Binaural rendering using higher-order stereophony.

Abstract

Higher-order stereophony is a new approach for spatial audio reproduction which extends classic two-channel stereophony to higher order soundfield reproduction and generalised multi-channel loudspeaker arrays. Higher order stereophony achieves accurate soundfield reproduction over a line by reproducing the degree m = 0 spherical harmonic soundfield coefficients only. The reproduction line is assumed to align with the interaural axis of a listener. This article addresses the extension of higher order stereophony to binaural reproduction. The technique is shown to exactly reproduce binaural signals when using a rigid sphere head-related transfer function model, and to reorder the energy of more generalised head-related transfer functions into spherical harmonic coefficients with degree index close to 0. To truncation order N, higher order stereophony requires only (N + 1) spherical harmonic coefficients compared to (N + 1)2 with higher order ambisonics, and the two techniques are compared through simulations and a listening test. Higher order stereophony is shown to perform similarly to higher order ambisonics under truncation to the same order, but using a smaller number of soundfield coefficients. For higher virtual source elevations, higher order stereophony performs worse than higher order ambisonics due to its ability to only reproduce axisymmetric head-related transfer functions.