ArticlePDF Available

Multiactuator Panels for Wave Field Synthesis: Evolution and Present Developments

Authors:

Abstract and Figures

A technical review is presented of the development of a special type of planar loudspeaker array for wave field synthesis, known as multiactuator panel. It consists of a thin, stiff panel with a small back volume, which vibrates by means of a group of mechanical exciters, each driven by a different signal. Multiactuator panels are used as alternatives to arrays of classic loudspeakers with conically shaped diaphragms for wave field synthesis, with added benefits such as low visual profile and diffuse radiation. However, the use of multiactuator panel arrays poses some problems and technical challenges that have been described in the literature. A historical review of the evolution from single-to multi-excited panels is given, followed by a technical discussion of the current developments in the field of wave field synthesis reproduction by means of multiactuator panels.
Content may be subject to copyright.
PAPERS
Multiactuator Panels for Wave Field Synthesis:
Evolution and Present Developments*
BASILIO PUEO
1
,AES Member, JOSE
´J. LO
´PEZ
2
,AES Member,
(basilio@ua.es) (jjlopez@dcom.upv.es)
JOSE
´ESCOLANO
3
,AES Associate Member, AND LARS HO
¨RCHENS
4
,AES Student Member
(escolano@ujaen.es) (l.horchens@tudelft.nl)
1
Communication and Social Psychology Department, University of Alicante, Alicante, Spain
2
iTeAM Institute, Universidad Politecnica de Valencia, Valencia, Spain
3
Telecommunication Engineering Department, Polytechnic School, University of Jaen, Jaen, Spain
4
Laboratory of Acoustic Imaging and Sound Control, Delft University of Technology, Delft, The Netherlands
A technical review is presented of the development of a special type of planar loudspeaker
array for wave field synthesis, known as multiactuator panel. It consists of a thin, stiff panel
with a small back volume, which vibrates by means of a group of mechanical exciters, each
driven by a different signal. Multiactuator panels are used as alternatives to arrays of classic
loudspeakers with conically shaped diaphragms for wave field synthesis, with added benefits
such as low visual profile and diffuse radiation. However, the use of multiactuator panel arrays
poses some problems and technical challenges that have been described in the literature. A
historical review of the evolution from single- to multi-excited panels is given, followed by a
technical discussion of the current developments in the field of wave field synthesis
reproduction by means of multiactuator panels.
0 INTRODUCTION
One of the important topics in sound reproduction has
always been the preservation of the spatial and temporal
characteristics of recorded sound events. Sound field
reproduction can be generated in essentially two ways: 1)
by providing signals directly to the ears that are similar to
those that would have occurred if there had been real
sound sources in the intended positions (this approach is
known as binaural reproduction [1], [2]), and 2) by
recreating the sound field over a more extended region of
space, which will be interpreted by the listener as being in
the environment intended to be simulated. Within this
second approach, known as holographic-based reproduc-
tion, there are the Ambisonic system [3] and wave field
synthesis, the latter being described in this paper.
Wave field synthesis is a spatial sound rendering
technique that generates a true sound field using
loudspeaker arrays [4], [5]. Wave fields are synthesized
based on virtual sound sources at some position in the
sound stage behind the loudspeakers or inside the
listening area. When using sound reproduction based on
wave field synthesis, sound fields can be generated in a
spatially and temporally correct way. Therefore the
acoustic perspective is perceived by listeners as in a real
sound event. This includes a realistic change of the sound
pressure level in an extended listening area where the
distance from the virtual source varies. The objective of
wave field synthesis to create an exact copy of the wave
field can only be fulfilled if an infinite, continuous
distribution of sound sources is used. In practical
implementations, however, the ideal source distribution
is replaced by discrete loudspeakers, which in turn causes
the array to have finite dimensions, thus limiting the
performance of the wave field synthesis systems.
Most practical wave field synthesis systems employ
arrays of dynamic loudspeaker drivers. However, these
drivers have disadvantages. To avoid back-to-front
cancellation in the low-frequency region, dynamic
loudspeakers need a housing with a relatively large
volume, leading to aesthetic difficulties in real listening
environments. Furthermore their pistonlike diaphragms
produce a nonuniform directivity consisting of a beam
normal to the loudspeaker cone and narrow lobes around
their main axes, which sharpens with increasing frequen-
cy. Alternatively to dynamic loudspeaker arrays, multi-
actuator panels are a special type of flat panel loudspeak-
ers that can also be used for wave field synthesis with
several advantages. Due to their ‘‘diffuse’’ radiation and
omnidirectionality, the listening area where individual
*Manuscript received 2009 April 9; revised 2010 January 11,
June 18, September 14, and October 22.
J. Audio Eng. Soc., Vol. 58, No. 12, 2010 December 1045
components of the sound pressure field merge correctly
into the desired wavefront is wider. Also the pressure
level decay with distance is less pronounced for typically
sized distributed-mode loudspeakers in comparison with
pistonlike loudspeakers. In addition multiactuator panels
can be integrated into room interiors because of their low
visual profile.
The aim of this work is to present a review of the past
and present developments in the field of wave field
synthesis reproduction by means of multiactuator panels.
In Section 1 an overview of the fundamentals of the wave
field synthesis spatial sound rendering technique is given,
considering real-world installations. Section 2 describes
the first attempts of using flat panel loudspeakers for
multichannel audio systems and the evolution from
single-excited panels to multiactuator panels. Section 3
explores the behavior of multiactuator panels in the
context of wave field synthesis reproduction, and
temporal and spatial characteristics are discussed. The
improvements in multiactuator panel technology to create
a sound field of quality comparable to that generated with
conventional loudspeakers are presented in Section 4.
Section 5 depicts the current multiactuator panel proto-
types that were designed and built as a result of the
research described in this paper.
1 WAVE FIELD SYNTHESIS
1.1 Introduction
Wave field synthesis has been introduced by Berkhout
[6] as a concept for sound reproduction without the sweet-
spot restrictions inherent in common multichannel
systems, as illustrated in Fig. 1. In two- (or more-)channel
stereo playback the spatial properties of the reproduced
field are determined by the characteristics of the
loudspeakers. The source localization is correct only in
a small area between the loudspeakers, shown by dashed
lines in Fig. 1(a). In wave field synthesis the wave
patterns of the sources to be reproduced are correctly
synthesized in time and space by an array of closely
spaced loudspeakers such that their localization is correct
for all listeners in the audience area. The acoustics of the
environment of the primary sources are also reproduced
since the sound reflections coming from mirror image
sources are accurately positioned in space.
The wave field synthesis concept is based on the
Huygens principle, stating that the propagation of a wave
through a medium can be described qualitatively by
adding the contributions of all secondary sources
positioned along a wavefront. This implies that when
the wave field on the boundary surface of a closed,
source-free volume is known in terms of pressure and
normal particle velocity, the sound pressure at any point
within that volume can be determined. Therefore the
surface can be interpreted as covered with a continuous
distribution of secondary dipole sources driven by the
local pressure generated by the primary sources, plus a
distribution of secondary monopole sources driven by the
local velocity, normal to the boundary. Together all these
secondary sources can be regarded as generating a field
within the closed volume. This field is identical to the
field that primary sources would have generated there.
The closed boundary can be replaced by a pseudoin-
finite plane with the primary sources at only one adjacent
half-space. Then the sound pressure at any point in the
other half-space can be determined either from the normal
velocity distribution in that plane, or from the pressure
distribution, as shown in [7].
In the context of wave field synthesis this plane can be
regarded as a dense planar array of monopole loudspeak-
ers (secondary sources), which are driven individually by
a signal equal to the local normal velocity generated by
the synthesized sound sources (primary sources). This
array creates a true copy of the primary sound field at the
source-free half-space, as depicted in Fig. 1(b). These
primary sound sources can be real as in a live event, or
they can be virtual sources if they are recorded. Also
loudspeakers with dipole characteristics could be used.
However, it has been shown [8] that by adapting the
driving signal, arrays with loudspeakers having nonideal
monopole or dipole characteristics can also be used for
wave field synthesis. A comprehensive derivation of the
underlying mathematics can be found in several studies,
such as [6], [4], [7].
Fig. 1. Illustration of basic difference between two-channel stereo and wave field synthesis. (a) Stereo playback. Proper sound
localization is shown between two dashed lines where sweet spot is located. (b) Wave field synthesis. P—primary source; S—
secondary source.
1046 J. Audio Eng. Soc., Vol. 58, No. 12, 2010 December
PUEO ET AL. PAPERS
1.2 Practical Implementations
The application of planar loudspeaker arrays, as
prescribed by the preceding theory, is—apart from the
huge computational power required to feed a large
number of loudspeakers—not practical aesthetically.
Therefore in common wave field synthesis practice linear
loudspeaker arrays are used which permit synthesizing
wave fields in the ear plane of the listeners, neglecting the
correct synthesis of elevated wave components for which
the human hearing mechanism is less sensitive [1]. For
this purpose two-dimensional array configurations are
used to synthesize the field of three-dimensional sources.
The derivation of the Rayleigh operators from the three-
dimensional version to two-dimensional arrays leads to
the so-called 2.5-dimensional versions [9].
Since the aim of wave field synthesis is the creation of
a true copy of a natural sound field, this aim can only be
fulfilled with certain restrictions in practice. Practical
implementations of the wave field synthesis technique are
based on loudspeaker arrays, which act as secondary
sound sources. The distribution of these sources is not
dense or infinitely continuous, but consists of a finite set
of individual discrete loudspeakers.
The first consequence is that the distance between
transducers Dxdefines a spatial sampling frequency.
Above this frequency spatial aliasing effects are to be
expected. The frequency limit of aliasing-free reconstruc-
tion is given by
fal ¼c
2Dxjsin amaxjð1Þ
where a
max
indicates a maximum observation angle. A
description of these aliasing artifacts can be found in
[10]–[12].
The second consequence is that real loudspeaker arrays
are always of finite length, causing truncation effects. A
finite array would cause diffraction at the edges, giving
rise to coloration, and would degrade localization to an
extent. This effect can be minimized by using a spatial
weighting function which would attenuate the outer
loudspeakers. However, attenuating edge loudspeakers
will reduce the size of the correctly reproduced wave
field. Thus the choice of the weighting function depends
on a tradeoff between the reduction of diffraction artifacts
and the size of the listening area.
Accordingtothisdiscussionandconsideringa
pragmatic approach to the problem, the appropriate
source for wave field synthesis requires the following.
In order to prevent spatial aliasing, the distance between
loudspeakers should be low, and therefore loudspeakers
with small dimensions should be used to fit into an
enclosure. On the other hand, small loudspeakers
typically have a problem reproducing the low-frequen-
cy band.
In order to minimize the artifacts caused by truncation
effects, the array should be large enough to cover a
given listening area. Together with the first require-
ment, this means that the total number of loudspeakers
should be very large.
According to the wave field synthesis theory, loud-
speakers substituting secondary sources must ideally act
as dipoles or monopoles. Although this is not
achievable in practice, it is advisable that loudspeakers
have a fairly constant directivity in all directions for a
reasonable frequency bandwidth.
