Title :
A mathematical relationship between full-band and multiband mel-frequency cepstral coefficients
Author_Institution :
Dept. of Comput. Sci., Hong Kong Univ. of Sci. & Technol., Clear Water Bay, China
Abstract :
Recently, it has been shown that robustness of automatic speech recognition (ASR) against band-limited additive noises may be improved by multiband ASR (MBASR) approaches. In an M-subband MBASR system, the channels in the full-band filterbank are divided into M subbands, usually of equal partitions, and subband mel-frequency cepstral coefficients (MFCCs) are computed from each filterbank partition using the discrete cosine transform. However, there is not as yet any analysis on the relationship between full-band and multiband MFCCs. In this letter, we show that the (Mj)th full-band MFCC is the sum of or difference between the Mjth multiband MFCCs multiplied by /spl radic/M.
Keywords :
cepstral analysis; discrete cosine transforms; speech recognition; ASR; M-subband MBASR system; automatic speech recognition; band-limited additive noise; discrete cosine transform; filterbank partition; full-band coefficients; full-band filterbank; mel-frequency cepstral coefficients; multiband coefficients; subband coefficients; Additive noise; Automatic speech recognition; Bandwidth; Cepstral analysis; Discrete cosine transforms; Filter bank; Hidden Markov models; Mel frequency cepstral coefficient; Noise robustness; Signal processing;
Journal_Title :
Signal Processing Letters, IEEE
DOI :
10.1109/LSP.2002.803007