Transaural virtual reality applied to front-back localization
Peter Xinya Zhang, William M. Hartmann · The Journal of the Acoustical Society of America · 2005
Human sound localization can be studied by measuring head-related transfer functions for different source locations and subsequently using the data to simulate different spatial locations using headphone presentation. This technique fails to adequately simulate the difference between front and back locations because the frequencies relevant to pinna cues are so high. A transaural (cross-talk cancellation) technique has been developed that is accurate up to 16 kHz. By controlling the levels and phases of spectral components, as measured by probe microphones in the ear canals, the technique eliminates significant differences between the baseline simulation and reality. The technique has been used, with modifications to the baseline simulation, to study the front-back information content in different frequency bands by eliminating the information while maintaining the energy. It has also been used to test for orthogonality between front-back localization and azimuthal localization based on simple interaural differences. The technique has verified that front-back discrimination cannot be mediated by interaural differences alone. By flattening the level spectrum in one ear while leaving the other ear unchanged it has been possible to demonstrate an ear advantage for front-back discrimination. [Work supported by the NIDCD, grant DC 00181.]