Extraction of speech signal based on Power Normalized Cepstral Coefficient and Mel Frequency Cepstral Coefficient: A comparison

摘要

Speech processing is emerged as one of the important application area of digital signal processing. Power Normalized Cepstral Coefficients (PNCC) and Mel Frequency Cepstral Coefficient (MFCC) are mainly used in feature extraction of speech signals. The problem of real time speaker segmentation in speech processing is enormous in which no prior knowledge about the number of speakers and the identities of speakers are available. In this paper the performances of the PNCC and MFCC are compared and the experimental results demonstrate that PNCC processing provides improvements in recognition accuracy in the presence of various types of noises and in environmental changes. The performance of PNCC method is robust to natural and unpredictable situations.