<!DOCTYPE article PUBLIC "-//NLM//DTD JATS (Z39.96) Journal Archiving and Interchange DTD v1.0 20120330//EN" "JATS-archivearticle1.dtd">
<article xmlns:xlink="http://www.w3.org/1999/xlink">
  <front>
    <journal-meta />
    <article-meta>
      <title-group>
        <article-title>A Minimalistic Sculptural Interfacefor Immersive Soundscape Presentation</article-title>
      </title-group>
      <contrib-group>
        <aff id="aff0">
          <label>0</label>
          <institution>Martin Rumori Institute of Electronic Music and Acoustics University of Music and Performing Arts Graz</institution>
          ,
          <country country="AT">Austria</country>
        </aff>
      </contrib-group>
      <fpage>25</fpage>
      <lpage>28</lpage>
      <abstract>
        <p>-Promenade proposes an interface for the semipersonal, interactive presentation of soundscapes that allows for the seamless integration of visual material. It combines the excitation of a flat surface using transducers with transaural cross-talk cancelled stereophonic projection. The unfolding of a wide spatial image within the sweet spot is contrasted by a collapsed, localised impression outside. This inherent interactive behaviour may be further extended by means of a distance sensor. Promenade emerged from an experimental, arts-based research project. It therefore targets contexts that feature custom audio content rather than aiming at general-purpose applications requiring optimal acoustic properties.</p>
      </abstract>
    </article-meta>
  </front>
  <body>
    <sec id="sec-1">
      <title>Exhibitions or other presentations of distinct objects often</title>
      <p>call for an unobtrusive sound projection means that affords the
visitors an immersive, mostly individual listening experience.</p>
      <p>Achieving such an experience requires spatial capabilities of
the chosen projection technique.</p>
      <p>In many such cases, headphones are used. Spatially rich
sound material may be projected in particular with binaural
signals. A major side effect is the inherent reconfiguration
of our multi-modal perception, as the surrounding auditory
domain is largely replaced by the projected one. Additionally,
headphones may be considered as cumbersome or unhygienic
in public exhibitions. Multichannel loudspeaker setups, in con- Figure 1. Promenade sound projection device.
trast, do not require physical engagement with the projection
medium. However, they do not facilitate individual listening
situations and they have to be mounted in an often complex
constellation around the target listening area.</p>
      <p>Promenade is a minimal auditory display prototype that
relates to such issues. It is important to notice that Promenade
has not been developed to overcome the limitations of other
projection means. Instead, it emerged from a series of
experiments in an arts-based research context as part of a sound
installation with a focus on aesthetic listening experience. Due
to its simplicity, its purpose and the idiosyncratic technology II. PROMENADE
it incorporates, Promenade is not a universal, general-purpose The Promenade sound projection device consists of a flat
spatial sound projection system but it manifests certain lim- surface of variable size with two electrodynamic exciters
itations. However, it might provide an appropriate and eleg- mounted on the back (Figure 1). Audio signals are fed to
ant solution in some of the circumstances described above. the exciters such that the surface acts as a loudspeaker
memPromenade’s qualities with respect to those circumstances brane. The physical properties of the surface material lead
are regarded retrospectively here in order to explore further to a certain radiation characteristics. The directional effect is
possible fields of application and obvious modifications for further emphasised by crosstalk-cancelled stereophony, which
different forms of appearance. creates a large auditory canvas that is much wider than the
This paper describes the basic idea of Promenade and its
originating context (section II), followed by introducing the
two main techniques involved, transaural crosstalk-cancelled
sound projection (section III) and the principle of flat panel
loudspeakers (section IV). Finally, some findings that arose
from practical experience are discussed, along with possible
modifications and extensions (section V).
