Pushing the Limits of Audio Spatialization with Emerging Architectures
Spatial audio refers to various systems and techniques used to diffuse sound sources or sound fields in space. While “surround sound” systems which are present in many homes typically involve 5 to 7 speakers, more advanced techniques such as Ambisonics and Wave Field Synthesis (WFS) imply the use of a much greater number of speakers. Those techniques are very efficient to recreate sound spatialization and are of high interest in fields like virtual and augmented reality, acoustic of concert room, archeoacoustics, etc. However, the number of involved speakers (limited by the computational power of the system) rarely exceeds 64, which is a critical factor for implementing more efficient spatial audio techniques limiting the quality of the simulation. The use of very powerful embedded platforms called Field Programmable Gate Arrays (FPGA) will allow us to target systems with hundreds of speakers, while providing better audio latency performances and synchronization. This work will rely on platforms and technologies currently developed as part of my thesis at INSA-Lyon and Inria (France). And it will benefit from the expertise and unique facilities of the Center for Computer Research in Music and Acoustics (CCRMA) at Stanford University.