ISSN : 1796-2048
Volume : 2    Issue : 2    Date : April 2007

Noisy Speech Feature Estimation on the Aurora2 Database using a Switching Linear Dynamic Mode
Jianping Deng, Martin Bouchard, and Tet Hin Yeap
Page(s): 47-52
Full Text:
PDF (333 KB)

This paper presents an approach to enhance speech feature estimation in the log spectral domain
under additive noise environments. A switching linear dynamic model (SLDM) is explored as a
parametric model for the clean speech distribution, enforcing a state transition in the feature space
and capturing the smooth time evolution of speech conditioned on the state sequence. Experimental
results using the Aurora2 database show that the new SLDM approach can improve speech
enhancement performance in terms of recognition accuracy.

Index Terms
speech feature enhancement, speech recognition, switching linear dynamic model, hidden Markov