• DocumentCode
    3391261
  • Title

    Analysis and enhancement of locally harmonic signals using adaptive multi-kernel methods

  • Author

    Dubost, Stéphanie ; Cappe, Olivier

  • Author_Institution
    Ecole Nat. Superieure des Telecommun., Paris, France
  • fYear
    1999
  • fDate
    1999
  • Firstpage
    163
  • Lastpage
    166
  • Abstract
    This paper is concerned with the analysis/synthesis and enhancement of signals that can be efficiently modeled as “quasi-harmonic”. We first provide a formal definition of quasi-harmonic signals and show that known estimation methods for such signals can be interpreted in a non-parametric local approximation framework. We then investigate the possibility of using “adaptive multi-kernel” estimation methods for which the duration of the analysis frame is tuned independently for each harmonic depending on the local behavior of the signal
  • Keywords
    Gaussian noise; adaptive estimation; adaptive signal processing; harmonic analysis; polynomial approximation; signal synthesis; speech processing; white noise; Gaussian white noise; adaptive estimation methods; adaptive multi-kernel methods; formal definition; local behavior; local polynomial approximation; locally harmonic signals analysis; locally harmonic signals enhancement; locally harmonic signals synthesis; nonparametric local approximation; quasi-harmonic signals; voiced speech; Frequency; Harmonic analysis; Integrated circuit modeling; Interpolation; Noise reduction; Parameter estimation; Signal analysis; Signal processing; Signal synthesis; Speech synthesis;
  • fLanguage
    English
  • Publisher
    ieee
  • Conference_Titel
    Applications of Signal Processing to Audio and Acoustics, 1999 IEEE Workshop on
  • Conference_Location
    New Paltz, NY
  • Print_ISBN
    0-7803-5612-8
  • Type

    conf

  • DOI
    10.1109/ASPAA.1999.810875
  • Filename
    810875