The classical front end analysis in speech recognition is a spectral analysis which parametrizes the speech signal into feature vectors; the most popular set of them is the Mel Frequency Cepstral Coefficients (MFCC). They ...