Skip to main content

Lund University Publications

LUND UNIVERSITY LIBRARIES

A hybrid design of beamformers for voice control devices

Yui, K.F.C ; Chan, K.Y ; Grbic, Nedelko LU and Nordholm, Sven (2012) In Pacific Journal of Optimization 8(3). p.533-544
Abstract
In this paper, a new approach to designing beamformers for voice control device is proposed. It is well-known that under a strong near-field noise with low signal-to-noise ratios (SNR), the performance of speech recognition is deteriorated significantly. However, designing the beamformer for enhancing speech recognition is a slow process and might not be adapted. easily to the changing noise environment. In order to lower the complexity of the design, we intend to exploit the combination of existing optimal beamformer designs, which can be implemented in parallel. These include the least-squares technique and the signal-to-noise ratio maximization technique. We show here that for a given pre-trained speech recognizer and for a finite set... (More)
In this paper, a new approach to designing beamformers for voice control device is proposed. It is well-known that under a strong near-field noise with low signal-to-noise ratios (SNR), the performance of speech recognition is deteriorated significantly. However, designing the beamformer for enhancing speech recognition is a slow process and might not be adapted. easily to the changing noise environment. In order to lower the complexity of the design, we intend to exploit the combination of existing optimal beamformer designs, which can be implemented in parallel. These include the least-squares technique and the signal-to-noise ratio maximization technique. We show here that for a given pre-trained speech recognizer and for a finite set of speech commands, neither method has a satisfactory performance in speech recognition accuracy under very low SNRs. However, since the two techniques have different characteristics in speech distortion and noise suppression, we show that it is possible to enhance the speech recognition accuracy by combining these two optimal designs. (Less)
Please use this url to cite or link to this publication:
author
; ; and
publishing date
type
Contribution to journal
publication status
published
subject
in
Pacific Journal of Optimization
volume
8
issue
3
pages
533 - 544
publisher
Yokohama Publishers
external identifiers
  • scopus:84874957516
ISSN
1348-9151
language
English
LU publication?
no
id
3fefb5f3-c09a-4c45-92fa-ea9167c716b3
date added to LUP
2016-06-23 14:16:23
date last changed
2024-01-04 08:54:42
@article{3fefb5f3-c09a-4c45-92fa-ea9167c716b3,
  abstract     = {{In this paper, a new approach to designing beamformers for voice control device is proposed. It is well-known that under a strong near-field noise with low signal-to-noise ratios (SNR), the performance of speech recognition is deteriorated significantly. However, designing the beamformer for enhancing speech recognition is a slow process and might not be adapted. easily to the changing noise environment. In order to lower the complexity of the design, we intend to exploit the combination of existing optimal beamformer designs, which can be implemented in parallel. These include the least-squares technique and the signal-to-noise ratio maximization technique. We show here that for a given pre-trained speech recognizer and for a finite set of speech commands, neither method has a satisfactory performance in speech recognition accuracy under very low SNRs. However, since the two techniques have different characteristics in speech distortion and noise suppression, we show that it is possible to enhance the speech recognition accuracy by combining these two optimal designs.}},
  author       = {{Yui, K.F.C and Chan, K.Y and Grbic, Nedelko and Nordholm, Sven}},
  issn         = {{1348-9151}},
  language     = {{eng}},
  number       = {{3}},
  pages        = {{533--544}},
  publisher    = {{Yokohama Publishers}},
  series       = {{Pacific Journal of Optimization}},
  title        = {{A hybrid design of beamformers for voice control devices}},
  volume       = {{8}},
  year         = {{2012}},
}