本研究介紹整合式空間聲場錄音及回放陣列系統應用於迴響環境的情況。以原有二維技術與架構為基礎，將此陣列系統拓展至三維陣列錄音及回放系統。三維陣列錄音及回放系統運作包含兩個階段：在第一個階段，利用混合式麥克風陣列分析將迴響聲場「編碼」，將聲場簡化為聲源方向以及訊號兩個物理量來表示，對於聲源的定位可以透過NT和MPDR、MUSIC兩種波束成型的方法來測定平面波的方向。聲源分離使用TIKR於聲源符合稀少性的條件下進行編碼。本研究中亦提出一項對於聲源反算TIKR正規化參數的選取方法。第二階段則是利用三層式喇叭陣列將第一階段中獲得的聲場資訊，聲源方位、訊號大小進行解碼以及重現。透過在三維空間的重建範圍內設置大量的假想控制點，來對欲重建的聲場做取樣，進而由這些控制點來反算出喇叭訊號。於真實房間中重建聲場時，使用兩種方法進行喇叭訊號的計算，一個方法是使用自由空間聲場反算，另一個方法則是使用房間環境反算。最後，利用定位和聆聽實驗進行整合式聲學陣列系統的驗證。

This research proposes the application of an integrated recording-rendering array system in a reverberation room. Based on the existing two-dimensional technology and architecture, the array system is extended to the hybrid microphone array (HMA) of recording and 3-layed loudspeaker array playback system. The three-dimensional array architecture of the system consists of two stages. In the sound field analysis stage (SFA), the hybrid microphone array is used to analyze the sound field by the plane-wave decomposition. This microphone array measures the direction of the plane wave through one-stage newton’s algorithm (NT), two-stage minimum power distortion-response (MPDR) beamforming and multiple signal classification (MUSIC) algorithms. TIKR and NT algorithms are utilized to separate the sound sources which the number of sound sources satisfy the sparsity assumption. On the other hand, for a real space environment where the number of sound sources does not satisfy the sparsity, we used the hybrid microphone array to record the signals and then encoded into many equally spaced signals. In this study, we also propose a two steps procedure to select the regularization parameters in TIKR algorithm. The second stage is the use of 3-layed loudspeaker array to decode and reproduce the sound field information obtained in the first stage, including the sound source directions and amplitudes. We reconstructed the three-dimensional space by setting a large number of hypothetical control points and reconstructed sound field to do sampling. Finally these control points were used to compute the loudspeaker driving signals. There are two methods use in the real room to reconstruct the sound field, freefield inversion and room response inversion. The integrated acoustic array system is validated with localization and listening tests.

摘要 i
ABSTRACT ii
誌謝 iii
CONTENTS iv
LIST OF FIGURES vi
LIST OF TABLES viii
Chapter 1 INTRODUCTION 1
Chapter 2 SOUND FIELD ANALYSIS (SFA) 4
2.1 Source localization and separation algorithms 4
2.1.1 The one-stage NT algorithm 4
2.1.2 Two-Stage MPDR and TIKR Algorithm 5
2.1.3 Two-Stage MUSIC and TIKR Algorithm 6
2.1.4 The choice of regularization parameter in TIKR algorithm 7
2.2 Hybrid microphone array design 9
2.2.1 Formulation of hybrid microphone array 9
Chapter 3 SOUND FIELD SYNTHESIS (SFS) 12
3.1 Room response measurement and interpolation 12
3.2 SFS by field matching 14
3.3 The arrangement of 3-layered loudspeaker array 15
Chapter 4 SIMULATIONS AND EXPERIMENTS 17
4.1 Numerical Simulations 17
4.1.1 SFA Simulations 17
4.1.2 SFS Simulations 18
4.2 Experimental Investigations 19
4.2.1 SFA Experiment 19
4.2.2 SFS Experiment 20
Chapter 5 CONCLUSIONS 37
REFERENCES 38

