python语音计算信噪比

import numpy as np
import librosa


#计算信噪比
def SNR_singlech(clean_file, original_file):
    clean, clean_fs = librosa.load(clean_file, sr=None, mono=True)#导入干净语音
    ori, ori_fs = librosa.load(original_file, sr=None, mono=True)#导入原始语音
    length = min(len(clean), len(ori))
    est_noise = ori[:length] - clean[:length]#计算噪声语音
    
    #计算信噪比
    SNR = 10*np.log10((np.sum(clean**2))/(np.sum(est_noise**2)))
    print(SNR)

SNR_singlech('sf1_clean.wav', 'sf1_noisy.wav')

上述代码仅限于已知干净语音与含噪声语音的情况