Steven D. Trautmann
Center for Comput. Res. in Music and Acoust., P.O. Box 9675, Stanford, CA 94309
Audio wavefront reconstruction is similar to visual holography in which optical wavefronts are reconstructed by reproducing phases and amplitudes to create the impression of an object being present when it is not there. In audio there are far fewer speakers than necessary to exactly reproduce a wavefront this way. In order to construct the closet possible approximation to a desired wavefront, information about the listening environment such as the number, placement and qualities of the loudspeakers and the nature of local reverberation can be used. The perceived quality of this approximation can be improved by using psychoacoustical properties of sound localization, such as treating amplitude as more vital than phase, ear, and head filtering, precedence effects and masking. Thus despite limited numbers of loudspeakers, improvements can be made in sound localization for quadraphonic and stereo systems, especially at low frequencies. This is done by using signal processing techniques including partial inverse filtering of the acoustical environment to cancel out unwanted room reverberation, and a criterion applied to a multidimensional system in order to minimize the maximum difference relative to the desired perception anywhere a listener might be located.