A robustness analysis of 3D audio using loudspeakers

It is well known that the effectiveness of 3D audio systems is critically dependent on the listener's head being in a known location. In this paper we analyze the fundamental role played by the loudspeaker positions in determining the robustness of the crosstalk canceler. Based on an extremely simple head model, we derive straightforward expressions for the loudspeaker positions that optimize the system robustness, which is measured by matrix condition numbers. These derived optimum positions are then compared with empirically-derived optimum positions obtained from actual HRTF (head related transfer function) measurements. The results indicate that our analytical expressions accurately predict the optimum loudspeaker positions.