Since a large number of loudspeakers should be
positioned around the listening area, it would be of
advantage if the loudspeakers were forming part of the
room interiors.
The next section investigates an alternative way to
create the field of wave field synthesis. This solution
makes use of so-called flat panel loudspeakers, which
overcome most of the disadvantages of the use of
dynamic loudspeakers for wave field synthesis.
2 FLAT PANEL LOUDSPEAKERS
2.1 Distributed-Mode Loudspeakers
A distributed-mode loudspeaker essentially consists of
a thin, stiff panel that vibrates in a complex pattern over
its entire surface driven by an electromechanic transducer,
called an exciter [13] (Fig. 2). The exciter is usually an
electrodynamic driver with the coil attached to the panel
and the magnet serving as a proof mass to produce an
inertia force. Exciters are carefully positioned onto the
flexible panel so as to promote a richness of excited
mechanical modes. The radiation of a distributed-mode
loudspeaker is based on the vibration of dense mechanical
resonance frequencies. Complex mechanoacoustic radia-
tion effects yield the typical omnidirectional radiation
pattern [14], [15]. As shown in Fig. 2, to prevent back-to-
front cancellations in the low-frequency range, distribut-
ed-mode loudspeakers need to be mounted in a housing
[16], [17] with absorbing materials.
Distributed-mode loudspeakers provide some advan-
Fig. 2. Distributed-mode loudspeaker. Electric wiring is
omitted for simplicity.
J. Audio Eng. Soc., Vol. 58, No. 12, 2010 December 1047
PAPERS MULTIACTUATOR PANELS FOR WFS
tages over pistonlike loudspeakers, which makes them
suitable as transducers for wave field synthesis. As slim
flat structures they can be integrated into the listening
environment more easily than the arrays of conventional
loudspeakers. For applications where aesthetics are
important they can be part of the room decoration as
walls or furniture. In addition distributed-mode loud-
speakers have an even power distribution over a wider
frequency band [18], [19]. This feature assures that there
will be fewer changes in pressure levels and their
associated artifacts in large listening areas. Moreover
room reflections produced by distributed-mode loud-
speakers are less correlated to the direct sound than those
radiated from pistonlike loudspeakers, and thus construc-
tive and destructive interference of sound is minimized
[20]–[22]. Another advantage is that localization preci-
sion as a function of room acoustics is improved by the
use of these acoustic radiators [23]. The proper sound
localization and the minimum degradation caused by
room acoustics were of crucial importance in considering
such radiators as alternatives to dynamic loudspeakers for
multichannel systems, as will be explained in the next
section.
2.2 From Single- to Multi-Exciter Panels
In 2000 Boone and Brujin tested for the first time the
applicability of distributed-mode loudspeakers for wave
field synthesis reproduction [24]. In single-excited
distributed-mode loudspeakers they reported that the first
wavefront, which is related to the first part of the impulse
response of the reproduction system, was as good as with
conventional loudspeakers.
However, at the beginning it was thought that these flat
radiators could alter the phase relationships between the
secondary sources (loudspeakers). A definite phase
response is of crucial importance in wave field synthesis.
To investigate this, a linear array of nine rectangular
single-excited distributed-mode loudspeakers with radiat-
ing surfaces of 180 mm by 125 mm was described in [25]
and is sketched in Figs. 3 and 4.
These experiments confirmed that distributed-mode
loudspeakers could be used for wave field synthesis since
Fig. 3. Three different ways to implement an array for wave field synthesis. (From [25].) (a) Loudspeaker array with oval
dynamic loudspeakers. (b) Array of individual distributed-mode loudspeaker panels. (c) Multi-exciter distributed-mode
loudspeaker.
Fig. 4. Oval dynamic loudspeaker of Fig. 3 (top) and single-excited prototype viewed from side (bottom). Distributed-mode
loudspeaker enclosure is lined with damping material to decrease reflections within enclosure. (From [25].)
1048 J. Audio Eng. Soc., Vol. 58, No. 12, 2010 December
PUEO ET AL. PAPERS
the localization cues were reproduced properly. As a
consequence, individual single-excited distributed-mode
loudspeakers reconstruct the wave field correctly and can
form an array for wave field synthesis equivalent to that of
dynamic loudspeakers. However, the spacing of the
secondary sources required by the wave field synthesis
algorithm to acquire a reasonably useful bandwidth forced
the size of the panels to be very small. For example, in order
to achieve an aliasing frequency of 1 kHz, the distance
between secondary sources must be approximately 170 mm.
However, small panels pose problems with sound
quality since they have a poor low-frequency response.
For that reason it was decided to extend the distributed-
mode loudspeaker technology to a single large panel with
multiple exciters, each driven by a different signal, and to
study whether exciters sharing the same vibrating surface
were acting as individual secondary sources. Fig. 5(b)
depicts such a configuration in contrast to the single-
excited transduction shown in Fig. 5(a).
Since transducers must be spaced very closely in order
to reconstruct the waveform up to a reasonable frequency,
a vibrating panel with low internal damping will result in
crosstalk. For that reason it is necessary to use panel
material with high internal damping. This allows each
exciter to excite only a small part of the panel around its
location. Therefore this configuration performed as a
wave field synthesis array since exciters behaved like
small individual loudspeakers. Conversely high internal
damping causes a loss of efficiency, because part of the
energy is dissipated within the panel and is not radiated as
acoustic energy.
Polycarbonate resin panels were chosen for the first
prototypes because of their high internal damping with
well-defined source localization. However, this material
had a very poor impulse response, showing a lot of
ringing [25]. Foam board became an alternative to
polycarbonate resin panels as it exhibited better impulse
response and maintained the independent behavior of the
exciters for wave field synthesis. As with polycarbonate
resin panels, digital filtering was still necessary to obtain
an adequate sound quality. This concept was then referred
to as a single-panel, multi-exciter array, a term that was
used in subsequent studies [26], [27].
3 MULTIACTUATOR PANELS FOR WAVE FIELD
SYNTHESIS
The experiments described in Section 2, and reported in
[28], [25] were conducted at the Delft University of
Technology (TU Delft) with the aim of designing a
loudspeaker reproduction system for the wave field
synthesis rendering technique in the framework of the
European IST project CARROUSO [29]. Although the
first publication announcing the project [30] did not
include any reference to flat panel loudspeakers, this type
of loudspeaker was used more and more frequently.
As a result of early investigations, the first multi-
actuator panel prototypes were shown at several confer-
ences, such as the AES 19th International Conference in
Schloss Elmau in 2001 June (see Fig. 6), the IFA in Berlin
in 2001 August, and the IST event in Du¨sseldorf in 2001
December.
Fig. 5. (a) Array of single-excited distributed-mode loudspeakers. (b) Multiactuator panels.
Fig. 6. First multiactuator panel prototypes shown at AES 19th International Conference on Surround Sound, Schloss Elmau,
Germany, 2001 June 21–24.
J. Audio Eng. Soc., Vol. 58, No. 12, 2010 December 1049
PAPERS MULTIACTUATOR PANELS FOR WFS
In 2004 Boone published a paper in the Journal where
all the refinements achieved during three years of
investigation were collected [31]. It was not until that
year that such arrays appeared in the technical literature
as multiactuator panels, or MAPs, to distinguish them
from the single-excited distributed-mode loudspeaker.
This time instead of using the early prototypes, the new
eight-exciter multiactuator panels from the CARROUSO
project were used in the experiments.
Two years later another significant publication in the
development of the multiactuator panel was the structural
acoustic analysis by means of a laser Doppler vibrometer
made at TU Delft by de Vries and other researchers [32].
This study confirmed experimentally the basis on which
multiactuator panels rely since the first publications,
namely, that acoustic radiation is generated almost
entirely by the structural near field around the excitation
point on the panel. The damping loss factor of the
multiactuator panel material must be large to produce
such behavior. In addition panel dimensions and exciter
positions played a role in the acoustic radiation of
multiactuator panels. It was then concluded that particular
choices of panel size and damping factor were needed to
meet the requirements of wave field synthesis.
In the following the results of these and other related
publications are presented to explore the behavior of
multiactuator panels in the context of wave field synthesis.
3.1 Temporal Behavior
An illustrative way to describe the temporal behavior
of a multiactuator panel is given by structural analysis
using laser vibrometry [32] and by near-field acoustic
holography measurements. Fig. 7 shows pressure mea-
surements of the TU Delft prototype. The details of these
Fig. 7. Three-dimensional representation of near-field radiation, TU Delft multiactuator panel prototype, measured at a
distance of 10 mm, grid spacing of 10 mm, for different time steps.
1050 J. Audio Eng. Soc., Vol. 58, No. 12, 2010 December
PUEO ET AL. PAPERS
measurements will be given in Section 5.1. An omnidi-
rectional microphone was used to obtain the pressure
distribution on a grid of 10-mm spacing at a distance of
10 mm from the panel surface. The extension of the
measurement grid was chosen to be several millimeters
larger than the panel. From the eight exciters mounted on
the panel, only the third from the right was driven in order
to determine the near-field impulse response of the
multiactuator panel.
The pressure maximum in the measurement area is
reached above the exciter at t¼1 ms, as shown in Fig.
7(a). The delay is due to the measurement chain.
Amplitudes of the plots are normalized to the maximum.
As time proceeds, structural waves start to propagate
through the panel. These waves are scattered by other
exciters and reflected from the panel boundaries, as can be
seen in Fig. 7(d)–(h). Because of high damping in the
material due to the dispersion of the bending wave, the
sound pressure in the measurement area drops signif-
icantly within a few milliseconds. The combination of
edge reflections and dispersion leads to a rather diffuse
panel radiation after about 3–4 ms, as shown in Fig. 7(f).
3.2 Directivity Characteristics
The near-field acoustic holography measurements of a
multiactuator panel can also be used in order to calculate
its directivity. Therefore the wave field in the measure-
ment plane is extrapolated to points in the far field using
the Rayleigh II integral [7]. Because of spatial sampling
the extrapolation is only valid up to the spatial aliasing
frequency. For the given grid spacing (10 mm) wave-
lengths down to 20 mm or, equivalently, frequencies up to
17 kHz, can be used.
Typical directivity characteristics of the multiactuator
panel are presented in Fig. 8 for several frequencies. The
number of lobes increases with frequency. Simultaneous-
ly lobes become more narrow. A significant drop in sound
pressure level can be observed for certain directions and
frequencies. Note that the frequency response may vary
considerably with frequency, as is also observed with
standard cone loudspeakers. However, the diffuseness of
multiactuator panels reduces interference at the listening
point if walls are nearby. This behavior is not so
pronounced with standard cone loudspeakers. The reason
is that this kind of diaphragm is optimized for rigid body
Fig. 8. Directivity of TU Delft multiactuator panel prototype for several frequencies. Left—cross sections of horizontal plane;
right—equivalent three-dimensional representations.
