Compact rigid-sphere microphone arrays are used to record 3D-sound scenes from one recording perspective. They are often used as the main microphone array in field recordings using Ambisonics. For time-invariant Ambisonic encoding, the array signals are typically first decomposed into spherical harmonic coefficient signals by a matrix, which are then equalized by radial filters, which need to be regularized by, e.g., a filter bank. This paper shows: (i) a simple way to determine common transducer responses for equalization based on comparison with an omnidirectional reference, (ii) a heuristic rule to define the filterbank frequencies depending on the Ambisonic order, diffuse-field-normalized weightings basic or max-rE, and available SNR, (iii) the appropriate handling of frequencies above the arrays spatial aliasing limit, and finally while the proposed solution supports loudspeaker playback: (iv) binaural playback requires a different equalization.
(See document for exact affiliation information.)
