Time-varying autoregressive modeling of audio and speech signals

Aki Harma; Marko Juntunen; Jari P. Kaipio

Time-varying autoregressive modeling of audio and speech signals

Aki Harma, Marko Juntunen, Jari P. Kaipio

Research output: Chapter in Book/Report/Conference proceeding › Conference article in proceeding › Academic › peer-review

Abstract

Conventional linear predictive techniques for modeling of speech and audio signals are based on an assumption that a signal is stationary within each analysis frame. However, natural signals are often continuously time-varying, i.e., nonstationary. Therefore this assumption might not be well justified. In this paper, we study a time-varying autoregressive (TVAR) modeling technique in which this restriction is relaxed. A frequency-warped formulation of the Subba Rao-Liporace TVAR algorithm is introduced in the article. The applicability of the presented methodology to various speech and audio signal processing tasks is illustrated and discussed. It is also shown that the TVAR scheme yields an efficient parametrization for time-varying sounds.

Original language	English
Title of host publication	European Signal Processing Conference - proceedings
Subtitle of host publication	10th European Signal Processing Conference, EUSIPCO 2000
Volume	2015-March
Edition	March
Publication status	Published - 31 Mar 2000
Externally published	Yes
Event	10th European Signal Processing Conference - Tampere, Finland Duration: 4 Sept 2000 → 8 Sept 2000 Conference number: 10

Publication series

Series	European Signal Processing Conference
ISSN	2219-5491

Conference

Conference	10th European Signal Processing Conference
Abbreviated title	EUSIPCO 2000
Country/Territory	Finland
City	Tampere
Period	4/09/00 → 8/09/00

Cite this

@inproceedings{d24b97c9b3ed40e4850c079e2d77bc22,

title = "Time-varying autoregressive modeling of audio and speech signals",

abstract = "Conventional linear predictive techniques for modeling of speech and audio signals are based on an assumption that a signal is stationary within each analysis frame. However, natural signals are often continuously time-varying, i.e., nonstationary. Therefore this assumption might not be well justified. In this paper, we study a time-varying autoregressive (TVAR) modeling technique in which this restriction is relaxed. A frequency-warped formulation of the Subba Rao-Liporace TVAR algorithm is introduced in the article. The applicability of the presented methodology to various speech and audio signal processing tasks is illustrated and discussed. It is also shown that the TVAR scheme yields an efficient parametrization for time-varying sounds.",

author = "Aki Harma and Marko Juntunen and Kaipio, {Jari P.}",

note = "Publisher Copyright: {\textcopyright} 2000 EUSIPCO.; 10th European Signal Processing Conference, EUSIPCO 2000 ; Conference date: 04-09-2000 Through 08-09-2000",

year = "2000",

month = mar,

day = "31",

language = "English",

volume = "2015-March",

series = "European Signal Processing Conference",

publisher = "European Signal Processing Conference, EUSIPCO",

booktitle = "European Signal Processing Conference - proceedings",

edition = "March",

}

Time-varying autoregressive modeling of audio and speech signals. / Harma, Aki; Juntunen, Marko; Kaipio, Jari P.
European Signal Processing Conference - proceedings: 10th European Signal Processing Conference, EUSIPCO 2000. Vol. 2015-March March. ed. 2000. 7075230 (European Signal Processing Conference).

Research output: Chapter in Book/Report/Conference proceeding › Conference article in proceeding › Academic › peer-review

TY - GEN

T1 - Time-varying autoregressive modeling of audio and speech signals

AU - Harma, Aki

AU - Juntunen, Marko

AU - Kaipio, Jari P.

N1 - Conference code: 10

PY - 2000/3/31

Y1 - 2000/3/31

N2 - Conventional linear predictive techniques for modeling of speech and audio signals are based on an assumption that a signal is stationary within each analysis frame. However, natural signals are often continuously time-varying, i.e., nonstationary. Therefore this assumption might not be well justified. In this paper, we study a time-varying autoregressive (TVAR) modeling technique in which this restriction is relaxed. A frequency-warped formulation of the Subba Rao-Liporace TVAR algorithm is introduced in the article. The applicability of the presented methodology to various speech and audio signal processing tasks is illustrated and discussed. It is also shown that the TVAR scheme yields an efficient parametrization for time-varying sounds.

AB - Conventional linear predictive techniques for modeling of speech and audio signals are based on an assumption that a signal is stationary within each analysis frame. However, natural signals are often continuously time-varying, i.e., nonstationary. Therefore this assumption might not be well justified. In this paper, we study a time-varying autoregressive (TVAR) modeling technique in which this restriction is relaxed. A frequency-warped formulation of the Subba Rao-Liporace TVAR algorithm is introduced in the article. The applicability of the presented methodology to various speech and audio signal processing tasks is illustrated and discussed. It is also shown that the TVAR scheme yields an efficient parametrization for time-varying sounds.

UR - http://www.scopus.com/inward/record.url?scp=84937060806&partnerID=8YFLogxK

M3 - Conference article in proceeding

AN - SCOPUS:84937060806

VL - 2015-March

T3 - European Signal Processing Conference

BT - European Signal Processing Conference - proceedings

T2 - 10th European Signal Processing Conference

Y2 - 4 September 2000 through 8 September 2000

ER -

Time-varying autoregressive modeling of audio and speech signals

Abstract

Publication series

Conference

Other files and links

Fingerprint

Cite this