適應性3D音效投影屏幕與藍芽耳機聲學最佳化設計

Abstract

面對耳機設計時各種相互衝突的特性需求，如靈敏度、失真率、頻寬以及微小化等，需要一種有系統且有效率的方法來達成我們的設計並符合設計需求。 利用微型揚聲器的參數鑒別法、電機聲模擬電路 (EMA)，建立藍芽耳機的聲學模擬平台。並且以此平台為基礎，利用模擬退火法 (simulated annealing)進行耳機腔體的最佳化設計，並能有效達到期望目標。實驗方面，利用Type 2人工耳（IEC 711 ear simulator）來得到實驗結果，並且顯示出經最佳化後的耳機腔體設計，能夠大大的提升其聲學品質，並且能夠符合3GPP2的規範。 一般而言，利用串音消除系統（cross talk cancellation system）為基礎的3D虛擬環繞音效，皆會產生甜點（sweet spot）的問題，也就是當聽者超出甜點的範圍，就無法聽到良好的音效。本文中提出一個適應性3D虛擬聲束操控（beam steering）系統，可使聽者利用雙聲道喇叭，即可在任意聆聽位置感受到絕佳的環繞音效。本系統包含了兩種技術，其中之一是利用攝影機做辨識的頭部追蹤系統，另外是聲束控制演算法，來進行聆聽位置的最佳化設定。利用適應性3D虛擬聲束操控系統，一個全新的平面喇叭投影屏幕誕生，並且可以使聆聽者在任何位置享受最佳的3D環繞音效。

Faced with various conflicting issues arising from sensitivity, distortion, bandwidth and miniaturization requirements for Bluetooth® earphones, it is desirable to develop a systematic way to attain the design that would meet these requirements. Based on a lumped parameter model, the design of the earphone enclosure is optimized using the simulated annealing (SA) algorithm. Experimental results obtained using a type 2 artificial ear (IEC 711 ear simulator) reveals that the optimized design has resulted in significant enhancement of performance, complying with the frequency response mask dictated in the 3GPP2 standard. This thesis proposes an adaptive three-dimensional (3D)-sound projection screen aimed at home theater applications. The screen serves both audio and visual rendering system that integrates numerous technologies including panel speakers, cross talk cancellation system (CCS) and video head tracking. In particular, the CCS enhances the localization impression when rendering 3D spatial sound with two loudspeakers. A camera is employed to adaptively track the head movement such that the robustness of CCS is ensured. The algorithms were implemented on a laptop computer in tandem with a digital signal processor, which handled the video and audio signal processing, respectively. Experimental results reveal that the proposed system is capable of rendering immersive spatial sound with robustness against listener’s head movement.