projection panel (see section III). The effect occurs at a central
listening position, facing the device from a distance of a few
ten centimetres up to about one and a half metres. As soon as
the sweet area is left, the spatial auditory image collapses and
sound seems to be restricted to the panel surface. Although
both exciters are mounted to the same surface, their signal
differences remain intact to a sufficient extent for providing
the effect of crosstalk-cancelled projection.</p>
      <sec id="sec-1-1">
        <title>A. Physical construction</title>
        <p>Promenade has been prototyped with two materials and in
two sizes so far. The first series uses finnboard, a lightweight
type of cardboard, with the dimensions 70 by 100 centimetres
(portrait) and 3 millimetres thickness. The second, more
compact series is made of screenboard, a more dense type of
cardboard with white cover layers, 43,5 by 64 centimetres and
2 millimetres thick. Other promising materials that have been
explored include ply wood of a few millimetres thickness, and
KAPA, a compound material combining a polyurethane foam
core with cardboard cover layers.</p>
        <p>The smaller prototype uses Visaton EX-60 R exciters
(60 millimetres in diameter), while the larger is equipped with
the more powerful Visaton EX-80 S (square shape, 80
millimetres edge length). White flat cable serves for feeding the
exciters and for suspending the device. Both the exciters and
the cables are glued to the panels. The projection devices are
fed by small standard amplifier boards installed separately.</p>
        <p>
          In the installation, the projected signals were generated from
a BeagleBone Black computer board equipped with the Bela
audio cape.1 Sound programming and convolution for crosstalk
cancellation was performed in SuperCollider.2 As cancellation
filters, the so-called BACCH filters by Edgar Choueiri [
          <xref ref-type="bibr" rid="ref1">1</xref>
          ] were
taken from the jconvolver convolution engine bundle by Fons
Adriaensen.3
        </p>
      </sec>
      <sec id="sec-1-2">
        <title>B. Sound installation and artistic context</title>
        <p>structures, processes, and realities around us, with the
singularity of particular spots in space that, when located, allow for an
exclusive access to individual layers of this environment. The
crosstalk-cancelled projection technique restricts these spots
to clear-cut boundaries but on the other hand it potentially
‘elevates’ each layer’s spatiality as a sublime, temporary
overlay to the global environment.</p>
        <p>The installation Promenade serves as an experiential starting
point for further investigations into the fields of anecdotal
music, narrative identity construction and strollology. All of
these areas may be reflected in conjunction with aesthetic
experience in our everyday life auditory surrounding and its
technical mediation. Departing from a workshop within the
project Promenade, a publication is currently prepared that
summarises the described approach and the ongoing process.</p>
      </sec>
    </sec>
    <sec id="sec-2">
      <title>III. TRANSAURAL CROSSTALK-CANCELLED SOUND</title>
      <p>PROJECTION</p>
      <sec id="sec-2-1">
        <title>A. Working principle</title>
        <p>The sound projection device has been conceived as part Transaural crosstalk-cancelled sound projection, often
abof a sound installation within the arts-based research project breviated XTC, has been developed for playing back binaural
Promenade. Anekdote, Alltag, Abbild (Promenade. Anecdote, signals, which normally require headphones, on a pair of
everyday life, image), hence its name.4 loudspeakers (see, e. g., [2, pp. 283–326]). While headphones</p>
        <p>Seven playback canvases (five in a later version, Figure 2) isolate the sound propagation paths from each of the earphones
were distributed in the exhibition space, each suspended from to the corresponding ear, respectively, loudspeaker projection
the ceiling. The projection surface was left blank, without any involves transaural crosstalk: the signal radiated from the left
further visual additions. Each of the devices played back a ear also reaches the right one and vice versa. However, the
soundscape composition based on field recordings. Listeners playback of binaural signals requires the separate presentation
were invited to explore and ‘visit’ the different soundscapes of each channel to the corresponding ear.
