• DocumentCode
    179723
  • Title

    Frequency domain acoustic echo canceller that handles asynchronous A/D and D/A clocks

  • Author

    Abe, Makoto ; Nishiguchi, Masayuki

  • Author_Institution
    SONY Corp., Tokyo, Japan
  • fYear
    2014
  • fDate
    4-9 May 2014
  • Firstpage
    5924
  • Lastpage
    5928
  • Abstract
    An implicit premise in using an acoustic echo canceller (AEC) is that the clock for A/D conversion for a microphone and the clock for D/A conversion for a loudspeaker work synchronously. Even a slight difference in sampling rate between the clocks critically degrades the echo cancelling performance. This paper describes a method of making an AEC in the frequency domain that can handle a mismatch in sampling rate between A/D and D/A conversion. The method recursively estimates the sampling-rate offset by a simple extension of the well-known LMS algorithm, and corrects it through two mechanisms, frame-step control and phase rotation, which obviate the need for any explicit resampling operation. Experimental results show that this method provides an echo suppression level comparable to a standard AEC without mismatch for an offset of up to 1000 ppm.
  • Keywords
    acoustic signal processing; analogue-digital conversion; clocks; digital-analogue conversion; echo suppression; least mean squares methods; loudspeakers; signal sampling; analog-digital clock; analog-digital conversion; asynchronous clock; digital-analog clock; digital-analog conversion; explicit resampling operation; frame step control; frequency domain acoustic echo canceller; least mean square algorithm; loudspeaker; phase rotation; sampling rate offset estimation; Acoustics; Clocks; Estimation; Frequency-domain analysis; Least squares approximations; Microphones; Standards; acoustic echo canceller; asynchronous clocks; sampling-rate mismatch; short-term Fourier transform; subband processing;
  • fLanguage
    English
  • Publisher
    ieee
  • Conference_Titel
    Acoustics, Speech and Signal Processing (ICASSP), 2014 IEEE International Conference on
  • Conference_Location
    Florence
  • Type

    conf

  • DOI
    10.1109/ICASSP.2014.6854740
  • Filename
    6854740