A new class of adaptive algorithm based on correlation functions for double-talk acoustic echo cancelling

  • Authors:

  • Mohammad Reza Asharif;Rui Chen;Katsumi Yamashita

  • Affiliations:

  • Department of Information Engineering, Faculty of Engineering, University of the Ryukyus, Okinawa, Japan and Department of Electrical and Computer Engineering, Faculty of Engineering, University o ...;Department of Information Engineering, Faculty of Engineering, University of the Ryukyus, Okinawa, Japan;Department of Electrical and Electronic Systems, Graduate School of Engineering, Osaka Prefecture University, Osaka, Japan

  • Venue:
  • Mathematics and Computers in Simulation
  • Year:
  • 2004

Quantified Score

Hi-index 0.00

Visualization

Abstract

Echo path estimation in echo cancelling for teleconference system is a problem in double talk condition. The correlation processing algorithm was defined by the authors to solve this problem. In this paper, a new fast implementation in the frequency domain with correlation initialization and step size normalization is introduced. The new frequency-domain ECLMS (FECLMS) algorithm requires only 2.7% of the ECLMS computational loads. The new algorithm is superior as compared with conventional one in double-talk with speech signal.