by moving in space, approaching the projection devices one Transaural crosstalk cancellation is based on time-shifted
by one and ‘entering’ the recorded space by positioning anti-phase signals emitted from the respective opposite
loudthemselves in the sweet area where the crosstalk-cancelled speaker that shall arrive at the near-side ear at the same time
projection effect unfolds. as the unwanted crosstalk signal from the far side speaker</p>
        <p>
          The installation seeks to combine a global sonic envir- in order to cancel it out. Obviously, the cancellation signal
onment that reflects the simultaneity of seemingly unrelated causes a crosstalk at the far side ear, too, which in turn
is eliminated by another correction signal emitted from the
21hhttttpp::////sbueplae.ricoollider.github.io opposite speaker. Several iterations of this process lead to the
3http://kokkinizita.linuxaudio.org/linuxaudio/ design of a recursive filter network that ultimately seeks to
4https://esc.mur.at/de/projekt/promenade-anekdote-alltag-abbild ensure an isolated projection of the ear signal pair [
          <xref ref-type="bibr" rid="ref3">3</xref>
          ].
        </p>
      </sec>
      <sec id="sec-2-2">
        <title>B. Binaural vs. stereophonic signals with XTC</title>
      </sec>
    </sec>
    <sec id="sec-3">
      <title>As mentioned above, transaural crosstalk cancellation has</title>
      <p>
        been developed for the projection of binaural signals on
loudspeakers, that is, ear signals involving interaural time (ITD)
and level differences (ILD) as well as elevation-dependent
spectral cues (see [
        <xref ref-type="bibr" rid="ref4">4</xref>
        ]). Nevertheless, XTC filters are also used
for the playback of ordinary loudspeaker stereophonic signals.
The latter application is known as Ambiophonics in consumer
entertainment contexts [
        <xref ref-type="bibr" rid="ref5">5</xref>
        ].
      </p>
      <p>Experimental findings of the Promenade project with
respect to stereophonic signals are discussed below (Section V).</p>
    </sec>
    <sec id="sec-4">
      <title>IV. FLAT PANEL LOUDSPEAKER PROJECTION</title>
    </sec>
    <sec id="sec-5">
      <title>The excitation of flat, solid matter surfaces as an alternative</title>
      <p>
        loudspeaker working principle has been investigated during
the past decades (see, e.g., [
        <xref ref-type="bibr" rid="ref6">6</xref>
        ]). Major advantages are the
appealing form factor requiring dramatically reduced mounting
depth, and the potential to create diffuse sound fields due to
the low directivity even at high frequencies.
      </p>
      <p>One important factor for exciting flat panels is the mounting
location of the exciter. Depending on the application, the
physical properties and the size of the panel, a basic question
is whether to exploit or to avoid the direct excitation of
predominant modes. Often, an out-of-centre location avoiding low
integer partition of the surface dimensions is recommended
for a more balanced response. In Promenade, exciters were
mounted according to the golden section as an experimentally
found trade-off of acoustic and visual qualities.</p>
      <p>Apart from the construction of high-end loudspeakers,
electrodynamic excitation of manifold objects for sound
projection is frequently applied in custom presentation media and
in sound art (see, e.g., SHHH SHHH SHHH5 by Amanda
Dunsmore, 2008, or Touched Echo6 by Markus Kison, 2007).
As described above, the exploration for Promenade departs
from the latter. In such realms the presented sound material is
usually created or adapted especially for the specific playback
medium. Consequently, the projection device does not need
to meet the same criteria as general-purpose loudspeakers,
especially with respect to linearity and neutrality, because
such deviations can be corrected in the adaption process. This
largely facilitates achieving a beneficial result with relatively
low-cost equipment.</p>
      <p>
        A special technical issue of the Promenade device is that
only one panel is used for projecting two audio channels at
the same time. The crucial question is whether the crosstalk
cancellation cues are propagated to a sufficient extent for
achieving a similar effect as with two small-spaced
loudspeakers or whether the inter-channel crosstalk in the panel degrades
the cues too much. A similar technique has been investigated
for the application of Wave Field Synthesis (WFS) using even
more than two exciters on the same panel, coined
MultiActuator Panels [
        <xref ref-type="bibr" rid="ref7">7</xref>
        ]. Later, the method has been generalised
for creating multiple sound sources with a single panel [
        <xref ref-type="bibr" rid="ref8">8</xref>
        ].
