python调用百度语音REST API

时间:2022-11-15 15:10:44

本文实例为大家分享了python调用百度语音REST API的具体代码,供大家参考,具体内容如下

(百度的rest接口的部分网址发生了一定的变化,相关代码已更新)

百度通过 REST API 的方式给开发者提供一个通用的 HTTP 接口,基于该接口,开发者可以轻松的获得语音合成与语音识别能力。SDK中只提供了PHP、C和JAVA的相关样例,使用python也可以灵活的对端口进行调用,本文描述了简单使用Python调用百度语音识别服务 REST API 的简单样例。

1、语音识别与语音合成的调用

注册开发者帐号和创建应用的过程就不再赘述,百度的REST API在调用过程基本分为三步:

  • 获取token
  • 向Rest接口提交数据
  • 处理返回数据

具体代码如下所示:

 

?
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
#!/usr/bin/python3
 
import urllib.request
import urllib
import json
import base64
class BaiduRest:
  def __init__(self, cu_id, api_key, api_secert):
    # token认证的url
    self.token_url = "https://openapi.baidu.com/oauth/2.0/token?grant_type=client_credentials&client_id=%s&client_secret=%s"
    # 语音合成的resturl
    self.getvoice_url = "http://tsn.baidu.com/text2audio?tex=%s&lan=zh&cuid=%s&ctp=1&tok=%s"
    # 语音识别的resturl
    self.upvoice_url = 'http://vop.baidu.com/server_api'
 
    self.cu_id = cu_id
    self.getToken(api_key, api_secert)
    return
 
  def getToken(self, api_key, api_secert):
    # 1.获取token
    token_url = self.token_url % (api_key,api_secert)
 
    r_str = urllib.request.urlopen(token_url).read()
    token_data = json.loads(r_str)
    self.token_str = token_data['access_token']
    pass
 
  def getVoice(self, text, filename):
    # 2. 向Rest接口提交数据
    get_url = self.getvoice_url % (urllib.parse.quote(text), self.cu_id, self.token_str)
 
    voice_data = urllib.request.urlopen(get_url).read()
    # 3.处理返回数据
    voice_fp = open(filename,'wb+')
    voice_fp.write(voice_data)
    voice_fp.close()
    pass
 
  def getText(self, filename):
    # 2. 向Rest接口提交数据
    data = {}
    # 语音的一些参数
    data['format'] = 'wav'
    data['rate'] = 8000
    data['channel'] = 1
    data['cuid'] = self.cu_id
    data['token'] = self.token_str
    wav_fp = open(filename,'rb')
    voice_data = wav_fp.read()
    data['len'] = len(voice_data)
    data['speech'] = base64.b64encode(voice_data).decode('utf-8')
    post_data = json.dumps(data)
    r_data = urllib.request.urlopen(self.upvoice_url,data=bytes(post_data,encoding="utf-8")).read()
    # 3.处理返回数据
    return json.loads(r_data)['result']
 
if __name__ == "__main__":
  # 我的api_key,供大家测试用,在实际工程中请换成自己申请的应用的key和secert
  api_key = "SrhYKqzl3SE1URnAEuZ0FKdT"
  api_secert = "hGqeCkaMPb0ELMqtRGc2VjWdmjo7T89d"
  # 初始化
  bdr = BaiduRest("test_python", api_key, api_secert)
  # 将字符串语音合成并保存为out.mp3
  bdr.getVoice("你好北京邮电大学!", "out.mp3")
  # 识别test.wav语音内容并显示
  print(bdr.getText("out.wav"))

2、调用pyaudio使用麦克风录制声音

python中的pyaudio库可以直接通过麦克风录制声音,可使用pip进行安装。我们可以通过调用该库,获取到wav测试语音。
具体代码如下所示:

?
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
#!/usr/bin/python3
# -*- coding: utf-8 -*-
 
from pyaudio import PyAudio, paInt16
import numpy as np
from datetime import datetime
import wave
 
class recoder:
  NUM_SAMPLES = 2000   #pyaudio内置缓冲大小
  SAMPLING_RATE = 8000  #取样频率
  LEVEL = 500     #声音保存的阈值
  COUNT_NUM = 20   #NUM_SAMPLES个取样之内出现COUNT_NUM个大于LEVEL的取样则记录声音
  SAVE_LENGTH = 8     #声音记录的最小长度:SAVE_LENGTH * NUM_SAMPLES 个取样
  TIME_COUNT = 60   #录音时间,单位s
 
  Voice_String = []
 
  def savewav(self,filename):
    wf = wave.open(filename, 'wb')
    wf.setnchannels(1)
    wf.setsampwidth(2)
    wf.setframerate(self.SAMPLING_RATE)
    wf.writeframes(np.array(self.Voice_String).tostring())
    # wf.writeframes(self.Voice_String.decode())
    wf.close()
 
  def recoder(self):
    pa = PyAudio()
    stream = pa.open(format=paInt16, channels=1, rate=self.SAMPLING_RATE, input=True,
      frames_per_buffer=self.NUM_SAMPLES)
    save_count = 0
    save_buffer = []
    time_count = self.TIME_COUNT
 
    while True:
      time_count -= 1
      # print time_count
      # 读入NUM_SAMPLES个取样
      string_audio_data = stream.read(self.NUM_SAMPLES)
      # 将读入的数据转换为数组
      audio_data = np.fromstring(string_audio_data, dtype=np.short)
      # 计算大于LEVEL的取样的个数
      large_sample_count = np.sum( audio_data > self.LEVEL )
      print(np.max(audio_data))
      # 如果个数大于COUNT_NUM,则至少保存SAVE_LENGTH个块
      if large_sample_count > self.COUNT_NUM:
        save_count = self.SAVE_LENGTH
      else:
        save_count -= 1
 
      if save_count < 0:
        save_count = 0
 
      if save_count > 0 :
      # 将要保存的数据存放到save_buffer中
        #print save_count > 0 and time_count >0
        save_buffer.append( string_audio_data )
      else:
      #print save_buffer
      # 将save_buffer中的数据写入WAV文件,WAV文件的文件名是保存的时刻
        #print "debug"
        if len(save_buffer) > 0 :
          self.Voice_String = save_buffer
          save_buffer = []
          print("Recode a piece of voice successfully!")
          return True
      if time_count==0:
        if len(save_buffer)>0:
          self.Voice_String = save_buffer
          save_buffer = []
          print("Recode a piece of voice successfully!")
          return True
        else:
          return False
 
if __name__ == "__main__":
  r = recoder()
  r.recoder()
  r.savewav("test.wav"

以上就是本文的全部内容,希望对大家的学习有所帮助,也希望大家多多支持服务器之家。

原文链接:https://blog.csdn.net/wolfblood_zzx/article/details/46418635