Detail publikačního výsledku

Separation of Two Delayed Acoustic Sources Using FIR Filters

EKSLER, V.

Originální název

Separation of Two Delayed Acoustic Sources Using FIR Filters

Anglický název

Separation of Two Delayed Acoustic Sources Using FIR Filters

Druh

Stať ve sborníku v databázi WoS či Scopus

Originální abstrakt

We present the advanced natural gradient adaptation method for adapting coefficients of inverse filters for blind source separation. In this paper an advanced method of batch learning algorithm with convergence measurement using kurtosis is discussed for the situation when two sources are delayed with respect to each other. We are interested in the two-channel case, in which we observe the outputs of a 2 x 2 linear time-invariant system. Real experimental results of acoustic source tests are shown and the separation of speech and music signal is present here to validate the algorithm functionality.

Anglický abstrakt

We present the advanced natural gradient adaptation method for adapting coefficients of inverse filters for blind source separation. In this paper an advanced method of batch learning algorithm with convergence measurement using kurtosis is discussed for the situation when two sources are delayed with respect to each other. We are interested in the two-channel case, in which we observe the outputs of a 2 x 2 linear time-invariant system. Real experimental results of acoustic source tests are shown and the separation of speech and music signal is present here to validate the algorithm functionality.

Klíčová slova v angličtině

Blind Source Separation, Natural Gradient, Speech Enhancement, Kurtosis

Autoři

EKSLER, V.

Vydáno

29.06.2005

Nakladatel

STU Bratislava

Místo

Bratislava, Slovak Republic

ISBN

80-227-2257-X

Kniha

Proceedings EC-SIP-M 2005

Strany od

280

Strany počet

6

BibTex

@inproceedings{BUT14793,
  author="Václav {Eksler}",
  title="Separation of Two Delayed Acoustic Sources Using FIR Filters",
  booktitle="Proceedings EC-SIP-M 2005",
  year="2005",
  pages="6",
  publisher="STU Bratislava",
  address="Bratislava, Slovak Republic",
  isbn="80-227-2257-X"
}