Audio analysis of statistically instantaneous signals with mixed Gaussian probability distributions

Ganesh R. Naik, Wenwu Wang

Research output: Contribution to journalArticlepeer-review

14 Citations (Scopus)

Abstract

In this article, a novel method is proposed to measure the separation qualities of statistically instantaneous audio signals with mixed Gaussian probability distributions. This study evaluates the impact of the Probability Distribution Function (PDF) of the mixed signals on the outcomes of both sub-and super-Gaussian distributions. Different Gaussian measures are evaluated by using various spectral-distortion measures. It aims to compare the different audio mixtures from both super-Gaussian and sub-Gaussian perspectives. Extensive computer simulation confirms that the separated sources always have super-Gaussian characteristics irrespective of the PDF of the signals or mixtures. The result based on the objective measures demonstrates the effectiveness of source separation in improving the quality of the separated audio sources.

Original languageEnglish
Pages (from-to)1333-1350
Number of pages18
JournalInternational Journal of Electronics
Volume99
Issue number10
DOIs
Publication statusPublished - Oct 2012
Externally publishedYes

Keywords

  • blind source separation
  • independent component analysis
  • kurtosis
  • probability distribution function
  • signal to interference ratio
  • sub-Gaussian
  • super-Gaussian

Fingerprint Dive into the research topics of 'Audio analysis of statistically instantaneous signals with mixed Gaussian probability distributions'. Together they form a unique fingerprint.

Cite this