IIR 濾波器在音場回音消除的應用

Abstract

以 FIR 濾波器設計 AEC(Acoustic Echo Cancerlation)時因為所需階數 (Order) 太大,造成電路實現不易,在本論文我們以 IIR 濾波器設計 AEC 希望達到降低濾波器階數的目的.首先我們以 MLS 測量房間的脈衝響應, 提供設計 AEC 時所需之真實數據.其次討論如何選用一種方法來求得 IIR 濾波器的係數,使濾波器可以穩定,又能準確的逼近房間的脈衝響應, 我們將重點放在討論 SM Method 和它的簡化(成為 New SM Method)我們 提供詳細的簡化過程,和模擬結果.最後我們以 SM 法 ,Prony 法 與 FIR 濾波器在各種階數之下比較三者在實際 AEC 的差異,我們會指出慎重選擇 IIR Model 將在 AEC 扮演一個重要角色. AEC (Accoustic Echo Cancellation) is usually implemented by FIR filters.However, such filters require up to thousands of orders to identify the corresponding room impulse responses. In this thesis, we attempts to use IIR filters to reduce the filter order.Firstly, we measure true room impulse response data based on the MLS method.Secondly, the IIR filter coefficients are estimated using the SM method which provides an accurate and stable model for the room impulse response. our major focus is to develop a new iterative SM (NSM) method to reduce computational complexity.Finally, computer simulations are performed to compare the performances of AEC among IIR and FIR filters.