Handbook for Sound Engineers

Acoustical Modeling and Auralization 233

composite data for a wide segment of the test results.
Another issue of concern when using headphone repro-
duction is that the apparent source location will move
with the listener’s head movements, something that
does not take place in the real world.

9.3.2.3 Multichannel Reproduction Using Loudspeakers

In this instance, the impulse response of the room is
divided into components that correspond to the general
locations in space from where the reflections originate
as shown in Fig. 9-26. Various systems have been
developed30,31 throughout the years ranging from just a
few speakers to hundreds of units driven by dozens of

separate channels. The advantage of the technique is

that the system relies on the listener’s own HRTF while

also allowing for head tracking effects. From the per-

spective of efficiency the approach can be implemented

with minimal hardware and software since the reflec-

tions can be categorized in terms of their direction of

arrival while the IR is being generated. The multichan-

nel reproduction technique can actually be imple-

mented from a physical scale model without the need

for computer tools by using delay lines and an analog

matrix system.^1 The reproduction system is, of course,

rather complicated since it requires substantial hardware

and an anechoic environment.

9.3.3 Real-Time Auralization and Virtual Reality

A real-time auralization system allows the user to actu-

ally move within the room and to hear the resulting

changes in the sound as they actually happen. This

approach requires the near-instantaneous computation

of the impulse response so that all parameters pertaining

to the direct sound and to the reflections can be com-

puted. In a recent implementation^32 the space is mod-

eled using an enhanced image method approach in

which a fast ray-tracing preprocessing step is taken to

check the visibility of the reflections at the receiver

location. The air absorption and the properties of the

surface materials are modeled using efficient digital fil-

ters, and the late reverberation is described using tech-

niques that give a smooth and dense reflection pattern

Figure 9-22. The basic auralization process.

Impulse response reflectograms

Reflectogram at 16 kHz

Reflectogram at 2 kHz

Reflectogram at 500 Hz

Reflectogram at 250 Hz

Reflectogram at 125 Hz

dB

Time

Time

dB

Anechoic music sample

Convolution engine (software

and/or hardware)

Playback

system

Reflectogram at 1 kHz

Figure 9-23. An example of a complete reflectogram at
1000 Hz.

Direct sound

Discrete reflections (early and late)

Late reverberation

80 dB

60 dB

40 dB

20 dB

0 ms 200 ms Time