5http://www.lit.ie/Dunsmore/linz/kunstraumwindows.html
6http://www.markuskison.de/touched_echo.html
      </p>
      <p>The experimental artistic context of Promenade did not
pursue a scientific approach to such a technique for transaural
sound projection. In multiple iterations, the used combination
of materials, dimensions and mount points of the exciters have
been adjusted such that the desired effect could be achieved
sufficiently well. A validation by acoustic measurements,
however, has not yet been carried out.</p>
    </sec>
    <sec id="sec-6">
      <title>V. DISCUSSION</title>
      <p>The presented sound projection device shows an
appropriate performance within its frame of origin, the Promenade
arts-based research project. There, it allows for experiencing
the often surprising, sometimes uncanny effect of
crosstalkcancelled sound projection and supports the exploration and
reflection of aesthetic experience in multi-layered spatial
auditory environments.</p>
      <p>Beyond the attached context, the principles exploited in
Promenade may be applied to other use cases. In the following,
some findings and possible directions are discussed, such as
the use of stereophonic signals, the integration with visual
presentations, extensions towards interactive applications, and
other form factors, materials, or objects.</p>
      <sec id="sec-6-1">
        <title>A. Crosstalk cancellation of stereophonic signals</title>
        <p>When developing the Promenade sound installation with
field recording material and XTC projection, it was found
that coincident stereophonic signals involving only level
differences but none of phase provide a more consistent and
stable spatial image than recordings in spaced omnidirectional
stereophony that is based on phase differences. There is no
well-researched explanation for this observation. It is likely
that both the listening conditions and the projection device are
not optimal for propagating phase differences to the listener’s
ears. In particular the fact that both exciters are working on
the same projection surface may degrade the fidelity of phase
reproduction, although a similar observation was made using
two separate loudspeakers.</p>
        <p>Overall, the discovery that intensity stereophony is well
suited to span the extremes of the projection base – from ear
to ear using headphones, nevertheless crossing the listener’s
head, from speaker to speaker in an ordinary stereophonic
setup, and along a well externalised, virtual stereo base of
about 150 degrees opening angle using XTC projection –
is most valuable for the production of appropriate material.
Following from that, field recordings were carried out in
mid/side stereophony, which allows for further benefits with
respect to adaptation and interaction (see subsection V-C).</p>
      </sec>
      <sec id="sec-6-2">
        <title>B. Visual integration</title>
        <p>In the Promenade sound installation, major emphasis was
put on listening to mediated auditory environments. Therefore,
the metaphor of a canvas was picked up that could have shown
a landscape but instead projected a soundscape while
remaining visually ‘silent’. However, the surface of the projection
medium may be easily used for a seamless integration of visual
and auditory material. In the simplest case, the cardboard
material may be printed or laminated. The backside may be
used as well, restricted by the mounted exciters.</p>
        <p>Alternatively, the panel may serve as a video projection
screen, including three-dimensional technology involving
polarised or shutter glasses. The projection angle and distance
have to be envisaged such that the spectator is not disturbing
the projection but may still position herself appropriately with
respect to crosstalk-cancelled sound projection.</p>
        <p>Using stereophonic signals, auditory events can be evoked
for the listener anywhere in the horizontal dimension within
the borders of the canvas and beyond. Therefore, horizontal
audio-visual congruence is well achievable, for example, for
annotating visual items in maps or alike. Psychoacoustic
effects may be exploited for suggesting vertical
correspondences. Depending on the stiffness of the projection panel,
the location of the exciters on the panel is audible in certain
frequency bands. In this case, a second pair of exciters may
serve as a panning means in the vertical dimension, at least to
a certain extent.</p>
        <p>Generally, any visual presentation that favours a fixed,
frontal and relatively close viewpoint can be expected to work
well in combination with a Promenade like sound projection
device.</p>
      </sec>
      <sec id="sec-6-3">
        <title>C. Interactivity, tracking, sensors</title>
        <p>From the beginning, the Promenade device has been
conceived with an additional optical distance sensor such that
