Files in this item



application/pdfECE499-Sp2009-hu.pdf (896kB)Restricted to U of Illinois
(no description provided)PDF


Title:Maximum Kurtosis Noise Cancellation Technique
Author(s):Hu, Dan
Contributor(s):Jones, Douglas L.
Subject(s):speech enhancement
noise cancellation
blind source separation
speech recovery algorithms
Abstract:Over the past decades, there has been considerable research attention on speech enhancement techniques. This study explores a few typical noise-cancellation algorithms and reviews an innovative method which combines approaches from the Blind Source Separation (BSS) class and the Beamforming class. Many speech-recovery applications deal with extracting a high-kurtosis speech source from low-kurtosis, non-stationary environmental noise. The speech-recovery technique discussed in this paper uses the maximum-kurtosis distortionless response (MKDR) algorithm in estimating the source-to-microphone response and applies the Minimum Variance Distortionless Response (MVDR) beamformer to recover the speech. This paper suggests a fast practical implementation of the MVDR algorithm and examines its speech-enhancement performance via Matlab simulation. Finally, we extend our discussion to commercial applications by constructing a two-microphone, small-aperture cell-phone model and test the algorithm with real-world recordings.
Issue Date:2009-05
Publication Status:unpublished
Peer Reviewed:not peer reviewed
Date Available in IDEALS:2014-01-22

This item appears in the following Collection(s)

Item Statistics