J. Audio Eng. Soc., Vol. 58, No. 12, 2010 December 1051
PAPERS MULTIACTUATOR PANELS FOR WFS
motion where the off-axis radiation is in phase or anti
phase with the on-axis response.
3.3 Spatial Frequency Response
The diffuseness of the distributed-mode operation in
multiactuator panels minimizes the drawbacks of room
interaction. Multiactuator panels radiate an even sound
pressure across a large listening area, and the reflections
on the room boundaries are less correlated with the direct
sound.
In the context of wave field synthesis this diffuse nature
has a direct consequence in the way the wave field is
reproduced. As discussed in Section 1.2, the region above
the aliasing frequency is characterized by oscillations in
both frequency and space. These interference patterns can
be experienced when pacing the listening area. However,
the diffuse behavior of multiactuator panels helps to
reduce these effects to some extent. Corteel et al. carried
out a comparison between multiactuator panels and
dynamic loudspeakers on an objective and a subjective
level for wave field synthesis [33]. The impact of the
diffuse behavior is evaluated by computing the spatial
response of a loudspeaker array comprising 48 ideal
omnidirectional loudspeakers with 150-mm spacing. The
wave field is sampled using a microphone array composed
of 96 omnidirectional microphones with 100-mm spacing,
located 2 m from the loudspeaker array. Two types of
filtering are used above the aliasing frequency (1.2 kHz):
1) a classical wave field synthesis filter, consisting of a
delayed and attenuated dirac pulse, which will be called
discrete filter (MEQ), and 2) a full diffuse filter, which is
generated from time-limited white noise that is generated
independently for each loudspeaker in order to obtain
uncorrelated outputs (FD). Fig. 9 displays the spatial
frequency response of the loudspeaker array for discrete
and diffuse filters.
For the discrete filter, where diffusion simulation is
switched off, there is a varying response through both
frequency and space. This disturbed response is charac-
terized by strong interference patterns, which can be
noticed by pacing the listening area. However, when the
diffusion feature of the filter is switched on, the
interference patterns appear not so pronounced but rather
chaotic, as illustrated in Fig. 9(b). Typically an irregular
interference pattern seems to increase the quality of the
listening experience. The diffusion, as simulated by the
filter experiment, can be found naturally in multiactuator
panels. As a consequence the inherent diffuse radiation of
multiactuator panels beneficially minimizes the percep-
tual impact of the aliased spatial spectrum. From an
objective point of view these results suggest that the
aliasing artifacts start to appear at a certain frequency, but
due to diffuseness they increase in a progressive way.
Another indicator of the quality of the spatial response,
which can be derived, is the analysis of coloration
(fluctuations in frequency). Coloration happens particu-
larly in wave field synthesis setups above the aliasing
frequency. The annoying interference patterns mentioned,
caused by practical applications of wave field synthesis,
can be seen as the result of an aliased spatial sampling
process. In [34] a methodology for the analysis of sound
field radiation in the space–time domain is applied to
observe the effects that spatial sampling with conven-
tional loudspeakers and with multiactuator panel loud-
speaker arrays have on the wave field. According to this
analysis the spatial aliasing artifacts at high frequencies
are characterized by a superposition of undesired plane
waves with different incidence angles at different
frequencies, causing the amended interference patterns.
As shown in Fig. 10, a graphical representation of spatial
(k
x
) versus temporal (k) frequencies facilitates the visual
identification of the aliased contributions [35].
One important conclusion in this study was that the
number of aliased contributions in practical multiactuator
panel implementations for wave field synthesis depends
not only on truncation effects due to the finiteness of the
panel, but also on the number of intermediate boundaries.
To illustrate this, Fig. 10(a) shows two equivalent
multiactuator panels having the same exciter distance Dx
l
¼0.126 m. The top one is composed of a single panel and
Fig. 9. Spatial frequency responses as a function of
diffusion. (From [33].) (a) Discrete filter with no diffusion.
(b) Diffuse filter.
1052 J. Audio Eng. Soc., Vol. 58, No. 12, 2010 December
PUEO ET AL. PAPERS
the bottom one comprises several small multiactuator
panels attached in landscape orientation. Consider a plane
wave, band-limited to 8.2 kHz, that is reproduced by these
two multiactuator panels and arrives at normal incidence
at a linear microphone array.
The space–time domain representation of both systems
is presented in Fig. 10(b). For the single multiactuator
panel a series of replicas are shifted 2p/0.126 ¼50 m
1
with respect to the nonaliased spectrum at k
x
¼0,
illustrated by bold lines. Among the six possible replicas
within the triangle representation, only four are visible
because the signal is band-limited to k¼150 m
1
(8.2
kHz).
To observe that aliasing contributions consist of new
plane waves with h
(al)
angles, let us consider an example
temporal frequency of 128 m
1
(7 kHz). As indicated in
Fig. 10(b) by a continuous line, at this particular
frequency four additional plane waves are reproduced
whose directions are 6238and 6518[see Fig. 10(c)]. In
general, as the temporal frequency augments, so does the
number of aliasing plane waves. Also, their aliasing
angles h
(al)
tend to narrow toward the desired angle, in
this case 08.
If the multiactuator panel comprises 15 exciters with a
total length of 2 m and the listening point is centered 1.2
m in front of the array, truncation effects will modify the
number of aliasing contributions. Since the listener
presents an angle of 6408with respect to the multi-
actuator panel limits, as shown in Fig. 10(c), a more
narrow triangle must be considered in the k
x
versus k
representation, shaded in gray in Fig. 10(b). Because of
this limited reproduction area, the example plane wave
discussed before (k¼128 m
1
,f¼7 kHz) will present
only two aliasing contributions at 6238, since the h¼
6518plane waves lie outside the gray triangle.
Second, consider the equivalent multiactuator panel of
Fig. 10(a) in which panels with two exciters each are used
to create the loudspeaker array. This time, for the range
plotted in Fig. 10(b), six new lines appear at spatial
frequencies of k
x
¼625, 675, and 6125 m
1
, in between
the lines already shown in the single multiactuator panel
at k
x
¼0, 650, and 6100 m
1
. The reason for this new
spatial pattern can be found in the multiactuator panel
boundary conditions since the panel frames are motion-
less, regardless of the input signal. Therefore a new
sampling spacing of Dx
p
¼0.252 m, the distance between
center panels, creates a series of replicas, shifted by 2p/
0.252 ¼25 m
1
. The practical consideration of this
additional sampling process is that new undesired plane
waves will arise on the reproduced wave field at angles of
611.28,635.88, and 677.68. Due to truncation effects
only the first two angles are reproduced in the listening
area, labeled ato din Fig. 10(c).
To summarize, in terms of acoustic behavior, large
panels are more advisable than small panels for wave field
synthesis reproduction. The use of small panels with
fewer exciters has undesired consequences on the
reproduced wave field. Since the contour of the panels
presents negligible motion when radiating, new aliasing
artifacts, at frequencies below that of larger panels or
dynamic loudspeaker arrays, will appear, which are a
function of the distance between panel centers. Then, to
avoid new aliasing artifacts, it is more convenient to split
the panels in a given reproduction scene into the least
number of sections possible.
4 IMPROVEMENTS TO ENHANCE
REPRODUCTION
This section deals with improvements in multiactuator
panel technology to create a sound field of quality
comparable to that generated with conventional loud-
speakers.
4.1 Practical Edge Boundary Conditions
The multiactuator panel is a plate whose radiation is
caused by the superposition of mechanical waves.
Standing-wave patterns due to incident and reflected
Fig. 10. Analysis of sound field created by two multiactuator panels with different numbers of exciters per panel. (a)
Multiactuator panels with same exciter spacing but different framing. (b) Space–time graphical representation. (c) Desired and
aliased sound fields at an example frequency of 7 kHz. Dimensions in meters.
J. Audio Eng. Soc., Vol. 58, No. 12, 2010 December 1053
PAPERS MULTIACTUATOR PANELS FOR WFS
waves from the boundaries allow modal vibrations to
appear.
In order to achieve an even spatial distribution of
loudspeaker sources, these sources should not be too close
to the mechanical boundaries of the plates. Unfortunately
to extend the reproduction stage beyond the dimensions
of multiactuator panels several multiactuator panels must
be attached in landscape orientation. In [32] the radiation
of the first exciter, which was only 80 mm from the edge,
was studied in an eight-exciter multiactuator panel. The
exciter performance showed some alteration with respect
to the central ones, but the interference was smaller than
expected.
In the technical literature the response of flat panel
loudspeakers has been investigated using various edge-
supporting techniques [36], [37]. Both free and simply
supported edge options do not present a viable technology
for a multiactuator panel frame. The free condition is a
free radiating panel with dipole radiation characteristics.
However, when observing the typical polar response of a
free distributed-mode loudspeaker, a reduction of pres-
sure in the plane of the panel is noticed because of the
cancellation effect of acoustic radiation at or near the
edges. For that reason the panel is usually placed in a
baffle, where radiation caused by the rear part of the panel
becomes contained. Simply supported conditions would
need a large supporting frame, which may cause
diffracted radiation. This problem can be solved with
the help of a special type of mixed boundary condition, as
suggested in [38].
Alternatively, elastic boundaries are a compromise
between supported and free edges and can form a viable
technology for a multiactuator panel frame. These
boundary conditions have been addressed in [39] through
an experimental comparison between clamped, free, and
three types of elastic boundaries. For that purpose the
wave field generated by an arrangement of five multi-
actuator panels comprising three exciters per panel with
spacing of Dx¼0.18 m was analyzed in the space–time
domain discussed earlier. This array configuration was
chosen to enhance the effect that different boundaries
may have on the system. For details of this prototype refer
to Section 5.4. Fig. 11 shows the wave fields for two
stimuli under two edge boundaries: 1) a clamped
condition, which prevents the deflection of the structure
at the supports, and 2) an elastic boundary with foam at
two sides of the panel.
The first experiment deals with the edge boundary
effect on the radiation of a single centered exciter, as
shown in Fig. 11(a). Ideally a broad-band point source is
composed of several plane waves for different angles and
frequencies, which would be depicted as a pure black
triangle. The clamped condition presents a distinct
distribution of energy within the triangle. When listening
to such a wave field, annoying pressure variations would
be perceived both in frequency and on moving inside the
listening area. On the other hand elastic boundaries
exhibit a moderately homogeneous response in the space
time frequency response with more resemblance to the
ideal black triangle.
The second stimulus is an on-axis plane wave, which is
generated to force the area near the boundaries to vibrate
at maximum so that the potential reflections on the edges
would be visible. This stimulus is presented in Fig. 11 (b).
The clamped condition shows the appearance of two new
undesired replicas of the plane wave field between those
related to the transducer spacing, which was discussed in
Section 3.3. However, with elastic boundaries the impact
of edge boundary conditions on the reproduced replicas is
decreased, as depicted at right in Fig. 11(b).
These experiments confirmed the general trend of using
elastic materials in the prototype panels, which will be
discussed in Section 5.
4.2 Equalization
As stated previously, multiactuator panel radiation is
based on the excitation of natural modes of vibration.
