Postfiltering Using Multichannel Spectral Estimation in Multispeaker Environments
- PDF / 3,467,810 Bytes
- 10 Pages / 600.05 x 792 pts Page_size
- 58 Downloads / 184 Views
Research Article Postfiltering Using Multichannel Spectral Estimation in Multispeaker Environments Hai Quang Dam, Sven Nordholm, Hai Huyen Dam, and Siow Yong Low Western Australian Telecommunications Research Institute (WATRI), Crawley, WA 6009, Australia Correspondence should be addressed to Hai Quang Dam, [email protected] Received 14 September 2006; Accepted 5 July 2007 Recommended by Douglas O’Shaughnessy This paper investigates the problem of enhancing a single desired speech source from a mixture of signals in multispeaker environments. A beamformer structure is proposed which combines a fixed beamformer with postfiltering. In the first stage, the fixed multiobjective optimal beamformer is designed to spatially extract the desired source by suppressing all other undesired sources. In the second stage, a multichannel power spectral estimator is proposed and incorporated in the postfilter, thus enabling further suppression capability. The combined scheme exploits both spatial and spectral characteristics of the signals. Two new multichannel spectral estimation methods are proposed for the postfiltering using, respectively, inner product and joint diagonalization. Evaluations using recordings from a real-room environment show that the proposed beamformer offers a good interference suppression level whilst maintaining a low-distortion level of the desired source. Copyright © 2008 Hai Quang Dam et al. This is an open access article distributed under the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
1.
INTRODUCTION
Multichannel beamforming techniques can be largely divided into three types, namely, fixed, optimum, and adaptive beamforming [1, 2]. For a fixed beamformer, the beamformer weights, which usually consist of FIR-filter weights, are designed to focus into a main source direction while suppressing signals from other undesired directions. This problem can be viewed as a multidimensional filter design problem [2]. As such, the weights are calculated based on information about the array geometry and the source localization with no statistical information about the signal’s environment or the required signals. Multichannel optimum filtering, on the other hand, requires statistical knowledge about the noise statistics, the environment, and the source statistics. The beamformer coefficients are optimized in such a manner that a focussed beam is steered to a desired source direction, whilst suppressing the contributions coming from other directions [2, 3]. Similar to the fixed beamformer case, the design also requires information about the location of the target signal and the array geometry. From those parameters, a spatial, spectral, and temporal filter is formed to match the beamforming requirement [4, 5].
Adaptive beamforming techniques are developed to track time-varying signal situations [6, 7]. A well-known technique is to combine the beamformer with an adaptive postfiltering technique. The
Data Loading...
