Source position interpolation of spatial room impulse responses

Abstract

Measured spatial room impulse responses (SRIRs) are often used for realistic six degrees-of-freedom (6DoF) virtual reality applications, as they allow for the high quality capture and reproduction of a room’s acoustics. Dense sets of SRIR measurements are time consuming to acquire, especially for multiple source and receiver combinations, and so interpolation of sparse measurement sets is required. This paper presents a method for interpolating between higher-order Ambisonic SRIRs with a fixed receiver position but different sound source positions, using a previous methodology. The method is based on linear interpolation with spectral equalisation and RMS compensation, though direct sound, early reflections and late reverberation are processed individually. In a numerical comparison to a basic linear interpolation, the proposed method is shown to more smoothly fade between source positions in root-mean-square amplitude and direction-of-arrival metrics.