Télécom Paris

Single Sensor Source Separation Using Multiple-Window STFT Representation
  • Laurent Benaroya, (MIST Technologies)
  • Raphaël Blouet, (MIST Technologies)
  • Cédric Févotte, (MIST Technologies)
  • Israel Cohen, (Technion, department of electrical engineering)
  • Sound enhancement and sound separation
Get the paper in PDF format
Acrobat Reader (version 5 minimum) is necessary to read this document.

The aim of this paper is to investigate the use of multiple-window Short-Time Fourier Transform representation for single sensor source separation. We propose to iteratively split the observed signal into target sources and residuals. Each target source is modeled as the sum of elementary components with known power spectral densities. The decomposition into source signals and residual signals employs a confidence measure, which is based on the Fisher information matrix of the expansion coefficients. We demonstrate the improved performance of the proposed method on mixtures of voice and music signals.

©2006 Télécom Paris/TSI
Edition : Télécom Paris -- 2006