利用雙喇叭產生3D立體音效

Abstract

在本論文中，吾人將探討如何利用雙喇叭來提供使用者3D立體音的感受 。而主 要的問題是串音(crosstalk)的干擾。吾人將使用順向( forward type)和逆向(feedbacktype)串音消除器來減少串音。順向串音 消除器包含了least square type和二種 directforward type。在least square type中吾人將焦點放在加多少的額外延遲在 desiresignal上以改 善串音消除效能。而 Direct forward type 的設計關鍵在於使用 minimum phase part 以保證串音消除器的穩定性。在逆向串音消除器中 ，吾人提出了 stability lemma以確保逆向串音消除器的穩定性。吾人也 會針對feedback type和 direct type 2 作擾動分析。而為了有數據來分 析串音消除的效能，吾人將使用串音壓制係數(CSF)和等化係數(EQ)作為 分析的依據。 In this thesis, we investigate the method to provide the listener virtual 3D sound perception via a loudspeaker pair. The most critical problem bothering us is the crosstalk disturbation. To overcome this inevitable problem, we investigate both forward and feedback types of crosstalk cancellers. The forward type includes the least square, and two types of the direct forward type. In the least square type, we focus on the desired extra delay to achieve the causalityand the better crosstalk cancellation. In the direct forward types, the keydesigning procedure is using the mimimum phase part of the denominator.The causality is also discussed. In the feedback type, we propose a stabilitylemma to guarantee the stalibily of the feedback type crosstalkcanceller. With this lemma, we can use pole-zero modeling to find out thedesired filters. Perturbation analysis of pole-zero modeling is proposedin comparing the direct forward type 2 and the feedback type. Crosstalk suppression factor (CSF) and the equalizationfactor (EQ) are used to compare their performance numerically in this thesis.