | python+百度语音识别API， 读取视频文件，将视频中的语音转换成文字输出
3 | 


--------------------------------------------------------------------------------
/out.wav:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/yuchengle/video-to-word/693bf2b1ceadadcfb85abbc901b5abca922791dd/out.wav


--------------------------------------------------------------------------------
/videotoword.py:
--------------------------------------------------------------------------------
 1 | #!/usr/bin/env python
 2 | # coding: utf-8
 3 | 
 4 | from moviepy.editor import VideoFileClip
 5 | from aip import AipSpeech
 6 | 
 7 | 
 8 | # # 视频转音频
 9 | 
10 | video_file = 'https://v2.addnewer.com/media/2020/08/1596606940740.mp4'
11 | audio_file = '/data/jupyter/yucl/videotoword/out.wav'
12 | video = VideoFileClip(video_file)
13 | video.audio.write_audiofile(audio_file,ffmpeg_params=['-ar','16000','-ac','1'])
14 | 
15 | 
16 | # # 音频转文本
17 | 
18 | '''
19 | 百度语音识别：https://console.bce.baidu.com/ai/?_=1597039403602#/ai/speech/overview/index
20 | '''
21 | #从百度AI开放平台创建应用处获取
22 | APP_ID = '***'
23 | API_KEY = '***'
24 | SECRET_KEY = '***'
25 | 
26 | client = AipSpeech(APP_ID, API_KEY, SECRET_KEY)
27 | 
28 | # 读取文件
29 | def get_file_content(audio_file):
30 |     with open(audio_file, 'rb') as fp:
31 |         return fp.read()
32 | 
33 | # 识别本地文件
34 | def get_text():
35 |     result = client.asr(get_file_content(audio_file), 'wav', 16000, {'dev_pid': 1537,})
36 |     #print(result)
37 |     text = result['result'][0]
38 |     return text
39 | 
40 | res = get_text()
41 | print (res)
42 | 
43 | '''紫熨斗全新的文体验波色因欧莱雅集团专利成本协同，玻尿酸生锈，石伟有一段话拽文，抬头纹和法令纹，眼部的全脸细纹，青森淡季功能鸭子熨斗。'''
44 | 


--------------------------------------------------------------------------------