Presentation 1996/5/16
A Method for predicting of perceived spectral sequences based on a FM-tracking model
Ikuyo MASUDA, Kiyoaki AIKAWA,
PDF Download Page PDF download Page Link
Abstract(in Japanese) (See Japanese page)
Abstract(in English) A method for extracting speech sound from interfering noises by tracking and predicting the temporal change of spectrum is presented. It is carried out by extending the FM-tracking model to spectral domain. First, spectrum is converted into a series of frequencies by using IFIS(Inverse Function of Integrated Spectrum). Next, the frequencies are tracked and predicted by using the FM-tracking model and they are converted into a spectrum again. Furthermore, phonemic restoration is replicated using this model. The strategy is that a spectrum is predicted based on trajectory by using proposed method, and real spectrum amplitude and predicted spectrum amplitude are compared, and then, an output spectrum is determined.
Keyword(in Japanese) (See Japanese page)
Keyword(in English) Phonemic restoration / Auditory scene analysis / Prediction / Signal separation / Noise reduction
Paper # SP96-2
Date of Issue

Conference Information
Committee SP
Conference Date 1996/5/16(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) A Method for predicting of perceived spectral sequences based on a FM-tracking model
Sub Title (in English)
Keyword(1) Phonemic restoration
Keyword(2) Auditory scene analysis
Keyword(3) Prediction
Keyword(4) Signal separation
Keyword(5) Noise reduction
1st Author's Name Ikuyo MASUDA
1st Author's Affiliation ATR Human Information Processing Research Laboratories()
2nd Author's Name Kiyoaki AIKAWA
2nd Author's Affiliation NTT Human Interface Research Laboratories
Date 1996/5/16
Paper # SP96-2
Volume (vol) vol.96
Number (no) 39
Page pp.pp.-
#Pages 8
Date of Issue