However, there are moderate fluctuations, which are
caused by the vibration pattern of the panel. The
acoustical response is caused by a complex superposition
of the bending wave excitations in the panel and can vary
strongly, even in small frequency bands. As a conse-
quence the typical multiactuator panel response may be
very poor and often needs to be equalized for a more
natural and uncolored response.
In the context of wave field synthesis reproduction, two
filtering processes have been proposed to enhance the
reproduced spectrum and to remove the room effect in the
sound field. The first group comprises filter techniques
that compensate for uneven frequency responses. In
multiactuator panel terminology these filters are called
spectral smoothness equalization. A second group of
filters equalize and compensate for reflections of the
room. Multichannel inverse filtering techniques are used,
which have also been extended to account for panel
reflections.
In wave field synthesis the use of these two filtering
approaches depends on the aliasing frequency. Multi-
channel techniques are not applied above aliasing
frequency, because otherwise severe spatial-dependent
distortion of the wave field will be introduced. On the
contrary, above that frequency spectral smoothness
equalization is used [40].
4.2.1 Spectral Smoothness Equalization
The typical frequency response of a multiactuator panel
is highly irregular because of a superposition of
temporarily and spatially arriving diffuse waves. The
transient response of panels with distributed-mode
operation is characterized by a long impulse response
tail, which in the frequency domain means that modes
exhibit a narrow bandwidth [38]. Therefore to compen-
sate for the complex and random characteristics, some
sort of filtering is advisable.
An interesting method, which combines linear-predic-
tive coding inverse filtering for room mode correction
1054 J. Audio Eng. Soc., Vol. 58, No. 12, 2010 December
PUEO ET AL. PAPERS
with discrete wavelet transform octave bands for
distributed-mode loudspeaker equalization, is presented
in [41]. The method proved to be appropriate for
equalizing peaks in the response rather than dips. Another
effective filtering is presented in [42], in which an FIR
filter varies coefficients for active noise control. In
general the resulting filtered response reported in other
experiments [43], [40] was still not flat, but it exhibited a
more uniform shape.
In another study [44] a filtering method that comes at a
very low computational cost and makes use of an
advantageous IIR scheme is tested with different multi-
actuator panels. For that purpose on-axis measurements
are taken in front of every individual exciter of the panel.
An inverse filter scheme is used to compensate their
uneven frequency responses. Considering a given target
bandwidth and the amplitude of the irregularities on the
responses, a structure of second-order section (SOS)
filters is computed for every exciter. Afterward filters are
applied to the multiactuator panels, and the equalized
frequency responses are measured in the laboratory. As an
example, Fig. 12 shows the experimental frequency
responses before and after applying the equalization
process, as well as the computed filter responses for the
three-exciter multiactuator panel prototype of Section 5.4.
The graph shows that the smoothed responses have
Fig. 11. Influence of edge boundary conditions in multiactuator panel radiation. (a) Point source from central exciter. (b) Axial
plane wave by means of wave field synthesis. Left—clamped; right—elastic boundary.
J. Audio Eng. Soc., Vol. 58, No. 12, 2010 December 1055
PAPERS MULTIACTUATOR PANELS FOR WFS
been equalized and are now almost flat. In order to give
an objective quality measurement of the flatness of the
responses, the logarithmic standard deviation r
log
is used
here as well as other multiactuator panel analyses [40].
For a perfectly flat frequency response r
log
will be equal
to 0 dB, which is the lower limit. The higher the value, the
more irregular is the frequency response. The computed
r
log
values for the unfiltered responses were 4.22, 3.73,
and 4.39 dB, whereas those for equalized responses
dropped to 1.49, 1.26, and 1.32 dB for the left, central,
and right exciters, respectively.
The efficiency of the method facilitates the application
to a large multichannel multiactuator panel system with
affordable hardware processing. Although spectral filter-
ing is generally intended for frequencies above aliasing,
the efficiency of this method opens the possibility of
also applying equalization for low frequency as a
prefiltering process prior to the multichannel room
equalization. Also, subjective tests have confirmed that
it is possible to reduce the order of the filter consider-
ably without sacrificing the perceived quality of the
equalization [44].
4.2.2 Reflection and Room Compensation
When a sound field is reproduced inside a listening
room, there can be deterioration in the frequency response
and in the spatial sound impression of the listener because
of the room effect. To solve this problem, reflection and
room compensation techniques aiming at reducing the
influence of the listening room are used. Multichannel
inverse filtering techniques offer control of the sound field
at a limited number of points in the listening space [45]. A
method based partly on multichannel inversion, which
aimed at controlling the free-field radiation patterns of
multiactuator panels, was proposed by Corteel and
coworkers [26], [27]. This method was later extended to
room compensation for wave field synthesis [46], [47].
The solution produced by multichannel equalization
techniques achieved both the synthesis of the reproduc-
tion objective as well as the compensation of reproduction
artifacts in a unique process. The latter included the
suppression of undesired reflections between multiactua-
tor panels that might disturb the perception of the target
sound field.
A similar method for room compensation in wave field
synthesis was presented by Spors and coworkers [48],
[49]. This study relies on circular microphone array
measurements which are decomposed into plane wave or
cylindrical harmonics [50]. Multichannel inversion is then
performed in this transformed domain. The control
remains efficient inside the circular array, with which
the sound field is sampled, but may suffer from artifacts
linked to both microphone and rendering system limita-
tions [51].
Alternatively another approach based on multichannel
inversion of multiple-input and multiple-output systems
was proposed as a possible practical solution to room
compensation in an extended area for wave field synthesis
reproduction systems [52]–[54]. For that purpose the
reproduction area was sampled with multiple micro-
phones over a square inside the listening area. This
measurement resulted in a set of impulse responses
between each loudspeaker and 196 listening points
separated by 50 mm. The data are used to calculate a
bank of inversion filters, which are applied just before
each individual loudspeaker section. In order to evaluate
the performance of the room compensation algorithm,
three single frequency sources have been positioned
obliquely in a U-shaped wave field synthesis array of 32
loudspeakers separated by 180 mm.
Fig. 12. Frequency responses of equalization process and experimental results for multiactuator panel prototype comprising
three exciters. (From [44].) – – – Exciter measurement, – – computed filter response, ––– measured filtered response.
Representations are shifted 30 dB to maintain clarity. (a) Left exciter. (b) Central exciter. (c) Right exciter.
1056 J. Audio Eng. Soc., Vol. 58, No. 12, 2010 December
PUEO ET AL. PAPERS
Fig. 13 shows the field rendered by point sources of 0.8
Hz, 1.5 kHz, and 1.8 kHz in a sampling area of 0.65 m by
0.65 m, with the wave field synthesis array having an
aliasing frequency of 1 kHz. In this way, the algorithm is
tested below and slightly above the aliasing frequency of
the system. Fig. 13(a) depicts simulated responses
produced by sources in an anechoic environment. In
Fig. 13(b) the experimentally measured field is presented,
which diverges from the simulated field because of room
reflections. This situation is typical in real wave field
synthesis setups, in which the room acoustics modify the
synthesized wave field. Finally Fig. 13(c) illustrates the
result of applying the multichannel compensation algo-
rithm to the real room responses. As can be seen in Fig.
13, the responses are quite similar to the simulated
response, even for frequencies above the aliasing
frequency.
5 CURRENT MULTIACTUATOR PANEL
PROTOTYPES
As a result of the research undertaken for the
CARROUSO project and other research institutions,
which have been discussed in this paper, most prototypes
were designed and built for research purposes. The six
multiactuator panel designs that were used in the
experimental setups in the technical literature are shown
to scale in Fig. 14.
5.1 Laboratory of Acoustic Imaging and Sound
Control, TU Delft
At the Laboratory of Acoustic Imaging and Sound
Control at TU Delft Boone and de Vries experimented
with the primitive single-exciter panel and finally
developed a multiactuator panel design that would appear
in several publications [28], [25], [31], [43], [40], [11],
Fig. 13. Fields rendered by source signals of frequencies 0.8 Hz, 1.5 kHz, and 1.8 kHz. (From [53].) (a) Original source
reproduced by wave field synthesis in free field. (b) Field measured in a real room. (c) Field after applying compensation
algorithm.
Fig. 14. Geometric details on multiactuator panel prototypes developed by some CARROUSO partners. (a) TU Delft. (b)
University of Erlangen–Nuremberg. (c), (d) IRCAM and Sonic Emotion AG. (e), (f) University of Alicante and Technical
University of Valencia. Dimensions in centimeters.
J. Audio Eng. Soc., Vol. 58, No. 12, 2010 December 1057
PAPERS MULTIACTUATOR PANELS FOR WFS
[32]. Fig. 14(a) depicts the geometrical details of the
design. It is a medium-sized panel of 1.34 m by 0.75 m
with eight exciters spaced 167.5 mm apart. This spacing
is an integer fraction of the panel width in order to allow
panels to be put next to each other in landscape
orientation. With such a configuration the first and last
exciters on a panel are attached in such a way as to have a
continuous evenly spaced distribution of the exciters. As
will be shown later, this technique is also used in other
multiactuator panel designs. The line of exciters is
slightly below the center of the panel to increase modal
excitation.
The panel is made from 5-mm PVC foam board core
with paper skins, which is sufficiently damped to let
exciters generate the main contribution of energy in a
region of approximately 100 mm in diameter. As
suggested in previous contributions, FIR filtering was
still necessary to smooth the acoustic response. The
selected exciters for this design were ELAC 82073, an 8-
Xexciter with a 37-mm diameter and a 20-W nominal
power output. Fig. 15 shows this multiactuator panel
during a measurement session in the anechoic room.
5.2 Laboratory of Multimedia Communications
and Signal Processing, University of Erlangen–
Nuremberg
Rabenstein and others at the Laboratory of Multimedia
Communications and Signal Processing at the University
of Erlangen–Nuremberg developed an eight-exciter multi-
actuator panel, although in the literature it was called
multi-exciter panel (MEP) [55], [56]. This model is
virtually the same configuration as that in Delft, the
difference being a slightly larger spacing of the exciters,
as can be seen in Fig. 14(b). It is a medium-sized panel of
1.376 m by 0.746 m with eight exciters spaced 171 mm
apart. Since the transducer spacing is slightly larger, so
should be the width of the panel to maintain the
functionality of attaching multiactuator panels in land-
scape orientation. For this design the panel material was
also PVC foam, and ELAC 82073 type J exciters were
also used. Fig. 16 presents a group of four of these
multiactuator panels.
5.3 IRCAM Paris and Sonic Emotion AG
The collaboration carried out by Corteel at IRCAM in
Paris and Pellegrini at Sonic Emotion AG in Switzerland
led to the manufacture of two multiactuator panel designs,
as shown in Fig. 14(c) and (d). The latter has appeared in
recent publications [33], [57]. Contrary to the preceding
models, these panels are commercialized by Sonic
Emotion under the name M3S Panels and have similar
transducer spacings and exciter models. The innovation
with respect to the latter designs is the use of high aspect
ratios for the panel of Fig. 14(c) and the introduction of
small panels with four exciters, shown in Fig. 14(d). In
addition to this, such panels are made of laminated
honeycomb paper, which is more efficient than foam. An
Fig. 15. Multiactuator panel prototype developed at Laboratory of Acoustical Imaging and Sound Control, Delft University of
Technology, The Netherlands.
