Presentation 2014-12-18
Some Properties of Iterative Wiener Filter for Speech Enhancement
Emiko ABE, Tetsuya SHIMAMURA,
PDF Download Page PDF download Page Link
Abstract(in Japanese) (See Japanese page)
Abstract(in English) In this paper, we show some properties of iterative Wiener filter for speech enhancement with SNR (Signal-to-Noise Ratio) changed by increasing the number of iterations. The iterative spectrum subtraction, which is one of the typical techniques of speech enhancement, is used for comparison. Additionally, we make a comparison in the case of changing the estimation method of the power spectrum, and in the case of using different iterative methods in iterative Wiener filter. We evaluated the accuracy in SNR for each iteration. It was found that the iterative Wiener filter provides the best performance at an early stage of the iterations. This means that it is not necessary to continue to increase the number of iterations. On the other hand, it was found that in order to obtain a high SNR in the iterative spectral subtraction method, it is necessary to continue to increase the number of iterations compared with the iterative Wiener filter. However, it is difficult to determine the appropriate number of iterations. We showed theoretically the difference obtaind in SNR for iterative Wiener filter and iterative spectral subtraction method.
Keyword(in Japanese) (See Japanese page)
Keyword(in English) speech enhancement / iterative Wiener filter / iterative spectral subtraction / SNR / LPC analysis
Paper # SIS2014-71
Date of Issue

Conference Information
Committee SIS
Conference Date 2014/12/11(1days)
Place (in Japanese) (See Japanese page)
Place (in English)
Topics (in Japanese) (See Japanese page)
Topics (in English)
Chair
Vice Chair
Secretary
Assistant

Paper Information
Registration To Smart Info-Media Systems (SIS)
Language JPN
Title (in Japanese) (See Japanese page)
Sub Title (in Japanese) (See Japanese page)
Title (in English) Some Properties of Iterative Wiener Filter for Speech Enhancement
Sub Title (in English)
Keyword(1) speech enhancement
Keyword(2) iterative Wiener filter
Keyword(3) iterative spectral subtraction
Keyword(4) SNR
Keyword(5) LPC analysis
1st Author's Name Emiko ABE
1st Author's Affiliation Faculty of Engineering, Saitama University()
2nd Author's Name Tetsuya SHIMAMURA
2nd Author's Affiliation Faculty of Engineering, Saitama University
Date 2014-12-18
Paper # SIS2014-71
Volume (vol) vol.114
Number (no) 370
Page pp.pp.-
#Pages 6
Date of Issue