Spatial Room Impulse Response Estimation from a Moving Microphone Array

Paper i proceeding, 2025

Estimating spatial room impulse responses (SRIRs) is essential for rendering virtual sound sources that realistically blend into a real-world environment in applications like augmented reality. Head-worn devices, such as smart glasses, are pivotal to enabling augmented reality experiences, but their continuous movement poses a unique challenge for the estimation of SRIRs. This study presents a method for estimating an SRIR representing the acoustic environment at a fixed reference point using signals recorded by a moving microphone array near that position. By leveraging the positional data of the microphone array, the proposed recursive least squares algorithm updates a signal model that describes the recorded sound pressure as a combination of the source signal and the SRIR at the reference point, transformed through rotation and translation in the circular harmonic domain. Simulation and measurement results demonstrate that the estimated SRIRs achieve high accuracy in the low and mid-frequency ranges -- exceeding the spatial aliasing frequency of the array -- while the translation distance and the order of the circular harmonic coefficients limit performance in the high-frequency range.