Presentation 1997/9/11
Noise Reduction of LSP parameters : Comparison between Neural Networks and K-L transform
Mitsuharu Hasumi, Tetsuo Funada, Yoshinori Morita,
PDF Download Page PDF download Page Link
Abstract(in Japanese) (See Japanese page)
Abstract(in English) In this paper, we investigate to reduce noise of LSP parameter with Neural Networks trained by non-noise LSP parameter. We quantize a vector of LSP parameters by scalar-quantizing the output of each unit of hidden layer and we find an appropriate number of scalar quantization bits to reduce noise. We investigate performance dependency on the number of middle layer units and compare between three-layer and five-layer Neural Networks. We also investigate to reduce noise with K-L transform and compare the performance between K-L transform and Neural Networks.
Keyword(in Japanese) (See Japanese page)
Keyword(in English) LSP parameter / noise reduction / neural networks / K-L transform
Paper # SP97-42
Date of Issue

Conference Information
Committee SP
Conference Date 1997/9/11(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 Speech (SP)
Language JPN
Title (in Japanese) (See Japanese page)
Sub Title (in Japanese) (See Japanese page)
Title (in English) Noise Reduction of LSP parameters : Comparison between Neural Networks and K-L transform
Sub Title (in English)
Keyword(1) LSP parameter
Keyword(2) noise reduction
Keyword(3) neural networks
Keyword(4) K-L transform
1st Author's Name Mitsuharu Hasumi
1st Author's Affiliation Faculty of Engineering, Kanazawa University()
2nd Author's Name Tetsuo Funada
2nd Author's Affiliation Faculty of Engineering, Kanazawa University
3rd Author's Name Yoshinori Morita
3rd Author's Affiliation Ishikawa National College of Technology
Date 1997/9/11
Paper # SP97-42
Volume (vol) vol.97
Number (no) 248
Page pp.pp.-
#Pages 8
Date of Issue