speakerfeatures.py

# -*- coding: utf-8 -*-
"""
Created on Mon Sep 14 19:26:59 2015

@author: Abhijeet Kumar
@code :  This program implemets feature (MFCC + delta)
         extraction process for an audio. 
@Note :  20 dim MFCC(19 mfcc coeff + 1 frame log energy)
         20 dim delta computation on MFCC features. 
@output : It returns 40 dimensional feature vectors for an audio.
"""

import numpy as np
from sklearn import preprocessing
import python_speech_features as mfcc

def calculate_delta(array):
    """Calculate and returns the delta of given feature vector matrix"""

    rows,cols = array.shape
    deltas = np.zeros((rows,20))
    N = 2
    for i in range(rows):
        index = []
        j = 1
        while j <= N:
            if i-j < 0:
                first = 0
            else:
                first = i-j
            if i+j > rows -1:
                second = rows -1
            else:
                second = i+j
            index.append((second,first))
            j+=1
        deltas[i] = ( array[index[0][0]]-array[index[0][1]] + (2 * (array[index[1][0]]-array[index[1][1]])) ) / 10
    return deltas

def extract_features(audio,rate):
    """extract 20 dim mfcc features from an audio, performs CMS and combines 
    delta to make it 40 dim feature vector"""    
    
    mfcc_feat = mfcc.mfcc(audio,rate, 0.025, 0.01,20,appendEnergy = True)
    
    mfcc_feat = preprocessing.scale(mfcc_feat)
    delta = calculate_delta(mfcc_feat)
    combined = np.hstack((mfcc_feat,delta)) 
    return combined
#    
if __name__ == "__main__":
     print "In main, Call extract_features(audio,signal_rate) as parameters"