On the Statistics of Spectrogram Reassignment Vectors

  • Authors:
  • Eric Chassande-Mottin;Patrick Flandrin;Francois Auger

  • Affiliations:
  • -;Laboratoire de Physique, Ecole Normale Supérieure de Lyon, 46 allée d‘Italie, 69364 Lyon Cedex 07, France;GE44, CRTT, Bd de l‘Université, BP 406, 44602 Saint-Nazaire Cedex, France

  • Venue:
  • Multidimensional Systems and Signal Processing - Special issue on recent developments in time-frequency analysis
  • Year:
  • 1998

Quantified Score

Hi-index 0.00

Visualization

Abstract

Reassignment is a non-linear techniquewhich can improve on the localization of a spectrogram by movingits values according to a suitable vector field. Statisticalproperties of spectrogram reassignment vectors are investigatedin detail. Closed form expressions are given when the observationconsists in a non-random component embedded in white Gaussiancircular noise, and when the analysis window is Gaussian. Anextension to arbitrary windows is also proposed and theoreticalclaims are supported by numerical simulations.