Fig. 16. Multiactuator panel prototype developed at Laboratory of Multimedia Communications and Signal Processing,
University of Erlangen–Nuremberg, Germany.
1058 J. Audio Eng. Soc., Vol. 58, No. 12, 2010 December
PUEO ET AL. PAPERS
arrangement of eight panels in a 32-channel configuration
is shown in Fig. 17.
5.4 University of Alicante and Technical
University of Valencia
Under the supervision of J. J. Lo
´pez the laboratories of
the University of Alicante and the Technical University of
Valencia in Spain jointly designed and built two multi-
actuator panel prototypes, depicted to scale in Fig. 14(e)
and (f). The performance of these two prototypes has been
discussed in several publications [58], [59], [34], [39],
[44], [60].
The two prototypes were designed to be arranged side
by side in a landscape orientation, as the multiactuator
panels mentioned before. The exciter spacing was set to
180 mm for both panels, thus enabling an aliasing
frequency of approximately 1 kHz. As shown in Fig.
14(e), the first prototype is a medium-size multiactuator
panel containing five exciters on a 0.90-m by 0.795-m
panel, and the second is a scaled version that only
incorporates three exciters. The panel of both prototypes
is a sandwich of polyester film bonded to an impregnated
paper honeycomb 5 mm thick using a thermoplastic
adhesive (cell size 4.8 mm). The exciters were Peerless
880101 25-mm-diameter dynamic transducers which
couple to the panel with eight small points per transducer.
Fig. 18 presents a group of three of these medium-size
multiactuator panels in a linear configuration.
6 CONCLUSIONS
The main aspects of multiactuator panels as a
reproduction system for wave field synthesis have been
discussed by revisiting some of the most relevant works
published on this subject. The evolution of the single-
excited distributed-mode loudspeaker to the current
multiactuator panels has been presented in sequence,
through bibliographic citations of literature from the
respective fields of audio engineering.
Multiactuator panels can be used alternatively to
dynamic loudspeaker arrays for wave field synthesis with
added benefits. Due to their low visual profile, they can be
integrated into the listening environment as part of the
room decoration. Furthermore the radiation of multi-
actuator panels is evenly distributed across the listening
area and over a wider frequency band. This characteristic
helps in merging the individual secondary sources of
wave field synthesis correctly for a large listening area. In
addition when multiactuator panels are used in non-
anechoic wave field synthesis setups, room reflections are
less correlated to the direct sound than those radiated
from pistonlike loudspeakers, and thus constructive and
destructive interference of sound is minimized. As a
consequence the localization precision is not degraded by
room acoustics as with dynamic loudspeaker arrays. The
proper sound localization and the minimum degradation
provided by multiactuator panels are important for wave
field synthesis installations.
However, since multiactuator panels are panels of finite
extent, excited mechanically on several points, there were
structural and geometric issues that had to be addressed.
Therefore the structural and acoustical behavior of
multiactuator panels was studied in the years following
their introduction. It was then confirmed that a single
panel excited in many points behaves like independent
loudspeakers in an array.
As a result of this research some prototypes were
designed and built by laboratories at universities and
companies. The research and development of multi-
actuator panels advanced to a degree of commercializa-
Fig. 17. Multiactuator panel prototype developed at IRCAM, Paris, France, and Sonic Emotion AG, Oberglatt, Switzerland.
Fig. 18. Multiactuator panel prototype developed at University of Alicante and Technical University of Valencia, Spain.
J. Audio Eng. Soc., Vol. 58, No. 12, 2010 December 1059
PAPERS MULTIACTUATOR PANELS FOR WFS
tion. Thanks to the technical developments achieved
during their evolution, multiactuator panels are today a
valid alternative to dynamic loudspeaker arrays for wave
field synthesis.
7 ACKNOWLEDGMENT
This work was supported by the Spanish Ministry of
Science and Technology under project TEC2009-14414-
C03-01, by the University of Alicante under project
GRE09-33 and FEDER funds, and by the Dutch
Technology Foundation STW, Applied Science Division
of NWO, and the Technology Program of the Ministry of
Economic Affairs.
The authors would like to thank Matthijs Ruoff for his
help with the near-field acoustic holography measurements
and Diemer de Vries, who provided insight and expertise
that greatly assisted the improvement of this manuscript.
8 REFERENCES
[1] J. Blauert, Spatial Hearing. The Psychophysics of
Human Sound Localization (MIT Press, Cambridge, NH,
1996).
[2] R. Begault, 3-D Sound for Virtual Reality and
Multimedia (Academic Press Professional, Boston, MA,
1994).
[3] M. Gerzon, ‘‘Periphony: With-Height Sound
Reproduction,’’ J. Audio Eng. Soc., vol. 21, pp. 2–10
(1973 Dec.).
[4] A. Berkhout, D. de Vries, and P. Vogel, ‘‘Acoustic
Control by Wave Field Synthesis,’’ J. Acoust. Soc. Am.,
vol. 93, pp. 2764–2778 (1993 May).
[5] D. de Vries, ‘‘Wave Field Synthesis,’’ Monographs
Audio Engineering Society, New York (2009).
[6] A. J. Berkhout, ‘‘A Holographic Approach to
Acoustic Control,’’ J. Audio Eng. Soc.(Engineering
Reports), vol. 36, pp. 977–995 (1988 Dec.).
[7] A. Berkhout, Applied Seismic Wave Theory
(Elsevier, Amsterdam, The Netherlands, 1987).
[8] D. de Vries, ‘‘Sound Reinforcement by Wavefield
Synthesis: Adaptation of the Synthesis Operator to the
Loudspeaker Directivity Characteristics,’’ J. Audio Eng.
Soc., vol. 44, pp. 1120–1131 (1996 Dec.).
[9] E. Start, ‘‘Direct Sound Enhancement by Wave
Field Synthesis,’’ Ph.D. dissertation, Delft University of
Technology, Delft, The Netherlands (1997).
[10] S. Spors and R. Rabenstein, ‘‘Spatial Aliasing
Artifacts Produced by Linear and Circular Loudspeaker
Arrays Used for Wave Field Synthesis,’’ presented at the
120th Convention of the Audio Engineering Society, J.
Audio Eng. Soc. (Abstracts), vol. 54, p. 689 (2006 July/
Aug.), convention paper 6716.
[11] E. Corteel, ‘‘Equalization in an Extended Area
Using Multichannel Inversion and Wave Field Synthesis,’’
J. Audio Eng. Soc., vol. 54, pp. 1140–1161 (2006 Dec.).
[12] S. Spors, ‘‘Investigation of Spatial Aliasing
Artifacts of Wave Field Synthesis in the Temporal
Domain,’’ in Proc. Ann. German Conf. on Acoustics
(DAGA) (2008 Mar.).
[13]N.HarrisandM.O.J.Hawksford,‘‘The
Distributed-Mode Loudspeaker (DML) as a Broad-Band
Acoustic Radiator,’’ presented at the 103rd Convention of
the Audio Engineering Society, J. Audio Eng. Soc.
(Abstracts), vol. 45, p. 1007 (1997 Nov.), preprint 4526.
[14] V. P. Gontcharov and N. P. R. Hill, ‘‘Diffusivity
Properties of Distributed Mode Loudspeakers,’’ presented
at the 108th Convention of the Audio Engineering
Society, J. Audio Eng. Soc. (Abstracts), vol. 48, p. 350
(2000 Apr.), preprint 5095.
[15] L. Cremer, M. Heckl, and B. A. Petersson,
Structure Borne Sound: Structural Vibrations and Sound
Radiation at Audio Frequencies, 3rd ed. (Springer, New
York, 2005).
[16] E. Y. Prokofieva, K. V. Horoshenkov, and N.
Harris, ‘‘The Acoustic Emission of a Distributed Mode
Loudspeaker Near a Porous Layer,’’ J. Acoust. Soc. Am.,
vol. 111, pp. 2665–2670 (2002 June).
[17] E. Prokofieva, ‘‘Effect of Porous Material on the
Diffusivity of an Unbaffled DML Panel,’’ presented at the
114th Convention of the Audio Engineering Society, J.
Audio Eng. Soc. (Abstracts), vol. 51, p. 410 (2003 May),
convention paper 5735.
[18] H. Azima and N. Harris, ‘‘Boundary Interaction of
Diffuse Field Distributed-Mode Radiators,’’ presented at
the 103rd Convention of the Audio Engineering Society,
J. Audio Eng. Soc. (Abstracts), vol. 45, p. 1020 (1997
Nov.), preprint 4635.
[19] B. Fazenda, M. Avis, and J. Davies, ‘‘Low
Frequency Room Excitation Using Distributed Mode
Loudspeakers,’’ in Proc. AES 21st Int. Conf. on
Architectural Acoustics and Sound Reinforcement (St.
Petersburg, Russia, 2002 June 1–3).
[20] N. Harris, V. Gontcharov, and M. O. J. Hawksford,
‘‘Measurement and Simulation Results Comparing the
Acoustics of Various Direct Radiators in the Presence of a
Dominant Specular Reflection,’’ presented at the 109th
Convention of the Audio Engineering Society, J. Audio
Eng. Soc. (Abstracts), vol. 48, p. 1104 (2000 Nov.),
preprint 5215.
[21] N. Harris and M. O. J. Hawksford, ‘‘Modeling
Room Interaction Effects for Pistonic and Distributed-
Mode Loudspeakers in Both the Frequency and Time
Domains,’’ presented at the 114th Convention of the Audio
Engineering Society, J. Audio Eng. Soc. (Abstracts), vol.
51, pp. 409, 440 (2003 May), convention paper 5732.
[22] N. Harris, ‘‘Modeling Acoustic Room Interaction
for Pistonic and Distributed-Mode Loudspeakers in the
Correlation Domain,’’ presented at the 117th Convention
of the Audio Engineering Society, J. Audio Eng. Soc.
(Abstracts), vol. 53, p. 80 (2005 Jan./Feb.), convention
paper 6180.
[23] N. Harris, S. Flanagan, and M. O. J. Hawksford,
‘‘Stereophonic Localization in Rooms, Comparing Con-
ventional and Distributed-Mode Loudspeakers,’’ present-
1060 J. Audio Eng. Soc., Vol. 58, No. 12, 2010 December
PUEO ET AL. PAPERS
ed at the 105th Convention of the Audio Engineering
Society, J. Audio Eng. Soc. (Abstracts), vol. 46, p. 1034
(1998 Nov.), preprint 4794.
[24] M. M. Boone and W. P. J. de Brujin, ‘‘On the
Applicability of Distributed Mode Loudspeaker Panels
for Wave Field Synthesis–Based Sound Reproduction,’’
presented at the 108th Convention of the Audio
Engineering Society, J. Audio Eng. Soc. (Abstracts),
vol. 48, p. 364 (2000 Apr.), preprint 5165.
