python中mp3转wav---安装、声音截取、音频分布
一、pydub和ffmpeg安装（失败）
1、直接pip安装
pip install pydub
pip install FFmpeg
运行下文程序时，报错：Couldn’t find ffmpeg or avconv - defaulting to ffmpeg, but may not work"
主要原因是没有找到 ffmpeg.exe。pip list查看安装版本为1.4。
二、单独下载ffmped，将bin文件并添加到环境变量（失败）
下载链接：https://ffmpeg.zeranoe.com/builds/
仍然报错：Couldn’t find ffmpeg or avconv - defaulting to ffmpeg, but may not work"
三、使用镜像进行安装（成功）
conda install -c https://conda.anaconda.org/menpo ffmpeg
conda-forge是编译的ffmpeg视频编解码库，它可直接对视频读存取
conda-forge是编译的ffmpeg视频编解码库，它可直接对视频读存取
四、mp3转wav代码
1、将.mp3转化为.wav格式
from scipy.io import wavfile
import wave
import numpy as np
import matplotlib.pyplot as plt
from pydub import AudioSegment
#定义转化格式函数
def trans_mp3_to_wav(filepath):
print(filepath)
song = AudioSegment.from_mp3(filepath)
song.export("now.wav", format="wav")
trans_mp3_to_wav(source_file_path)
2、录音详情
wave_path = "now.wav"
file = wave.open(wave_path)
# print('---------声音信息------------')
# for item in enumerate(file.getparams()):
# print(item)
a = file.getparams().nframes # 帧总数
f = file.getparams().framerate # 采样频率
sample_time = 1 / f # 采样点的时间间隔
time = a / f # 声音信号的长度
sample_frequency, audio_sequence = wavfile.read(wave_path)
# print(audio_sequence) # 声音信号每一帧的“大小”
x_seq = np.arange(0, time, sample_time)
plt.plot(x_seq, audio_sequence, 'blue')
plt.xlabel("time (s)")
plt.show()
3、截取录音片段
def get_part_wav(main_wav_path, start_time, end_time,part_wav_path):
start_time = int(start_time)
end_time = int(end_time)
sound = AudioSegment.from_mp3(main_wav_path)
word = sound[start_time:end_time]
word.export(part_wav_path, format="wav")
main_wav_path = 'index.mp3'
part_wav_path = 'part.wav'
start_time = 155800
end_time = 160020
get_part_wav(main_wav_path, start_time, end_time,part_wav_path)
参考文献：
1、python 安装 ffmpeg包–简便方法
2、python处理wav音频文件：音频信息，读取内容，获取时长，切割音频，pcm与wav互转