Spatial Room Impulse Response Identification from Rotating Equatorial Microphone Arrays

Spatial room impulse responses (SRIRs) facilitate the rendering of virtual sound sources to realistically augment a real-world acoustic scene, for example in the context of augmented reality. With the integration of microphone arrays into robotic systems and head-worn devices, there arises the need to identify the SRIRs from arrays amidst rotation. This work establishes the groundwork for addressing this issue through the introduction of a recursive least squares filter that adaptively identifies the SRIR from signals captured by a rotating equatorial microphone array and a given reference signal. The method not only compensates for the rotation of the array but exploits the rotation to increase the directional accuracy by updating the SRIR estimates in the space-continuous circular harmonic domain. Results from simulations and measurements show that the accuracy of the SRIR estimates from the method using a rotating array of four irregularly spaced microphones is comparable to measured SRIRs from static arrays with a substantially greater number of equally spaced microphones.