# -*- coding:utf-8 -*-
# -*- author:zzZ_CMing CSDN address:https://blog.csdn.net/zzZ_CMing
# -*- 2018/07/12; 15:19
# -*- python3.5
import pyaudio
import wave
input_filename = "input.wav" # 麦克风采集的语音输入
input_filepath = "音频存储位置" # 输入文件的path
in_path = input_filepath + input_filename
def get_audio(filepath):
aa = str(input("是否开始录音? (是/否)"))
if aa == str("是") :
CHUNK = 256
FORMAT = pyaudio.paInt16
CHANNELS = 1 # 声道数
RATE = 11025 # 采样率
RECORD_SECONDS = 5
WAVE_OUTPUT_FILENAME = filepath
p = pyaudio.PyAudio()
stream = p.open(format=FORMAT,
channels=CHANNELS,
rate=RATE,
input=True,
frames_per_buffer=CHUNK)
print("*"*10, "开始录音:请在5秒内输入语音")
frames = []
for i in range(0, int(RATE / CHUNK * RECORD_SECONDS)):
data = stream.read(CHUNK)
frames.append(data)
print("*"*10, "录音结束\n")
stream.stop_stream()
stream.close()
p.terminate()
wf = wave.open(WAVE_OUTPUT_FILENAME, 'wb')
wf.setnchannels(CHANNELS)
wf.setsampwidth(p.get_sample_size(FORMAT))
wf.setframerate(RATE)
wf.writeframes(b''.join(frames))
wf.close()
elif aa == str("否"):
exit()
else:
print("无效输入,请重新选择")
get_audio(in_path)
get_audio(in_path)
说明:
- 这是一个可循环采样的脚本,由于输出文件名都是
input.wav
,所以在你保存音频的地址下永远只生成一个音频文件(后一个会覆盖前一个); - 可以结合科大讯飞语音听写的SDK包,完成语音转文本:
# 采集音频
get_audio(in_path)
sys.stdout.write("you ask>> ")
# 科大语音听写:语音转文本输出 请参考上一篇博客
input_seqs = keda_API.XF_text(in_path, audio_rate)