[25] W. van Rooijen, ‘‘Distributed Mode Loudspeakes
for Wave Field Synthesis,’’ Master’s thesis, Faculty of
Applied Science, Technical University Delft, Delft, The
Netherlands (2001).
[26] E. Corteel, U. Horbach, and R. Pellegrini,
‘‘Multichannel Inverse Filtering of Multiexciter Distrib-
uted Mode Loudspeakers for Wave Field Synthesis,’’
presented at the 112th Convention of the Audio
Engineering Society, J. Audio Eng. Soc. (Abstracts),
vol. 51, p. 511 (2002 June), convention paper 5611.
[27] U. Horbach, E. Corteel, and D. de Vries, ‘‘Spatial
Audio Reproduction Using Distributed Mode Loudspeak-
er Arrays,’’ in Proc. AES 21st Int. Conf. on Architectural
Acoustics and Sound Reinforcement (St. Petersburg,
Russia, 2002 June 1–3).
[28] M. M. Boone, W. P. J. de Bruijn, and W. van
Rooijen, ‘‘Recent Developments on WFS for High-
Quality Spatial Sound Reproduction,’’ presented at the
110th Convention of the Audio Engineering Society, J.
Audio Eng. Soc. (Abstracts), vol. 49, pp. 539, 540 (2001
June), convention paper 5370.
[29] ‘‘CARROUSO—Creating, Assessing and Render-
ing in Real Time of High Quality Audio-Visual
Environments in MPEG-4 Context,’’ (Information Society
Technologies, European Union, (2001), http://cordis.
europa.eu/ist/ka3/iaf/projects/carrouso.htm (last seen
2008 May).
[30] S. Brix, T. Sporer, and J. Plogsties, ‘‘CARROU-
SO—A European Approach to 3D Audio,’’ presented at
the 110th Convention of the Audio Engineering Society,
J. Audio Eng. Soc. (Abstracts), vol. 49, p. 528 (2001
June), convention paper 5314.
[31] M. M. Boone, ‘‘Multi-Actuator Panels (MAPs) as
Loudspeaker Arrays for Wave Field Synthesis,’’ J. Audio
Eng. Soc., vol. 52, pp. 712–723 (2004 July/Aug.).
[32] M. Kuster, D. de Vries, D. Beer, and S. Brix,
‘‘Structural and Acoustic Analysis of Multiactuator
Panels,’’ J. Audio Eng. Soc., vol. 54, pp. 1065–1076
(2006 Nov.).
[33] E. Corteel, K. V. Nguyen, O. Warusfel, T.
Caulkins, and R. Pellegrini, ‘‘Objective and Subjective
Comparison of Electrodynamic and MAP Loudspeakers
for Wave Field Synthesis,’’ in Proc. AES 30th Int. Conf.
on Intelligent Audio Environments (Saariselka
¨, Finland,
2007 Mar. 15–17).
[34] B. Pueo, J. J. Lo
´pez, J. Escolano, and S. Bleda,
‘‘Analysis of Multiactuator Panels in the Space–Time
Wavenumber Domain,’’ J. Audio Eng. Soc., vol. 55, pp.
1092–1106 (2007 Dec.).
[35] E. G. Williams, Fourier Acoustics. Sound
Radiation and Nearfield Acoustic Holography (Academic
Press, London, 1999).
[36] E. Prokofieva, K. V. Horoshenkov, and N. Harris,
‘‘Intensity Measurements of the Acoustic Emission from a
DML Panel,’’ presented at the 112th Convention of the
Audio Engineering Society, J. Audio Eng. Soc. (Ab-
stracts), vol. 50, p. 511 (2002 June), convention paper
5609.
[37] M. R. Bai and T. Huang, ‘‘Development of Panel
Loudspeaker System: Design, Evaluation and Enhance-
ment,’’ J. Acoust. Soc. Am., vol. 109, pp. 2751–2761
(2001 June).
[38] J. A. S. Angus, ‘‘Distributed Mode Loudspeaker
Resonance Structures,’’ presented at the 109th Conven-
tion of the Audio Engineering Society, J. Audio Eng. Soc.
(Abstracts), vol. 48, p. 1105 (2000 Nov.), preprint 5217.
[39] B. Pueo, J. J. Lo
´pez, and J. Escolano, ‘‘Edge
Boundary Conditions Impact on the Radiation of Multi-
actuator Panels for Multichannel Audio Reproduction,’’
Acta Acustica, vol. 94, pp. 754–764, (2008 Sept/Oct.).
[40] J. van Dorp and D. de Vries, ‘‘Wave Field
Synthesis Using Multi-Actuator Panel: Further Steps to
Optimal Performance,’’ in Proc. AES 28th Int. Conf. on
the Future of Audio Technology—Surround and Beyond
(Pitea
˚, Sweden, 2006 June 20–July 2).
[41] K. Chiao, N. Harris, and C. Kyriakakis, ‘‘A New
Approach to Speaker/Room Equalization,’’ presented at
the 109th Convention of the Audio Engineering Society,
J. Audio Eng. Soc. (Abstracts), vol. 48, p. 1105 (2000
Nov.), preprint 5221.
[42] H. Zhu, R. Rajamani, J. Dudney, and K. Stelson,
‘‘Active Noise Control Using a Distributed Mode Flat
Panel Loudspeaker,’’ ISA Trans., vol. 42, pp. 475–484
(2003 July).
[43] J. van Dorp, ‘‘Wave Field Synthesis Using Multi
Actuator Panel Loudspeakers: Design and Application of
Various Digital Filtering Algorithms,’’ Master’s thesis,
Faculty of Applied Science, TU Delft, The Netherlands
(2005).
[44] B. Pueo, J. J. Lo
´pez, G. Ramos, and J. Escolano,
‘‘Efficient Equalization of Multi-Exciter Distributed
Mode Loudspeakers,’’ Appl. Acoust., vol. 70, pp. 737–
746 (2009 May).
[45] P. A. Nelson, F. Ordun
˜a-Bustamante, and H.
Hamada, ‘‘Multichannel Signal Processing Techniques in
the Reproduction of Sound,’’ J. Audio Eng. Soc., vol. 44,
pp. 973–989 (1996 Nov.).
[46] E. Corteel and R. Nicol, ‘‘Listening Room
Compensation for Wave Field Synthesis. What Can Be
Done?’’ in Proc. AES 23th Int. Conf. on Signal Processing
in Audio Recording and Reproduction (Copenhagen,
Denmark, 2003 May 23–25).
[47] R. van Zon, E. Corteel, D. de Vries, and O.
Warusfel, ‘‘Multiactuator Panel (MAP) Loudspeakers:
How to Compensate for Their Mutual Reflections?’’
presented at the 116th Convention of the Audio
J. Audio Eng. Soc., Vol. 58, No. 12, 2010 December 1061
PAPERS MULTIACTUATOR PANELS FOR WFS
Engineering Society, J. Audio Eng. Soc. (Abstracts), vol.
52, p. 795 (2004 July/Aug.), convention paper 6052.
[48] S. Spors, A. Kuntz, and R. Rabenstein, ‘‘Approach
to Listening Room Compensation with Wave Field
Synthesis,’’ in Proc. AES 24th Int. Conf. on Multichannel
Audio (Banff, Alberta, Canada, 2003 June 26–28).
[49] S. Spors, H. Buchner, and R. Rabenstein,
‘‘Efficient Active Listening Room Compensation for
Wave Field Synthesis,’’ presented at the 116th Conven-
tion of the Audio Engineering Society, J. Audio Eng. Soc.
(Abstracts), vol. 52, p. 812 (2004 July/Aug.), convention
paper 6119.
[50] E. Hulsebos, D. de Vries, and E. Bourdillat,
‘‘Improved Microphone Array Configurations for Aural-
ization of Sound Fields by Wave-Field Synthesis,’’ J.
Audio Eng. Soc., vol. 50, pp. 779–790 (2002 Oct.).
[51] S. Spors, M. Renk, and R. Rabenstein, ‘‘Limiting
Effects of Active Room Compensation Using Wave Field
Synthesis,’’ presented at the 118th Convention of the
Audio Engineering Society, J. Audio Eng. Soc. (Ab-
stracts), vol. 53, p. 681 (2005 July/Aug.), convention
paper 6400.
[52] J. J. Lo
´pez, A. Gonza
´lez, and L. Fuster, ‘‘Room
Compensation in Wave Field Synthesis by Means of
Multichannel Inversion,’’ in Proc. IEEE Workshop on the
Application of Signal Processing to Audio and Acoustics
(WASPAA’05) (New Paltz, NY, 2005 Oct.).
[53] L. Fuster, J. J. Lo
´pez, A. Gonza
´lez, and P. D.
Zuccarello, ‘‘Room Compensation Using Multichannel
Inverse Filters for Wave Field Synthesis Systems,’’
presented at the 118th Convention of the Audio
Engineering Society, J. Audio Eng. Soc. (Abstracts),
vol. 53, p. 681 (2005 July/Aug.), convention paper 6401.
[54] L. Fuster, J. J. Lo
´pez, A. Gonza
´lez, and P. Faus,
‘‘Time and Frequency Domain Room Compensation
Applied to Wave Field Synthesis,’’ in Proc. 8th Int. Conf.
on Digital Audio Effects (DAFx05) (Madrid, Spain, 2005
Sept.).
[55] D. Seuberth, ‘‘Kompensation von Flachlautspre-
chern fu¨r ein Wellenfeldsynthesesystem,’’ Master’s thesis,
Multimedia Communications and Signal Processing,
University of Erlangen–Nuremberg, Germany (2003).
[56] S. Spors, D. Seuberth, and R. Rabenstein, ‘‘Multi-
Exciter Panel Compensation for Wave Field Synthesis,’’
in Proc. Ann. German Conf. on Acoustics (DAGA) (2005
Mar.).
[57] E. Corteel, ‘‘Synthesis of Directional Sources
Using Wave Field Synthesis, Possibilities, and Limita-
tions,’’ EURASIP J. Appl. Signal Process., vol. 2007, p.
18 (2007).
[58] B. Pueo, J. Escolano, S. Bleda, and J. J. Lo
´pez,
‘‘An Approach for Wave Field Synthesis High Power
Applications,’’ presented at the 118th Convention of the
Audio Engineering Society, J. Audio Eng. Soc. (Ab-
stracts), vol. 53, pp. 688, 689 (2005 July/Aug.),
convention paper 6428.
[59] B. Pueo, S. Bleda, J. Escolano, and J. J. Lo
´pez,
‘‘Analysis of Spatial Resolution of Multiactuator Panels,’’
presented at the 120th Convention of the Audio
Engineering Society, J. Audio Eng. Soc. (Abstracts),
vol. 54, p. 695 (2006 July/Aug.), convention paper 6733.
[60] B. Pueo, ‘‘Analysis and Enhancements of Multi-
actuator Panels for Wave Field Synthesis Reproduction,’’
Ph.D. dissertation, Technical University of Valencia,
Valencia, Spain (2008).
