Time and frequency filtering for speech recognition with real noises

Abstract

very speech recognition system requires a signal representation that parametrically models the temporal evolution of the speech spectral envelope. Current parameterizations involve, either explicitly or implicitly, a set of energies from frequency bands which are often distributed in a mel scale. The computation of those energies is performed in diverse ways, but it always includes smoothing of basic spectral measurements and non-linear amplitude compression. Several linear transformations are then applied to the two-dimensional time-frequency sequence of energies before entering the HMM pattern matching stage. In this paper, a recently introduced technique that consists of ®ltering that sequence of energies along the frequency dimension is presented, and its resulting parameters are compared with the widely used cepstral coe cients. Then, that frequency ®ltering transformation is jointly considered with the time ®ltering transformation that is used to compute dynamic parameters, showing that the ¯exibility of this combined (ti ng) approach can be used to design a robust set of ®lters. Recognition experiment results are reported which show the potential of ti ng for an enhanced and more robust HMM speech recognition. Ó 2001 Elsevier Science B.V. All rights reserved.