• DocumentCode
    2801684
  • Title

    Acoustic echo cancellation with post-filtering in subband

  • Author

    Lu, Xiaojian ; Champagne, Benoit

  • Author_Institution
    Dept. of Electr. & Comput. Eng., McGill Univ., Montreal, Que., Canada
  • fYear
    2003
  • fDate
    19-22 Oct. 2003
  • Firstpage
    29
  • Lastpage
    32
  • Abstract
    A new structure is proposed for an acoustic echo cancellation (AEC) system which explores post-filtering techniques in the subband. The post-filtering approach used seeks to minimize near-end speech distortion while further suppressing the residual echo to a predefined level. The new AEC system seamlessly combines this post-filtering technique and adaptive filters in a computationally efficient subband structure based on oversampled uniform DFT banks. Compared to its full-band counterpart, the proposed subband algorithm has a much lower computational complexity while achieving a comparable performance in terms of acoustic echo suppression. These observations are supported by experimental results. An improved practical design of the prototype filter for use with DFT filter banks is also introduced.
  • Keywords
    acoustic signal processing; adaptive filters; channel bank filters; computational complexity; discrete Fourier transforms; echo suppression; speech processing; AEC; DFT filter banks; acoustic echo cancellation; acoustic echo suppression; adaptive filters; computational complexity; near-end speech distortion minimization; oversampled uniform DFT banks; subband post-filtering; Acoustic distortion; Acoustical engineering; Adaptive filters; Computational complexity; Echo cancellers; Filter bank; Frequency domain analysis; Nonlinear filters; Prototypes; Speech;
  • fLanguage
    English
  • Publisher
    ieee
  • Conference_Titel
    Applications of Signal Processing to Audio and Acoustics, 2003 IEEE Workshop on.
  • Print_ISBN
    0-7803-7850-4
  • Type

    conf

  • DOI
    10.1109/ASPAA.2003.1285801
  • Filename
    1285801