我使用的服务是：识别-REST_API

调用的接口是：http://vop.baidu.com/server_api

参数设置情况：

stt_para = {'format':'wav', 'rate':8000, 'channel':1, 'cuid':'7C-E9-D3-11-1E-69', 'token':'24.1d33d3ebea11d4********64.282335-1335399','len':speech_len ,'speech':speech}

开发平台和机型：windows,python3.6

复现步骤、现象及其他描述：

import requests
import base64
import json

#tok = requests.get(tokurl)

tts_url = 'http://tsn.baidu.com/text2audio'
tts_para = {'tex':'你好凯文','lan':'zh','tok':'24.1d33d3ebea*********63664.282335-1335399','ctp':1,'cuid':'0A-00-27-00-00-12'}
tts_data = requests.post(tts_url, data=tts_para)

with open('hello.wav', 'wb') as f:
f.write(tts_data.content)

input()

stt_url = 'http://vop.baidu.com/server_api'
headers = {'content-type': 'application/json;charset=utf-8'}
with open('hello.wav','rb') as f:
bdata = f.read()
_speech = base64.b64encode(bdata)
speech = _speech.decode() #这一步是为了转成字符串，不然会报TypeError: Object of type 'bytes' is not JSON serializable
   speech_len = len(bdata)
stt_para = {'format':'wav', 'rate':8000, 'channel':1, 'cuid':'7C-E9-D3-11-1E-69', 'token':'24.1d33d3ebea11d4***********64.282335-1335399','len':speech_len ,'speech':speech}
stt_data = requests.post(stt_url, data=json.dumps(stt_para),headers=headers)
input()

合成语间没问题，但用这个合成的语音去识别老是报3301

{"err_msg":"recognition error.","err_no":3301,"sn":"883003405861493389441"}