Abstract:

Natural sound is usually recorded in head-related stereophony by ear microphones of a dummy head in order to ensure a virtual reality, e.g., for subjective sound-evaluation procedures. However, in many applications the use of a dummy head is disturbing. Additionally, it is difficult to change the geometry or shape of a dummy head in order to adjust a recording individually. The spatial information of the sound field, which is perceived in appropriate stereophonic reproduction, is encoded in the two signals of the ear microphones by recording the modifications of the wave field from the near-field diffraction of the obstacle ``dummy head.'' A diffraction pattern can be approximated by a sufficient number of monopole sources at appropriate locations, and adjusted in amplitudes and phases. This method of source substitution is used for an arrangement of pressure microphones. Each microphone signal is processed twice and added in the stereophonic channels. If locations and transfer functions are optimized, the two channels represent an approximation of the ear-microphone signals of the dummy-head's (or real head's) diffraction information. The preliminary results of the optimization and listening experiments are reported. [Work supported by the German Graduate School on Psychoacoustics.]