University of Southampton
Signal processing strategies to improve low-frequency performance of multi-channel systems in cuboid reverberant spaces.
Abstract
dc:description.abstractThis thesis addresses problems in the reproduction of low-frequency audio, in home listening environments, using digital signal processing. Focusing on cuboid rooms, it covers two primary areas. Firstly, controlling the sound field within an enclosure through a least-squares approach to filter design, secondly, predicting this field from a minimal number of acoustic measurements. It concludes that filters designed using a four-channel least-squares approach, based on pseudo-real-world data, and created using the proposed enhanced analytical model, show a measurable improvement in the real world. A literature review is presented of the physical properties of sound within an enclosure, various previous approaches addressing low-frequency issues, and methods for understanding and predicting the sound field. From this review, it is clear that the proposed approach for predicting the sound field within an enclosure is unique. The plane wave cancellation method offers a good solution but with limitations, and this is very thoroughly examined. The target is clearly to create a system objectively as effective as the plane wave cancellation method but without its limitations. The least-squares filter approach is introduced, and shown through numerical simulations to outperform the plane wave cancellation method, without its limitations. It is also shown that the effectiveness of the least-squares filter approach seen in numerical simulations, transfers to real-world data. Finally, methods are proposed for extracting room dimensions and source and receiver locations, from the frequency analysis of a small number of real-world acoustic measurements. The results are combined with Green’s function, to create an enhanced analytical model, that can give pseudo-real-world measurements, accurate enough to replace the need for real-world measurements, when designing least-square filters
Degree
thesis:*- Name dc:type.qualificationname
- Ph.D.
- Level dc:type.qualificationlevel
- doctoral
- Grantor dc:publisher.institution
- University of Southampton
- Year dc:date.issued
- 2023
Author and committee
dc:creator, dc:contributor.*- Author dc:creator
-
- Bell, Thomas Oliver
- Advisor dc:contributor.advisor
-
- Fazi, Filippo
Rights
- Language dc:language
- English, eng