A comparison of LPC and FFT-based acoustic features for noise robust ASR

Abstract

Within the context of robust acoustic features for automatic speech recognition (ASR), we evaluated mel-frequency cepstral coefficients (MFCCs) derived from two spectral representation techniques, i.e. the fast Fourier transform (FFT) and linear pre­ dictive coding (LPC). ASR systems based on the two feature types were tested on a digit recognition task using continuous density hidden Markov phone models. System performance was determined in clean acoustic conditions as well as in differ­ ent simulations of adverse acoustic conditions. The LPC-based MFCCs outperformed their FFT counterparts in most of the ad­ verse acoustic conditions that were investigated in this study. A tentative explanation for this difference in recognition perfor­ mance is given.