THE AUTHORS
B. Pueo J. J. Lo
´pez J. Escolano L. Ho
¨rchens
Basilio Pueo was born in Spain in 1973. He received
B.S. and M.S. degrees in telecommunications engineering
in 2004 and 2008, respectively. In 1998 he received a
Ph.D. degree in telecommunications engineering from the
Technical University of Valencia, Spain.
Since 1999 he has held a teaching position at the
University of Alicante, Spain. His main research interests
include multichannel audio techniques and wave field
synthesis, with emphasis on the application of audiovisual
technology to the media.
Dr. Pueo has published about 30 papers at international
conferences and in journals. He was papers cochair of the
118th Convention of the Audio Engineering Society in
Barcelona and treasurer of the AES Spanish Section in
2006. He is currently the treasurer of the AES Spanish
Section.
1062 J. Audio Eng. Soc., Vol. 58, No. 12, 2010 December
PUEO ET AL. PAPERS
Jose
´Javier Lo
´pez was born in Valencia, Spain, in 1969.
He received a telecommunications engineering degree in
1992 and a Ph.D. degree in 1999, both from the Technical
University of Valencia.
Since 1993 he has been involved in education and
research in the Communications Department of the
Technical University of Valencia, where at present he is
a professor, teaching undergraduate and graduate courses
on signal processing and digital audio processing. His
current research activity is centered on digital audio
processing in the areas of three-dimensional sound, wave
field synthesis, acoustic spaces modeling, efficient
filtering structures for loudspeaker correction, algorith-
mics in signal processing, and development of multime-
dia software in real time.
Dr. Lo
´pez has published over 100 papers in interna-
tional technical journals and at renowned conferences in
the fields of audio and acoustics and has led several
research projects. He was workshop cochair at the 118th
Convention of the Audio Engineering Society in Barce-
lona and has served on the committee of the AES Spanish
Section for 5 years, at present being its secretary.
X
Jose
´Escolano was born in Valladolid, Spain, in 1977.
He received Bachelor’s and Master’s degrees in telecom-
munications engineering in 2002 and 2004, respectively.
In 2008 he received a Ph.D. degree from the Technical
University of Valencia, Spain.
Since 2006 he has been an assistant professor at the
University of Jaen, Jaen, Spain. In 2004 he was a visiting
researcher in the Multimedia Communications and Signal
Processing Group, Erlangen–Nuremberg University,
Germany, for three months, and during the period of
2005–2007, he was with the Acoustic Technology Group
of the Technical University of Denmark, Copenhagen,
partially on a Marie Curie grant from the EC. During the
summers of 2009 and 2010 he joined the Architectural
Acoustics Group at Rensselaer Polytechnic Institute,
Troy, NY, as visiting postdoctoral fellow. His main
research interests focus on room acoustic simulation,
numerical modeling, multichannel audio techniques, and
electroacoustic transduction.
Dr. Escolano has published about 50 papers at
international conferences and in professional journals.
He is a member of AES, ASA, and IEEE.
X
Lars Ho
¨rchens was born in Germany in 1979. He
received a diploma in media technology from Ilmenau
University of Technology, Germany, in 2005. Subse-
quently he joined the Laboratory of Acoustical Imaging
and Sound Control at Delft University of Technology,
The Netherlands, where he received a Ph.D. degree in
2010.
He is currently pursuing postdoctoral research in the
same group. His research interests include inverse
problems, such as acoustical imaging, and the analysis
of flexural wave fields with applications in nondestructive
testing, loudspeaker design, and building acoustics.
Dr. Ho
¨rchens is a committee member of the AES
Netherlands Section.
J. Audio Eng. Soc., Vol. 58, No. 12, 2010 December 1063
PAPERS MULTIACTUATOR PANELS FOR WFS
2010 JCR Science Edition
Rank in Category: JOURNAL OF THE AUDIO ENGINEERING SOCIETY
Journal Ranking
For 2010, the journal JOURNAL OF THE AUDIO ENGINEERING SOCIETY has an Impact Factor of 0.483.
This table shows the ranking of this journal in its subject categories based on Impact Factor.
Category Name Total Journals
in Category Journal Rank
in Category Quartile
in Category
ACOUSTICS 30 23 Q4
ENGINEERING, MULTIDISCIPLINARY 87 51 Q3
Category Box Plot
For 2010, the journal JOURNAL OF THE AUDIO ENGINEERING SOCIETY has an Impact Factor of 0.483.
This is a box plot of the subject category or categories to which the journal has been assigned. It provides information
about the distribution of journals based on Impact Factor values. It shows median, 25th and 75th percentiles, and the
extreme values of the distribution.
Key
A - ACOUSTICS
B - ENGINEERING,
MULTIDISCIPLINARY
Acceptable Use Policy
Copyright © 2014 Thomson Reuters.
... Par conséquent, une nouvelle technologie utilisant plusieurs actionneurs et appelée Multi-Actuator Panels (MAP) ou dans le cas de plaques plus grandes, Large Multi-Actuator Panels (LaMAP) aété conçue pour surmonter ces limitations [22]. Cette technologie aété utilisée avec succès dans le contexte de la WFS [6,5,2]. Toutefois, cette technologie présente encore certains inconvénients, tels que le comportement modal aux basses fréquences et les interférences entre les sources, ce qui entraîne des distorsions et donc des problèmes de qualité sonore. ...
... De plus, dans [22], uneétude complète de la réponse temporelle en vitesse du panneau est donnée, avec l'excitation de deux actionneurs différents, ainsi qu'une modélisation mécanique du panneau. Ces résultats sont repris dans [2], et réexpliqués avec de nouveaux essais. Il en résulte que d'après figure 2.18, le comportement temporel est cohérent avec celui expliqué juste au-dessus. ...
... En effet, toujours dans [2], l'auteur reprend encore les travaux de [22] pour les calculs de directivité. La directivité est mesurée en deux temps. ...
Thesis
L’objectif de cette thèse est de proposer et de valider des solutions alternatives aux haut-parleurs électrodynamiques classiquement utilisés dans l’automobile qui soient compatibles avec le rendu sonore spatialisé. L’idée directrice est de faire directement vibrer et rayonner des garnitures intérieures équipées d’actionneurs (électrodynamiques ou piézoélectriques) en y focalisant des ondes de flexion. Trois méthodes de focalisation de la littérature ont ainsi été formalisés dans un cadre commun et adaptées aux applications audio. Une étude paramétrique a été menée pour les comparer numériquement et expérimentalement à l’aide d’indicateurs de performances dédiés. La méthode de focalisation la plus efficace est l’inversion spatio-temporelle de l’opérateur de propagation. Elle nécessite un apprentissage préalable de la dynamique de la structure expérimentalement couteux. Une validation expérimentale des capacités de focalisation de cette méthode a été effectuée sur une portière de voiture. L’apprentissage par le biais d’un jumeau numérique au lieu de données expérimentales a de plus été validé. Les variations de température entre 10°C et 60°C existantes dans un habitacle de voiture ont ensuite été considérées. Les propriétés dynamiques de la structure hôte (constituée de polypropylène en général) varient en effet largement dans cette plage de température. Une stratégie de compensation de la température basée sur un jumeau numérique entrainant un réseau de neurones a été mise en place et validée numériquement. De plus, les bruits et vibrations dus au moteur et à la route perturbent également la focalisation. Une commande en boucle fermée a été mise en place afin de garantir les performances de l’algorithme de focalisation tout en éliminant ces perturbations.
... Concurrent to this emerging interest for audio technologies on the part of flight vehicle manufacturers that are concerned by the sound quality and sound environment of their products, the audio community is years in advance for the application of sound field reproduction, multichannel equalization, and panel actuators [6][7][8][9][10][11][12][13][14][15][16][17]. In this context this paper presents an illustrative application of audio technologies to an emerging field of research and development that could represent a promising expansion for the audio community. ...
... Although this represents an innovation, it also brings specific constraints and difficulties. Namely, the use of the basic trim panels as sound reproduction sources makes the task more difficult because they are not intended for sound reproduction and they are not ideal panel loudspeakers [15]. To reduce the risk of audible squeaks and rattles, we introduced a bass management system so that trim panels do not receive low-frequency content [18,19]. ...
Article
Sound environment reproduction of various flight conditions in an aircraft mock-up is a valuable tool for the study, prediction, demonstration, and jury testing of interior aircraft sound quality. To provide a faithful reproduced sound environment, time, frequency, and spatial characteristics should be preserved as much as possible. Physical sound field reproduction approaches for spatial sound reproduction are mandatory to immerse the listener's body in the proper sound field so that localization cues are recreated. For sound field reproduction inside cabin mock-up, the reported approach relies on multichannel equalization using least-meansquare formulation. In this paper a modified multichannel equalization procedure is proposed to simplify the selection of the regularization amount. The paper presents physical evaluations of reproduced sound fields on the basis of real flight recordings using an 80-channel microphone array and 41 reproduction sources in the cabin mock-up. Physical evaluations of reproduced spatial sound distributions are also provided on the basis of acoustic source maps obtained from acoustic imaging.
... It consists in driving several audio exciters bonded on the MAP with different audio signals. Such technologies have been successfully used in the context of WFS [5,6] and thus have shown their ability to provide satisfying results for spatial audio rendering purposes. However, it still suffer from the modal behavior of the plate and the cross talk cancellation of the audio exciters. ...
Article
Advanced automotive audio applications are more and more demanding with respect to the visual impact of loudspeakers while still requiring more and more channels for high quality spatial audio rendering. Removing classical heavy and large electrodynamic loudspeakers and using car interior plate-like structures driven by state of the art spatial sound algorithms appear as a promising solution to tackle both issues. However, to meet spatial audio rendering constraints, the bending waves generated within car interior plate-like structures must be focused at a given position and to a certain extent within the host structure. Theoretically, this means being able to invert in a robust manner the spatio-temporal wave propagation operator for the generated bending waves to fit a given target shape. The propagation operator inversion method considered here is the spatio-temporal inverse filtering (STIF) method based on the knowledge of the propagation operator on a regular spatial grid over the structure at a given temperature. However, in a car interior a high temperature variations exist and can adversely impact the performances of the STIF method, mainly because dynamical properties of the host structure (built up with polypropylene in most cases) largely vary within this temperature range. Even if the STIF method has already been adapted and assessed in the context of automotive audio reproduction, no study dealing with the effects of temperature on the STIF method and providing potential mitigation procedures avoiding experimental measurements at each temperature has been reported. To address that issue, the influence of temperature on the behavior of a polypropylene plate is first experimentally quantified. A model updating method is used to build a finite-element model of the plate taking into account temperature effects. This digital twin of the host-structure is then used to assess the influence of the temperature on the STIF method. A neural network based controller is finally trained and validated on the digital twin in order to compensate for the effects of temperature on STIF filters. Obtained results demonstrate that this procedure successfully allows to compensate for temperature effects on the STIF method applied to polypropylene plate with very limited experimental needs, thus paving the way through an industrial development of such approaches.
