melspectrogram_snore.py

import numpy as np
from librosa.feature import melspectrogram
import os
from scipy import signal, misc
from scipy.io import wavfile
import matplotlib.pyplot as plt


save_file = "../snore_melspectrogram_1/"

if not os.path.exists(save_file):
    os.mkdir(save_file)

if not os.path.exists(save_file+"train"):
    os.mkdir(save_file+"train")

if not os.path.exists(save_file+"devel"):
    os.mkdir(save_file+"devel")

if not os.path.exists(save_file+"test"):
    os.mkdir(save_file+"test")

i_train = 0
i_devel = 0
i_test = 0

for signal in os.listdir("../Snore_dist/wav"):

    sig_vec = signal.split('_')
    sig_name = sig_vec[0]
    sig_number = int(sig_vec[1].split(".")[0])
    print ("This is the current signal to be processed: "+signal)
    sigbag = wavfile.read("../Snore_dist/wav/"+signal)
    sig = sigbag[1]
    #sig = sig / np.amax(abs(sig))
    #sig = wiener(sig, mysize=1600)

    if sig_name == "train":
        i_train = sig_number
    elif sig_name == "devel":
        i_devel = sig_number
    elif sig_name == "test":
        i_test = sig_number

    #sig = sig[:36000]
    #sig = sig / np.amax(abs(sig))
    if sig.shape[0] >= 44000:
        pass
    else:
        while sig.shape[0] < 44000:
            sig = np.concatenate((sig, sig[0:44000-sig.shape[0]]))
    ms = melspectrogram(y=sig, sr=16000,n_fft=350, hop_length=175, power=2, n_mels=176)

    # save the resultant gradient matrix as image
    if sig_name == "train":
        plt.imsave(save_file+"train/"+str(i_train)+'.png',ms,cmap='viridis')
    elif sig_name == "devel":
        plt.imsave(save_file+"devel/"+str(i_devel)+'.png',ms,cmap='viridis')
    elif sig_name == "test":
        plt.imsave(save_file+"test/"+str(i_test)+'.png',ms,cmap='viridis')

    if sig_name == "train":
        print ("This is train "+str(i_train)+" finished!")
    elif sig_name == "devel":
        print ("This is devel "+str(i_devel)+" finished!")
    elif sig_name == "test":
        print ("This is test "+str(i_test)+" finished!")