需求描述

  1. 使用python3调用百度tts，生成mp3文件
  2. 调用ffmpeg将mp3转为8k8bit的wav文件，用于FreeSWITCH的播放

实现步骤

申请百度tts账号

参考：链接

python实现代码

baidutts.pytts.py

文件：baidutts.py

  1. # coding=utf-8
  3. import sys
  4. import json
  6. from urllib.request import urlopen
  7. from urllib.request import Request
  8. from urllib.error import URLError
  9. from urllib.parse import urlencode
  10. from urllib.parse import quote_plus
  12. IS_PY3 = sys.version_info.major == 3
  14. # 替换你的 API_KEY
  15. API_KEY = 'zqyXXkMmH0grrVh7N5tPPxxx'
  17. # 替换你的 SECRET_KEY
  18. SECRET_KEY = 'KV7iADpCqm19q8xj9XkvaKZuHEorXxxx'
  20. TTS_URL = 'http://tsn.baidu.com/text2audio'
  22. """  TOKEN start """
  24. TOKEN_URL = 'http://openapi.baidu.com/oauth/2.0/token'
  26. """
  27.     获取token
  28. """
  31. def fetch_token():
  32.     params = {'grant_type': 'client_credentials',
  33.               'client_id': API_KEY,
  34.               'client_secret': SECRET_KEY}
  35.     post_data = urlencode(params)
  36.     if (IS_PY3):
  37.         post_data = post_data.encode('utf-8')
  38.     req = Request(TOKEN_URL, post_data)
  39.     try:
  40.         f = urlopen(req, timeout=5)
  41.         result_str = f.read()
  42.     except URLError as err:
  43.         print('token http response http code : ' + str(err.code))
  44.         result_str = err.read()
  45.     if (IS_PY3):
  46.         result_str = result_str.decode()
  48.     result = json.loads(result_str)
  50.     if 'access_token' in result.keys() and 'scope' in result.keys():
  51.         if not 'audio_tts_post' in result['scope'].split(' '):
  52.             print('please ensure has check the tts ability')
  53.             exit()
  54.         return result['access_token']
  55.     else:
  56.         print('please overwrite the correct API_KEY and SECRET_KEY')
  57.         exit()
  60. """  TOKEN end """
  63. def get_tts(myText, fileName):
  64.     token = fetch_token()
  66.     tex = quote_plus(myText)  # 此处TEXT需要两次urlencode
  68.     params = {'tok': token, 'tex': tex, 'cuid': "quickstart",
  69.               'lan': 'zh', 'ctp': 1}  # lan ctp 固定参数
  71.     data = urlencode(params)
  73.     req = Request(TTS_URL, data.encode('utf-8'))
  74.     has_error = False
  75.     try:
  76.         f = urlopen(req)
  77.         result_str = f.read()
  79.         headers = dict((name.lower(), value) for name, value in f.headers.items())
  81.         has_error = ('content-type' not in headers.keys() or headers['content-type'].find('audio/') < 0)
  82.     except URLError as err:
  83.         print('http response http code : ' + str(err.code))
  84.         result_str = err.read()
  85.         has_error = True
  87.     save_file = "error.txt" if has_error else fileName
  89.     with open(save_file, 'wb') as of:
  90.         of.write(result_str)
  92.     if has_error:
  93.         if IS_PY3:
  94.             result_str = str(result_str, 'utf-8')
  95.         print("tts api  error:" + result_str)
  97.     print("file saved as : " + save_file)

文件：tts.py

  1. from baidutts import get_tts
  2. import os
  4. text = "你好，你来自什么地方，那个地方漂亮吗"
  5. filemp3 = 'media/text.mp3'
  6. # 输出的8k8bit的文件
  7. finalwav = 'media/my.wav'
  9. # 使用百度tts生成对应的mp3文件
  10. get_tts(text, filemp3)
  12. # 将mp3转为8k8bit的wav文件
  13. cmd = f'ffmpeg -i {filemp3} -ar 8000 -ac 1 -acodec pcm_u8 {finalwav}'
  14. os.system(cmd)
  16. # 删除生成的中间文件
  17. os.remove(filemp3)