FeatureExtraction.py

#!/usr/bin/env python
# coding: utf-8

# In[ ]:


import librosa
import librosa.display
import numpy as np
import pandas as pd
import scipy
from scipy.stats import skew

import matplotlib.pyplot as plt

SAMPLE_RATE = 44100
def get_mfcc(name, path):
    data, _ = librosa.core.load(path + name, sr = SAMPLE_RATE)
    assert _ == SAMPLE_RATE
    try:
        ft1 = librosa.feature.mfcc(data, sr = SAMPLE_RATE, n_mfcc=30)
        ft2 = librosa.feature.zero_crossing_rate(data)[0]
        ft3 = librosa.feature.spectral_rolloff(data)[0]
        ft4 = librosa.feature.spectral_centroid(data)[0]
        ft5 = librosa.feature.spectral_contrast(data)[0]
        ft6 = librosa.feature.spectral_bandwidth(data)[0]
        ft1_trunc = np.hstack((np.mean(ft1, axis=1), np.std(ft1, axis=1), skew(ft1, axis = 1), np.max(ft1, axis = 1), 
                               np.median(ft1, axis = 1), np.min(ft1, axis = 1)))
        ft2_trunc = np.hstack((np.mean(ft2), np.std(ft2), skew(ft2), np.max(ft2), np.median(ft2), np.min(ft2)))
        ft3_trunc = np.hstack((np.mean(ft3), np.std(ft3), skew(ft3), np.max(ft3), np.median(ft3), np.min(ft3)))
        ft4_trunc = np.hstack((np.mean(ft4), np.std(ft4), skew(ft4), np.max(ft4), np.median(ft4), np.min(ft4)))
        ft5_trunc = np.hstack((np.mean(ft5), np.std(ft5), skew(ft5), np.max(ft5), np.median(ft5), np.min(ft5)))
        ft6_trunc = np.hstack((np.mean(ft6), np.std(ft6), skew(ft6), np.max(ft6), np.median(ft6), np.max(ft6)))
        return pd.Series(np.hstack((ft1_trunc, ft2_trunc, ft3_trunc, ft4_trunc, ft5_trunc, ft6_trunc)))
    except:
        print('bad file')
        return pd.Series([0]*210)

def MFCC_spectrogram(y, sr):
    # 提取頻域音頻資料 MFCC spectrogram feature  
    melspec = librosa.feature.melspectrogram(y, sr, n_fft=1024, hop_length=512, n_mels=128)  
    # 轉成 log scale 已呈現頻譜資料  
    logmelspec = librosa.power_to_db(melspec)  
    # 繪圖  
    plt.figure()  
    librosa.display.specshow(logmelspec, sr=sr, x_axis='time', y_axis='mel')  
    plt.title('Features of unknow frog by MFCC (Mel Frquency Cepstral Coefficients)')  
    plt.show()



def FeatureExtraction(fname):
    # 創建文件檔
    test= pd.DataFrame({'fname':[fname]})
    # 特徵提取
    test_data = test['fname'].progress_apply(get_mfcc, path='')
    test_data = test_data.values
    return test_data