... Jeon et al. [90] used the localization factor described in [37] to optimally place exciters on MAPs for beamforming. A review of MAPs for WFS was published in 2010 [91]. ...
Article
Full-text available
The underlying physics and the design of loudspeakers that radiate sound through the bending vibrations of elastic panels, here referred to generically as flat-panel loudspeakers, are reviewed in this paper. The form factor, reduced weight, and aesthetic appeal of flat-panel speakers have made them a topic of interest for more than 90 years, but these advantages have been overshadowed by acoustical shortcomings, specifically the uneven frequency response and directivity in comparison to conventional cone-radiator loudspeakers. Fundamentally, the design challenges of flat-panel speakers arise from the intrinsically large number of mechanical degrees of freedom of a panel radiator.Anumber of methods have been explored to compensate for the acoustical shortcomings of flat-panel speakers, such as employing inverse filters, equalization, canceling mechanical resonances with actuator arrays, and modifying the panel material, shape, structure, and boundary conditions. Such methods have been used in various combinations to achieve significant audio performance improvements, and carefully designed flat-panel loudspeakers have been rated in blind listening tests as competitive with some prosumer-grade conventional loudspeakers. This review presents a brief historical account of the evolution of flat-panel loudspeakers and summarizes the essential physics and design methodologies that have been developed to optimize their fidelity and directional response.
... Of course, covering a complete living room, concert hall, theater or cinema with a surrounding loudspeaker array is a drastic impairment of the interior decoration. To solve this problem, research currently goes in two directions: The first concept is to apply multiactuator panels (MAPs) [24,27,61]. Due to their flat, inconspicuous nature, they do not harm interior decoration as much as conventional loudspeakers. ...
Chapter
Wave field synthesis enables acoustic control in a listening area by systematic regulation of loudspeaker signals on its boundary. This chapter starts with an overview including the history of wave field synthesis and some exemplary installations. Next, the theoretic fundamentals of wave field synthesis are detailed. Technical implementations demand drastic simplifications of the theoretical core, which come along with restrictions of the acoustic control as well as with synthesis errors. Simplifications, resulting synthesis errors as well as the working principles of compensation methods and their effects on the wave field are extensively discussed. Finally, the current state of research and development is addressed.
... Concurrently to this emerging interest for audio technologies on the part of flight vehicle manufacturers that are concerned by the sound quality and sound environment of their products, the audio community is years in advance for the application of sound field reproduction, multichannel equalization, and panel actuators [6][7][8][9][10][11][12][13]. In this context, this paper presents an illustrative application of audio technologies to an emerging field of research and development that could represent a promising expansion for the audio community. ...
Conference Paper
Full-text available
Reproduction in a mock-up of aircraft cabin noise in various flight conditions is an interesting tool for the prediction, optimization, demonstration and jury testing of interior aircraft sound quality. To provide a faithfully reproduced sound environment, time, frequency and spatial characteristics of the actual sound field should be preserved. This communication presents a cabin sound reproduction system in a full-scale mock-up of a CRJ 1000 Bombardier aircraft cabin. The reference sound field is based on a spatial capture of the CRJ cabin sound in various flight conditions using a 80 microphone array. The reproduction approach presented here is based on multichannel equalization using regularized least-mean-square minimization. The system involves inertial actuators mounted on the mock-up trim panels as well as bass shakers and a sub-woofer. The communication will discuss the system design as well as objective and subjective evaluation. http://pub.dega-akustik.de/IN2016/data/articles/001043.pdf
Experiment Findings
Full-text available
La metodología de este proyecto responde a el enfoque mixto, donde se tomaron elementos tanto cuantitativos (mediciones objetivas) como cualitativos (mediciones subjetivas), para el estudio del panel multiactuador (diseño y construcción) y codificación B-Format para la reproducción sonora envolvente a través del mismo.
Article
Advanced automotive audio applications are more and more demanding with respect to the visual impact of loudspeakers while still requiring more and more channels for high quality spatial sound rendering. The use of arbitrary plate-like structures driven by electromagnetic actuators or by piezoelectric elements appears as a promising solution to tackle both issues. However, to meet spatial rendering audio constraints (i.e. to be as close as possible to omnidirectional piston-like sources), the generated bending waves must be focused at a given position and to a certain extent within the host plate which can be of arbitrary shape, material, and thickness. Theoretically, this means being able to invert the spatio-temporal wave propagation operator for the generated bending waves to fit a given target shape. There are several methods (modal control, time-reversal, and propagating waves operator inversion) that allow to focus bending waves in a media. However, there is scarce work on their adaption and performances assessment in the context of audio applications. These methods depend differently on the available knowledge of wave propagation in the plate (theoretical, partial spatial or full spatial knowledge) and are here investigated to perform this task. Their performances are assessed with respect to several aspects: geometrical complexity, thickness, and material damping of the host structure, number and type of actuators, position and extent of the focusing area. The various methods are presented in a unified theoretical framework and they are compared by means of two key performance indexes (focus localization error and spatial contrast). An experimental validation on a relevant industrial case is also carried out and learning through a digital twin instead of time consuming experimental data investigated. This work falls within the framework of research which tries to bridge the gap between laboratory research and industrial deployment of this kind of technologies.
Article
Full-text available
A methodology for the analysis of the radiation characteristics and spatial performance of loudspeaker arrays for wave field synthesis (WFS) reproduction is presented. It is based on the wavenumber domain analysis, where the source radiation is decomposed into plane waves for arbitrary angles of incidence. The method deals with the measurement and analysis of the radiation performance, evaluation of the spatial aliasing frequency, and associated sampling artifacts for linear loudspeaker arrays. A detailed description of the parameters that modify spatial aliasing artifacts, such as array directivity and truncation effects by geometry, is also given. The method is validated at the laboratory with two multiactuator panel arrangements and a dynamic loudspeaker array, all presenting the same transducer spacing. Simulations and experimental results are discussed through several case studies, comparing dynamic loudspeaker arrays and multiactuator panels in WFS operation. Moreover, a study of the consequences for the wavefield when splitting the panels to accommodate a lower number of exciters is also addressed.
Article
Wave Field Synthesis is a 3D audio reproduction system, which allows synthesizing a realistic sound field in a wide area by using arrays of loudspeakers. However, the listening room adds echoes not considered by the wave field synthesis system, thus deteriorating the spatial e®ect. This paper proposes a new room compensation approach based on a multichannel inverse filter bank calculated to compensate the room e®ects at selected points within the listener area. Time domain algorithms are proposed to accurately compute the bank of inverse filters. Di®erent laboratory experiments have been carried out to validate the method.
Article
In sound rendering systems using loudspeakers, the listening room adds echoes not considered by the, reproduction system, thus deteriorating the rendered audio signal. Specifically Wave Field Synthesis is a 3D audio reproduction system, which allows syn thesizing a realistic sound field in a wide area by using arrays of loudspeakers. This paper proposes a room compensation approach based on a multichannel inverse filter bank calculated to compen sate the room effects at selected points within the listening area. Time domain and frequency domain algorithms are proposed to ac curately compute the bank of inverse filters. A comparative study between these algorithms by means of laboratory experiments is presented.
Article
A new concept for multichannel loudspeakers is introduced for application in wave field synthesis (WFS). It is a multi-actuator panel (MAP), which consists of a damped acoustic radiation panel with a number of exciters that are used to generate the WFS wave field. It is first shown that distributed-mode loudspeaker (DML) technology can be applied successfully to WFS. There are, however, some drawbacks to apply standard DML panels for this application. This led to the development of the MAPs. It is shown from theory and is confirmed by measurements that these panels can be designed in such a way that they are ideally suited for WFS sound reproduction.
Conference Paper
Multiactuator panels are a possible solution to satisfying the requirement of a large number of loudspeaker channels inherent in wavefield synthesis. The structural acoustic behavior of multiactuator panels has been measured with a laser Doppler vibrometer, and acoustic radiation simulation has been performed using a discretized Rayleigh I integral. The analysis showed that, due to large structural damping, the acoustic radiation is generated almost entirely by the structural near field around the excitation point on the panel, but it is influenced to a large extent by the panel dimensions and the exciter position. The radiation principle of a distributed-mode loudspeaker from the literature is briefly contrasted.
Article
Wavefield synthesis is a concept that enables the generation of sound fields with natural properties in time and space. The concept is applied to the reinforcement of direct sound. Wavefield synthesis is based on Rayleigh's representation theorem. In their original form, the discretized Rayleigh integrals prescribe the use of planar arrays of monopole- or dipole-type loudspeakers. It is shown that the theoretical synthesis operator can be modified such that a linear array of loudspeakers with arbitrary directivity characteristics can be used to generate a reinforced wavefield with natural properties in time and space. Since the derivation of the synthesis operator is based on array theory, defective characteristics of individual array loudspeakers cannot, or only in part, be corrected for by adapting the driving signal for these loudspeakers.
Article
Wave field synthesis (WFS) targets the synthesis of the physical characteristics of a sound field in an extended listening area. This synthesis is, however, accompanied by noticeable reconstruction artifacts. They are due to both loudspeaker radiation characteristics and approximations to the underlying physical principles. These artifacts may introduce coloration, which must be compensated for over the entire listening area. Multichannel equalization techniques allow for the control of the sound field produced by a loudspeaker array at a limited number of positions. The control can be extended to a large portion of space by employing a new method that combines multichannel equalization with a linear microphone array-based description of the sound field and accounts for WFS rendering characteristics and limitations. The proposed method is evaluated using an objective coloration criterion. Its benefits compared to conventional equalization techniques are pointed out for both ideal omnidirectional loudspeakers and multi-actuator panels.
Conference Paper
This article deals with the direct comparison of WFS rendering using either MAP or electrodynamic loudspeakers on an objective and a subjective level. Objective criteria are used to evaluate coloration and localisation cues that are perceived in an extended listening area. It is shown in a first listening test that the reduced spatial coherence of MAP loudspeakers partly explains the perceived differences between the two loudspeaker technologies. This "diffuse" behavior can be artificially produced on electrodynamic loudspeakers using a diffuse filtering. The proposed diffuse filtering may also limit rendering artifacts above the spatial aliasing frequency. Finally, it is shown in a second listening experiment that MAP loudspeakers favor distance perception compared to electrodynamic loudspeakers.
Conference Paper
Multi-Actuator Panel (MAP) loudspeaker arrays are advantageously used for wave field synthesis applications in the multi-media domain. However, due to their large dimensions they reflect each others sound waves, thus blurring the synthesized wave field. Furthermore, the MAPs have a very irregular initial frequency response, resulting in a highly coloured and unnatural sound. Filters for reflection elimination and equalization have been derived from impulse responses measured under anechoic conditions. These filters are based on inversion in the frequency domain, using a singular value decomposition method to avoid instabilities. It is shown how such filters can also be designed based on simulated data, thus avoiding the rather elaborate panel transport and measurement procedure. The performance of several filter types has been evaluated and compared physically and perceptually. The simulation-based filters appear to be a promising tool for further MAP performance optimization.