(back to SpatialLocalizationPaper)

In the previous section we discussed how the preservation of HRTFs applied to a sound by recording it from within a dummy head can lead to a headphone listener localizing sounds outside the head. The advent of digital signal processing (DSP) makes it feasible to simulate these HRTFs electronically. This may sound like an uninteresting application with the ease of recording "true" binaural signals so low these days, but it is of great concern to those who wish to model aural landscapes that may not be accessible to recording, or may not exist at all. A complete cross-section of this field of research lies beyond the scope of this paper, but we can certainly whet the appetite.

One application of so-called "binaural technology" is room simulation. A wealth of research in the early 1990's has created systems that input:


They can then compute the amount and quality (equalization) of sounds that would reach the listener directly and through several orders of reflection. Most commonly the sound sources are "ray-traced," calculating the movement of several vectors as they are reflected across the room. One then takes an anechoic recording, applying the impulse response of the room reflections and convolving it with the HRTFs (Blauert, 377-8).

Truly virtual environments are conceptually similar to simulated rooms, but the practical amount of work required is much greater. Neither the listener's position nor the environment itself can be assumed constant, so calculations must be done effectively on the fly. Mainstream computer games began implementing these ideas in the late 1990's with the release of soundcard chipsets by (the now-defunct) Aureal Inc. that included preprogrammed binaural DSP algorithms. Game developers would specify the environment characteristics above at each visual frame cycle, and the soundcard would render the sounds appropriately when headphones or 2-way or 4-way speaker setups were selected. Microsoft soon created a standardized (if bare-bones) way for games to interface with cards, but in recent years development in this area has greatly slowed, especially compared with the rapid progress in visual 3D simulation.

onward to LocalizationConclusion