Publication detail

Application of Mel-scale Filter bank for Noise Estimation in Speech Processing

MOTLÍČEK, P., BURGET, L.

Original Title

Application of Mel-scale Filter bank for Noise Estimation in Speech Processing

English Title

Application of Mel-scale Filter bank for Noise Estimation in Speech Processing

Type

conference paper

Language

en

Original Abstract

The paper addresses the problems of noise estimation algorithm for front-end distributed speech recognition (DSR) system. We propose the method, where Mel-scale filter bank is applied for purpose of spectral smoothing. The selection of noise estimation algorithm is obviously influenced by the following noise suppression system. In our experiments, the noise suppression algorithm has been a-priori given. The memory size limitation have played important role in our noise estimation experiments.

English abstract

The paper addresses the problems of noise estimation algorithm for front-end distributed speech recognition (DSR) system. We propose the method, where Mel-scale filter bank is applied for purpose of spectral smoothing. The selection of noise estimation algorithm is obviously influenced by the following noise suppression system. In our experiments, the noise suppression algorithm has been a-priori given. The memory size limitation have played important role in our noise estimation experiments.

Keywords

noise estimation, distributed speech recognition

RIV year

2002

Released

10.05.2002

Publisher

Slovak University of Technology in Bratislava

Location

Bratislava

ISBN

80-227-1700-2

Book

12th International Czech-Slovak Scientific conference Radioelektronika 2002

Pages from

1

Pages to

4

Pages count

4

URL

Documents

BibTex


@inproceedings{BUT10275,
  author="Petr {Motlíček}",
  title="Application of Mel-scale Filter bank for Noise Estimation in Speech Processing",
  annote="The paper addresses the problems of noise estimation algorithm
for front-end distributed speech recognition (DSR) system. We propose
the method, where Mel-scale filter bank is applied for purpose of
spectral smoothing. The selection of noise
estimation algorithm is obviously influenced by the following noise
suppression system. In our experiments, the noise suppression
algorithm has been a-priori given. The memory size limitation  have
played important role in our noise estimation experiments.",
  address="Slovak University of Technology in Bratislava",
  booktitle="12th International Czech-Slovak Scientific conference Radioelektronika 2002",
  chapter="10275",
  institution="Slovak University of Technology in Bratislava",
  year="2002",
  month="may",
  pages="1",
  publisher="Slovak University of Technology in Bratislava",
  type="conference paper"
}