Probabilistic time-frequency source-filter decomposition of non-stationary signals

Author

Badeau, Roland ; Plumbley, Mark D.

Author_Institution

Inst. Mines-Telecom, Telecom ParisTech, Paris, France

fYear

2013

fDate

9-13 Sept. 2013

Firstpage

1

Lastpage

5

Abstract

Probabilistic modelling of non-stationary signals in the time-frequency (TF) domain has been an active research topic recently. Various models have been proposed, notably in the nonnegative matrix factorization (NMF) literature. In this paper, we propose a new TF probabilistic model that can represent a variety of stationary and non-stationary signals, such as autoregressive moving average (ARMA) processes, uncorrelated noise, damped sinusoids, and transient signals. This model also generalizes and improves both the Itakura-Saito (IS)-NMF and high resolution (HR)-NMF models.

Keywords

autoregressive moving average processes; filtering theory; matrix decomposition; probability; time-frequency analysis; transient analysis; ARMA process; Itakura-Saito NMF model; TF probabilistic model; autoregressive moving average; damped sinusoid; high resolution NMF model; nonnegative matrix factorization; probabilistic nonstationary signal modelling; probabilistic time-frequency source filter decomposition; transient signal; uncorrelated noise; Autoregressive processes; Convolution; Hafnium; Mathematical model; Probabilistic logic; Time-domain analysis; Time-frequency analysis; Non-stationary processes; Nonnegative matrix factorisation; Probabilistic modelling; Source-filter models; Time-frequency analysis;

fLanguage

English

Publisher

ieee

Conference_Titel

Signal Processing Conference (EUSIPCO), 2013 Proceedings of the 21st European

Conference_Location

Marrakech

Type

conf

Filename

6811602