Directional analysis with microphone array mounted on rigid cylinder for directional audio coding

A spatial-sound processing technique, Directional Audio Coding (DirAC), provides a parametric representation of a sound field, containing data of the direction of arrival and the diffuseness in frequency channels depending on time. Such a representation can be obtained by a directional analysis from a square array of four omnidirectional microphones, which enables reliable direction estimation only within a limited frequency range. This paper proposes inserting a rigid cylinder into the array to cast an acoustic shadow and cause scattering. This produces prominent inter-microphone differences that are utilized in two alternative direction estimation methods, based either on energy gradients or vector quantization, having different computational complexities. Both methods estimate the direction reliably for almost over the entire audible frequency range, which is corroborated with objective measurements and subjective listening tests.