the playback can react interactively, depending on whether,
how close and for how long a listener is present in front
of the canvas. For the installation versions involving several
projection panels, actual use of the sensors has been dismissed
so far. It turned out to be more desirable in the first place to
explore the field of sweet spots distributed in space which,
when being entered, change the listening experience
dramatically. This effect may be regarded as an inherent ‘interactive’
behaviour of the projection dispositif, although no technical
reactivity is involved.</p>
        <p>Nevertheless, involving the sensor bears several interesting
use cases. If multiple projection panels are present, the sensors
may serve as a simple tracking device in order to activate or
emphasis one or more panels rather than others, both in a
decentralised (each panel controls itself without influencing
others) or a centralised manner. Furthermore, the spatial
arrangement of sound material could be extended towards an
enterable, interactive radio play. Different individual listeners
cannot be distinguished, though, which would require more
sophisticated tracking means.</p>
        <p>The sensor information may also affect the projection
technique itself. In the Promenade installation, gradual transitions
take place from monaural signals (centered in the projection
panel) over ordinary stereophony (located roughly in between
the exciter positions, thus a partition of the panel) to
crosstalkcancelled projection. As mid/side-stereophony is used,
modulating the contribution of the side signal allows for further
controlling the perceived width of the resulting sonic image.
Such transitions may be triggered or emphasised by evaluating
the listener’s distance or overall presence via the sensor.</p>
      </sec>
      <sec id="sec-6-4">
        <title>D. Other form factors, materials, or objects</title>
      </sec>
    </sec>
    <sec id="sec-7">
      <title>Theoretically, combining two-channel electrodynamic excit</title>
      <p>ation of solid matter with transaural crosstalk cancellation is
possible with various objects and materials of any form or
extension. For example, architectural prototypes, engineering
models or basic commodity might be enhanced by an
integrated auditory layer that invites for spatial immersion.</p>
      <p>The determining factors remain those that have been
consulted while conceiving Promenade:
• the physical properties of the excited material with
respect to sound propagation,
• appropriate mounting locations for the exciters and fitting
form factors of the object and the exciters, and
• sufficiently isolated radiation of both channels for
achieving the desired effect of crosstalk cancellation.</p>
    </sec>
    <sec id="sec-8">
      <title>VI. CONCLUSION</title>
      <p>In this paper, the Promenade spatial sound projection device
has been presented, along with its originating context, a sound
installation as part of the arts-based research project of the
same name. The device has been presented as combining the
two techniques of transaural crosstalk cancellation for inviting
a rich spatial auditory experience with the excitation of solid
objects for an unobtrusive appearance, differing from that of
ordinary loudspeakers. The presentation form of a canvas,
though not visually occupied in the original installation, invites
for manifold applications that involve audio-visual
correspondence. Further extensions include the use of distance sensors
for interactive functionality, both with respect to the presented
narrative and to the evoked spatial impression.</p>
    </sec>
    <sec id="sec-9">
      <title>ACKNOWLEDGEMENTS</title>
    </sec>
    <sec id="sec-10">
      <title>The research project Promenade is supported by the Vice Rector’s office of research at University of Music and Performing Arts Graz, the Cultural Office of the City of Graz, and ESC Media Arts Laboratory, Graz.</title>
    </sec>
  </body>
  <back>
    <ref-list>
      <ref id="ref1">
        <mixed-citation>
          [1]
          <string-name>
            <given-names>E. Y.</given-names>
            <surname>Choueri</surname>
          </string-name>
          , “
          <article-title>Optimal crosstalk cancellation for binaural audio with two loudspeakers</article-title>
          ,”
          <year>2010</year>
          . [Online]. Available: https://www.princeton.edu/ 3D3A/Publications/BACCHPaperV4d.pdf
        </mixed-citation>
      </ref>
      <ref id="ref2">
        <mixed-citation>
          [2]
          <string-name>
            <given-names>B.</given-names>
            <surname>Xie</surname>
          </string-name>
          ,
          <string-name>
            <surname>Head-Related Transfer</surname>
          </string-name>
          Function and
          <article-title>Virtual Auditory Display</article-title>
          . Plantation: J.
