Optimising Metasurface Retrieval by Hill Descending

Abstract

Delivering a uniform acoustic experience in a space is a well-known challenge. Today, this is achieved stacking many digitally controlled wavefront shape loudspeakers. However, since the directivity of each speaker varies with the frequency, the sound waves do not mate consistently, resulting in a non-uniform sound pressure level, changing with both the frequency and the position of the listener.Here we show how this can be addressed using acoustic metasurfaces, capable of locally manipulating both phase and intensity of a sound wave in a small footprint by changing their geometrical parameters. When coupled at the outlet of a sound source, they could create meta-filters, passively embedding part of the required signal processing, creating sound homogeneity or providing sound effects. The question is: given a desired wavefront, how can the parameters be tuned efficiently? In this study, a method for retrieving an acoustic metasurface through the hill descending algorithm is presented. The algorithm is tested on a simple optimization problem, searching for the best metamaterials combination providing a desired coverage in a reflection-free listening area. Input data are retrieved by finite-element modelling a simplified sound system in a reflection-free venue. Further research is required to validate this method toward multi-objective optimisation.