2021-11-03

##将批量音频文件 .ogg文件转换成.wav.或是其他音频文件转换成.wav

from librosa.util.exceptions import LibrosaError
from pydub import AudioSegment
from pydub.utils import make_chunks
from pydub.silence import detect_silence
import os
import librosa
import random
import numpy as np

out_path = "zhuan"
file_path = "./zyang/xiaojun"
ss = os.listdir(file_path)
file = []
for files in ss:
    # if files.endswith('.wav'or'.mp3'):
        file.append(files)
for i in range(len(ss)):
    
      sound = AudioSegment.from_file(file_path+'/'+file[i])
      
      sound.export(out_path+'/'+file[i].split('.')[0]+'.wav', format="wav")

##将音频文件重采样，将采样频率由48000->16000.

from logging import _srcfile
from librosa.util.exceptions import LibrosaError
from pydub import AudioSegment
from pydub.utils import make_chunks
from pydub.silence import detect_silence
import os
import librosa
import random
import numpy as np
import torchaudio

out_path = "./speech_commands/zhuan16"
file_path = "./speech_commands/zhuan"
ss = os.listdir(file_path)
file = []
for files in ss:
    # if files.endswith('.wav'or'.mp3'):
        file.append(files)
for i in range(len(ss)):

    waveform, sr = torchaudio.load(file_path+'/'+file[i])
    waveform = torchaudio.transforms.Resample(orig_freq=48000, new_freq=16000)(waveform)

    torchaudio.save(out_path+'/'+file[i], waveform, sample_rate=16000)