立體聲迴音消除計數之研究與設計

Abstract

本論文主要是對立體聲迴音消除做整體的介紹，並探討數種適應性演算法在立體聲迴音消除之收斂速度、穩定性及運算量複雜度的表現。一般的立體聲迴音消除器若要以有限響應濾波器實現，其龐大的硬體運算量以及適應性濾波器係數的錯誤收斂是其最主要的兩個問題。為了解決這些問題，我們利用了一些現有的時域、及次頻適應性訊號處理演算法來降低兩個輸入訊號間的交互關連性，而使得適應性濾波器在收斂時能夠收斂至正確的房間響應係數。另一方面，我們也改進了一些現有的演算法，以降低其運算量。

In this thesis, the theories of stereophonic acoustic echo cancellation (SAEC) were studied. The performance in terms of convergence rate, stability and computational complexity of several adaptive algorithms applied to stereophonic acoustic echo canceller were evaluated. For implementing a stereophonic acoustic echo canceller, the heavy computational complexity and misconvergence of the adaptive filters are the main problems in SAEC system. To solve these problems, we tried to utilize several existing time-domain and subband adaptive filtering algorithms to degrade the cross-correlation between the two input signals so that the filters can converge to the real impulse response coefficients during the adaptation. On the other hand, we also tried to modify some of the algorithms in order to reduce the computational complexity.