          <string-name>
            <surname>Ross</surname>
          </string-name>
          ,
          <year>2013</year>
          .
        </mixed-citation>
      </ref>
      <ref id="ref3">
        <mixed-citation>
          [3]
          <string-name>
            <given-names>R.</given-names>
            <surname>Glasgal</surname>
          </string-name>
          , “
          <volume>360</volume>
          °
          <article-title>localization via 4.x RACE processing,” in 123rd AES Convention</article-title>
          . AES,
          <year>2007</year>
          .
        </mixed-citation>
      </ref>
      <ref id="ref4">
        <mixed-citation>
          [4]
          <string-name>
            <given-names>J.</given-names>
            <surname>Blauert</surname>
          </string-name>
          , Spatial Hearing. Cambridge Mass./London: MIT Press,
          <year>1997</year>
          .
        </mixed-citation>
      </ref>
      <ref id="ref5">
        <mixed-citation>
          [5]
          <string-name>
            <given-names>R.</given-names>
            <surname>Glasgal</surname>
          </string-name>
          , “Ambiophonics:
          <article-title>The synthesis of concert-hall sound fields in the home,” in 99th AES Convention</article-title>
          . AES,
          <year>1995</year>
          .
        </mixed-citation>
      </ref>
      <ref id="ref6">
        <mixed-citation>
          [6]
          <string-name>
            <given-names>G.</given-names>
            <surname>Bank</surname>
          </string-name>
          and
          <string-name>
            <given-names>N.</given-names>
            <surname>Harris</surname>
          </string-name>
          , “
          <article-title>The distributed mode loudspeaker - theory and practice,” in 13th Microphone &amp; Loudspeakers AES Conference</article-title>
          . AES,
          <year>1998</year>
          .
        </mixed-citation>
      </ref>
      <ref id="ref7">
        <mixed-citation>
          [7]
          <string-name>
            <surname>M. M. Boone</surname>
          </string-name>
          ,
          <article-title>“Multi-actuator panels (MAPs) as loudspeaker arrays for wave field synthesis</article-title>
          ,
          <source>” Journal of the Audio Engineering Society</source>
          , vol.
          <volume>52</volume>
          , no.
          <issue>7</issue>
          /8, pp.
          <fpage>712</fpage>
          -
          <lpage>723</lpage>
          ,
          <year>2004</year>
          .
        </mixed-citation>
      </ref>
      <ref id="ref8">
        <mixed-citation>
          [8]
          <string-name>
            <given-names>M. C.</given-names>
            <surname>Heilemann</surname>
          </string-name>
          ,
          <string-name>
            <given-names>D.</given-names>
            <surname>Anderson</surname>
          </string-name>
          , and
          <string-name>
            <given-names>M. F.</given-names>
            <surname>Bocko</surname>
          </string-name>
          , “
          <article-title>Sound-source localization on flat-panel loudspeakers</article-title>
          ,
          <source>” Journal of the Audio Engineering Society</source>
          , vol.
          <volume>65</volume>
          , no.
          <issue>3</issue>
          , pp.
          <fpage>168</fpage>
          -
          <lpage>177</lpage>
          ,
          <year>2017</year>
          .
        </mixed-citation>
      </ref>
    </ref-list>
  </back>
</article>