[1]M. R. Bai and C. H. Kuo, “Acoustic Source Localization and Deconvolution-Based Separation,” J. Acoust. Soc. Am., 135(4), 2358 (2014).
[2]M. R. Bai, Y. H. Hua, C. H. Kuo and Y. H. Hsieh, “An Integrated Analysis- Synthesis Array System for Spatial Sound Fields,” J. Acoust. Soc. Am., 137(23), 1366-1376 (2015).
[3]M. R. Bai, H. Hsu and J.C. Wen, “Spatial sound field synthesis and upmixing based on the equivalent source method,” J. Acoust. Soc. Am., 135(1), 269-282 (2014).
[4]M. Miyoshi and Y. Kaneda, “Inverse filtering of room acoustics,” Proc. IEEE 36(2), 145-152 (1988).
[5]Ina Kodrasi, Stefan Goetze and Simon Doclo, “Regularization for parital multichannel equalization for speech dereverberation,” Proc. IEEE 21(9), 1879-1890 (2013).
[6]M. A. Gerzon, “Ambisonics in Multichannel Broadcasting and Video,” J. Audio Eng. Soc., 33(11), 859-871(1985).
[7]T. Sporer, “Wave Field Synthesis – Generation and Reproduction of Natural Sound Environments”, in Proc. of the 7th Int. Conf. on Digital Audio Effects, 133-138 (2004).
[8]S. Spors, R. Rabenstein and J. Ahrens, “The Theory of Wave Field Synthesis Revisited”, Audio Eng. Soc., Conv. 124, (2008).
[9]D. de Vries, “AES Monograph: Wave Field Synthesis,” Audio Eng. Soc. 1-95 (2009), New York: Springer.
[10]F. M. Fazi, “Sound Field Reproduction,” Ph.D. thesis, Science and Mathematics, University of Southampton, UK (2010).
[11]L. Romoli, P. Peretti, S. Cecchi, L. Palestini and F. Piazza, “Real-Time Implementation of Wave Field Synthesis for Sound Reproduction Systems,” IEEE Asia Pacific Conf., Macao, 430-433 (2008).
[12]S. Enomoto, Y. Ikeda, S. Ise, and S. Nakamura, “Three-dimentional sound field reproduction and recording systems based on boundary surface control principle,” Proceedings of the 14th International Conf. on Auditory Display (2008).
[13]Y. Li, “Time-Space Analysis, Synthesis, and Room Response Modeling of Virtual-Reality Sound Fields Using Microphone and Loudspeaker Arrays,” M.S. thesis, Department of Power Mechanical Engineering, National Tsing Hua University (2016).
[14]M. R. Bai, J. G. Ih, and J. Benesty Acoustic Array Systems: Theory, Implementation and Application. Wiley/IEEE Press, Singapore, 177-270 (2013).
[15]R. Wait, The numerical Solution of Algebraic Equations, John Wiley & Sons, 1979.
[16]C.W. Groetsch, “The theory of Tikhonov regularization for Fredholm equations of the first kind,” Pitman Advanced Pub. Program, Boston (1984).
[17]J. B. Allen and D. A. Berkley, “Image Method for Efficiently Simulating Small-Room Acoustics,” J. Acoust. Soc. Am., 65(4), 943-950(1979).
[18]M. Grant and S. Boyd, cvx, v.1.21 MATLAB software for disciplined convex programming available at http://cvxr.com/cvx .
[19]C. Chung, “Iterative Compressive Sensing (CS) algorithms for solving acoustic source characterization problems with point or block sparsity constraints,” M.S. thesis, Department of Power Mechanical Engineering, National Tsing Hua University, (2017).
[20]ITU-T Recommendation, “Perceptual evaluation of speech quality (PESQ), An Objective Method for End-to-End Speech Quality Assessment of Narrow-Band Telephone Networks and Speech Codecs,” ITU-T Recommendation 862 (2001).
[21]R. P. Brent, “Algorithms for Minimization without Derivatives,” Prentice-Hall, Inc., Englewood Cliffs, New Jersey, 48-75 (1973).
[22]P. C. Hansen, “Analysis of Discrete Ill-Posed Problems by Means of the L-curve”, Society for Industrial and Applied Mathematics (1992).
[23]P. C. Hansen and D. P. O’Leary, “The use of the l-curve in the regularization of discrete ill-posed problems”, Society for Industrial and Applied Mathematics (1993).