Modal based solutions for the acquisition and rendering of large spatial soundfields

Abstract

This thesis explores the design and analysis of acoustic signal processing algorithms for the acquisition and rendering of spatial soundfields over large regions. The principal function of spatial sound systems is to record and reproduce desired soundfields in real time. Although there exist numerous approaches to the above problem, the development of immersive audio over large spatial regions is impeded by several fundamental and technological limitations. Therefore, there is a growing interest in deriving new signal processing algorithms for the recreation of large acoustic events, along with a wide range of applications envisioned for the future. Our approach is to study the modal based solution to the classical wave equation, and exploit its intrinsic properties to develop new recording/reproduction techniques. Novel designs for microphone/loudspeaker arrays are derived where each unit records/reproduces lower order spatial sound samples of the larger soundfield of interest. These samples are collectively combined to generate the final outcome using a set of modal analysis tools referred to as the addition theorems. An efficient model of the underlying structure of reverberation is also incorporated to improve the fidelity of sound reproduction. Several outcomes resulting from this approach are: (i) derivation of a novel microphone array design for spatial sound recording with a substantial reduction in microphone numbers, (ii) derivation of a novel loudspeaker array design for spatial sound reproduction with a substantial reduction in loudspeaker numbers and, (iii) a new approach to acoustic equalization based upon an efficient parameterization of the room transfer function. Compared to the conventional approaches to sound recording, the proposed microphone design is shown to be more robust to measurement noise. Likewise, compared to the existing surround sound systems, the proposed loudspeaker design is shown to be highly advantageous with the added ability to cancel the outgoing soundfield. Application and performance of the above methods are illustrated using appropriate simulation examples. Show more