Third International Conference on Spoken Language Processing (ICSLP 94)

Yokohama, Japan
September 18-22, 1994

Mel-Generalized Cepstral Analysis - A Unified Approach to Speech Spectral Estimation

Keiichi Tokuda (1), Takao Kobayashi (2), Takashi Masuko (2), Satoshi Imai (2)

(1) Department of Electrical and Electronic Engineering, Tokyo Institute of Technology, Tokyo, Japan
(2) Precision and Intelligence Laboratory, Tokyo Institute of Technology, Yokohama, Japan

The generalized cepstral analysis method is viewed as a unified approach to the cepstral method and the linear prediction method, in which the model spectrum varies continuously from all-pole to cepstral according to the value of a parameter 7. Since the human ear has high resolution at low frequencies, introducing similar characteristics to the model spectrum, we can represent speech spectrum more efficiently. From this point of view, this paper proposes a spectral estimation method which uses the spectral model represented by mel-generalized cepstral coefficients. The effectiveness of mel-generalized cepstral analysis is demonstrated by an experiment of HMM-based isolated word recognition.

Full Paper

Bibliographic reference.  Tokuda, Keiichi / Kobayashi, Takao / Masuko, Takashi / Imai, Satoshi (1994): "Mel-generalized cepstral analysis - a unified approach to speech spectral estimation", In ICSLP-1994, 1043-1046.