APPLICATIONS OF LEAST-SQUARES FIR FILTERS TO VIRTUAL ACOUSTIC SPACE

A virtual acoustic space (VAS) employs the localization cues specified by the direction-dependent 'free-field to eardrum transfer function' (FETF) to synthesize sound-pressure waveforms present near the tympanu m. The combination of a VAS and the earphone delivery of synthesized w aveforms is useful to study parametrically the neural mechanisms of di rectional hearing. The VAS-earphone procedure requires accurate FETF e stimation from free-field measurements and appropriate compensation fo r the undesirable spectral characteristics of the closed-field earphon e sound delivery and measurement systems. Here we describe how special ly designed finite-impulse-response (FIR) filters improve these two op erations. The coefficients of an FIR filter are determined using a lea st-squares error criterion. The least-squares FIR filter is implemente d entirely in the time domain and avoids the usual problems with divis ion inherent in a frequency domain approach. The estimation of an FETF by a least-squares FIR filter is veracious since its impulse response can recover signals that were recorded near the eardrum in the free f ield with a very high fidelity. The correlation coefficient between re corded and recovered time waveforms typically exceeds 0.999. Similarly , least-squares FIR filters prove excellent in compensating closed-fie ld sound systems since comparisons of waveforms delivered by a compens ated earphone to their corresponding predistorted signals yield correl ation coefficients that exceed 0.99 on average.