Author (s)
Zoboli Enrico
Affiliation
Politecnico di Milano
Publication date
2022
Abstract
Emerging spatial audio technologies require immersive acoustic and audio quality. Microphone arrays are key factors in the spatial audio ecosystem, being able to capture spatial features of acoustic sound fields. Over the years, the proposed solutions resulted to be limited in spatial performance for upcoming applications. Advanced and affordable technologies, such as Micro Electrical Mechanical System (MEMS) and Automotive Audio Bus (A2B), constitute a promising solution for future microphone arrays. Spherical harmonics and Ambisonics are adopted as a functional framework to represent sound fields and evaluate spatial audio technologies and devices. The advancements in the microphone array field allowed the derivation of the mapping between microphone signals and spherical harmonics without the employment of analytical equations. The derived technique is based on an advanced characterization of the microphone array through impulse response measurements, leading to a matrix of Finite Impulse Response (FIR) filters to perform the mapping. This procedure allowed the employment of non-spherical three-dimensional microphone arrays. In this thesis, the development and the realization of a new three-dimensional A2B-based MEMS microphone array is presented. The target 5th Ambisonics order defined the minimum amount of microphone sensors necessary for the project. The entire system is based on triangular shaped planar microphone arrays that inspired two possible arrangement geometries: the dodecahedron and the truncated octahedron. The two arrangements have been studied through Finite Element Method (FEM) simulations, assessing their spatial performance. The truncated octahedron has shown the best results and it has been physically realized though 3D printing. The realized microphone array has been acoustically characterized through an advanced measuring technique. The measured data has been employed for deriving the necessary FIR filters to map the microphone array signal into spherical harmonics. Finally, the spatial performance has been assessed leading to the definition of acceptable frequency ranges of the realized device.
Full paper
https://www.politesi.polimi.it/handle/10589/194870
Keywords
MEMS microphone array, automotive audio bus (A2B), spherical harmonics, higher-order ambisonics, spatial audio capture, FIR filter mapping, finite element method (FEM) simulation