Abstract
The principles of time-intensity panning and of using non-coincident microphone arrays for recording multichannel audio have been well established. Unlike previous studies which focus on lower-order microphones or assume ideal directivity patterns, this paper presents a practical design of higher-order non-coincident microphone arrays for multichannel recording and reproduction based on time-intensity panning. In this design, higher-order microphones are implemented using linear differential microphone arrays (LDMAs). Furthermore, a comprehensive framework for differential beamformer design and array topology optimization is developed, which minimizes the error between the actual beampattern and the ideal directivity pattern while ensuring the beamformer robustness. Simulation results demonstrate the effectiveness of the proposed microphone array and the associated differential beamformer.