Computer Aided Sound System Design 1359
large-format touring line arrays and digitally controlled
loudspeaker columns but also with the increasing use of
inexpensive DSP technology employed for multiway
loudspeaker systems. Some of the issues conflicting
with above points 1–5 are listed in the following.
- A large line array system of some meters, height
cannot be measured adequately in its far field in addi-
tion to the fact that a large number of line array appli-
cations actually happens to take place mainly in the
near field. Therefore the simulation of a whole line
array as a point source is not valid within reasonable
error ranges.
- Another problem often encountered is insufficient
angular resolution. Loudspeaker columns and multi-
way loudspeakers exhibit significant lobing behavior
in the frequency ranges where multiple acoustic
sources interact at similar strength. Often too coarse
angular measurements fail to capture these fine struc-
tures and thus cause erroneous simulation results due
to aliasing/sampling errors.
- While in many cases the phase of the sound pressure
radiated by a simple loudspeaker is negligible at least
if it is considered on-axis and the run-time phase is
compensated for, the same is not true for most real-
world systems. On the one hand, for multiway
systems one cannot generally define a single point
where the measured phase response vanishes for all
frequencies and angular directions. This is the
problem of the so-called acoustic center for a set of
sound sources. In such cases the measured phase data
will typically show a run-time phase component that
depends on angle and frequency, no matter where the
point of rotation is. On the other hand, the inherent
phase response plays an important role in describing
the radiation behavior that is influenced by diffrac-
tion about the edges of the loudspeaker case, that is,
at angles of 60 degrees and more off-axis.
- Loudspeaker systems become increasingly configu-
rable, so that the user can adapt them to a particular
application. Typical examples include almost all
touring line arrays where the directional behavior is
defined mechanically by the splay angles between
adjacent cabinets, and in loudspeaker columns or
multiway loudspeakers, where the radiation charac-
teristics can be changed electronically by manipu-
lating the filter settings.
- In advanced computer simulations of sound rein-
forcement systems in venues, geometrical calcula-
tions must be performed. This is required to obtain
exact knowledge of which part of the audience might
be shadowed by obstacles between the sound sources
and the receivers. Geometric considerations are also
needed in ray-tracing calculation in order to find
reflections and echoes. For both processes, the reduc-
tion of a physically large loudspeaker system to a
point source can lead to significant errors. Depending
on the choice of the reference point for the source,
particular reflections might not be found or are exag-
gerated, or a large fraction of the audience area might
be seemingly shadowed by a very small object.
In addition to the above, a set of minor problems is
also evident. This includes the definition of maximum
power handling capabilities of multi-input systems that
are represented by a single point source. The avail-
ability of case drawings to help in the mechanical
design and the clear indication of the reference point
that was used for the measurements are is important.
As a result of the obvious contradictions, a variety of
proposed solutions emerged in the later 1990s. This
development happened partially by the loudspeaker
manufacturers and partially by the creators of simula-
tion software as well. To resolve the problem of large-
format loudspeaker systems, a subdivision into smaller
elements is required to be able to measure them and use
them for prediction purposes. To properly model the
coherent interaction between these elements, complex
measurement data, including both magnitude and phase
data, is needed.
The most prominent solutions can be summarized as
follows. Instead of measuring a whole system, so-called
far-field cluster balloons were calculated based on the
far-field measurement of individual cabinets or groups
of loudspeakers.^30 To describe individual sound sources,
phase data was introduced in addition to the magnitude-
only balloon data.31,32 Mathematical models providing
phase information implicitly were applied, such as
minimum phase or elementary wave approaches as well
as 2D sound sources.^17 However, these first approaches
lacked generality and thus their implementation into
existing simulation software packages was specific,
difficult, or even impossible.
The situation was resolved first by the concept of the
loudspeaker DLL (dynamic link library), which serves
essentially as a programmable plug-in for simulation
software.^33 Another concept, namely the GLL (generic
loudspeaker library), introduced a new loudspeaker data
file format that is significantly more flexible than the
conventional data formats, and is designed to resolve
most of their apparent contradictions.^34 We will review
both approaches in the next section as they have turned
out to be a standardized way to model complex loud-
speaker systems.
