Conference PaperPDF Available

SeamLess Integration of Spatial Sound Reproduction Methods



The presented software system aims at a combined use of different spatial sound reproduction methods, such as Wave Field Synthesis and Ambisonics, in a robust, user-friendly workflow. The rendering back-end is based on free and open source components, running on multiple Linux servers, thus allowing the operation of large loudspeaker setups. Using a send-based signal routing paradigm with an OSC message distribution software, the individual rendering engines can be combined seamlessly and extended with additional methods. Content can be created and played back using digital audio workstation projects which are unaware of the reproduction systems and make use of OSC automation plugins. This ensures a straightforward transfer of content between different sites and speaker configurations. Due to its adaptability, the proposed system is considered a potential solution for comparable setups with larger numbers of loudspeakers.
SeamLess Integration of Spatial Sound Reproduction Methods
Henrik von Coler
TU Berlin
Paul Schuladen
TU Berlin
Nils Tonn¨
TU Berlin
The presented software system aims at a combined use
of different spatial sound reproduction methods, such as
Wave Field Synthesis and Ambisonics, in a robust, user-
friendly workflow. The rendering back-end is based on free
and open source components, running on multiple Linux
servers, thus allowing the operation of large loudspeaker
setups. Using a send-based signal routing paradigm with
an OSC message distribution software, the individual ren-
dering engines can be combined seamlessly and extended
with additional methods. Content can be created and played
back using digital audio workstation projects which are
unaware of the reproduction systems and make use of OSC
automation plugins. This ensures a straightforward trans-
fer of content between different sites and speaker config-
urations. Due to its adaptability, the proposed system is
considered a potential solution for comparable setups with
larger numbers of loudspeakers.
Different methods for spatial sound reproduction, such as
Wave Field Synthesis (WFS), Higher Order Ambisonics
(HOA) or panning approaches like Vector Base Ampli-
tude Panning (VBAP), have individual strengths and draw-
backs. Depending on the requirements and resources, ei-
ther can be preferable for a specific application. Multiple
methods can also be combined to be used either alternating
or in parallel. This results in versatile systems for varying
or extended use cases with the ability of creating highly
immersive sonic experiences.
1.1 TU Studio
For several years, the TU Studio at TU Berlin operates a
sound field synthesis studio for music production and re-
search. It is equipped with a 192 channel WFS system,
a 21 channel loudspeaker dome for HOA and other meth-
ods, plus a classic ring of eight loudspeakers. With few ex-
ceptions, as for example in demos, these systems are usu-
ally used independently. One reason for this is the lacking
availability of comparable systems and the high effort of
installing them temporarily for concerts or shows. How-
ever, a recently built public listening room at the Hum-
boldt Forum in Berlin offers new possibilities and requires
Copyright: ©2021 1 et al. This is an open-access article distributed
under the terms of the Creative Commons Attribution License 3.0 Un-
ported, which permits unrestricted use, distribution, and reproduction in
any medium, provided the original author and source are credited.
an improved workflow for everyday operation. This paper
introduces the SeamLess software system, using the TU
Studio and the Humboldt Forum as example systems.
1.2 Humboldt Forum Listening Room
The Humboldt Forum is the new home of several museums
and a cultural platform at the center of Berlin. Located in
the reconstructed Berlin Palace, it encompasses amongst
others the Ethnological Museum of Berlin, the Museum
of Asian Art and the Humboldt-Lab. For immersive pre-
sentations of relevant audio content, the Ethnomusicology
branch of the museum commissioned a multichannel sys-
tem, combining a 2-dimensional WFS System and a 3-
dimensional HOA System.
A previous project of the Audio Communication Group
dealt with the planning, operation and evaluation of a multi-
channel loudspeaker system at the original location of the
Ethnological Museum in Berlin Dahlem [1]. This prede-
cessor was a 21 channel Ambisonics system. The software
solution was based on VST plugins by Matthias Kronlach-
ner [2] inside Reaper.
1.3 Requirements for the Combined System
The combination of different reproduction systems requires
a new software concept, since the fully integrated DAW
solution can not be extended to drive the WFS with the
given number of channels. The result is a distributed soft-
ware system, relying on task-specific components, with a
special focus on the everyday use in exhibitions with low
In the filed of computer music, respectively experimen-
tal electronic music and electroacoustic music, most com-
posers and performers are used to individual tools and work-
flows. Especially for spatialization this offers the most
powerful means when controlling a large number of sound
source positions and other sound qualities. Despite power-
ful plugin based environments, spatial practice can still be
considered a domain for technically skilled or experienced
computer musicians. However, since the system at Hum-
boldt Forum should be accessible to composers and artists
from various backgrounds, a seamless operation as in usual
DAW handling is aimed at. Further, the everyday use in a
museum routine demands a server-capable, headless ren-
dering system with a robust automation concept for sound
source movements. All processes need to be started and
organized in fully automated routines, in order to allow the
operation by the museum staff and ensure synchronicity
with other media processes.
Since the listening room in the museum is open to the
public during daytime, it is necessary to prepare content
in other studios. Projects thus need to be interchangeable,
ignoring differences in the hardware configuration. These
recurring problems are addressed by the SeamLess system.
1.4 Spatialization with WFS and HOA
Recent spatial audio systems usually work with the con-
cept of virtual sound sources. This is also referred to as
the object based [3] approach, in contrast to channel based
approaches, which consider specific loudspeaker setups.
In object based spatialization, audio signals are linked to
dynamic spatial source positions. The actual loudspeaker
signals are then generated in real time by applying these
source positions within the rendering algorithms. With this
concept, the audio material is detached from the presenta-
tion format and the content can be transferred to different
loudspeaker setups or systems. This applies to Ambison-
ics, WFS and VPAB.
Figure 1 shows the model of a virtual point sound source
with spherical coordinates, as used within the system. The
position is defined by the angles azimuth (α) and elevation
() and a distance (d) from the origin, which is the center
of the listening space. Cartesian coordinates, with the di-
mensions x, y and z can also be processed by the proposed
Figure 1. Virtual sound source with spherical coordinates and normal
projection N.
Table 1 shows the full list of source attributes and their
applicability in the SeamLess system. Only the HOA sys-
tem is capable of rendering the three-dimensional position
of the source. Since the WFS system is working in two
dimensions, the elevation, respectively the z component of
the position, is ignored and only the x-y components are
rendered. The Doppler effect is an inherent part of the
WFS rendering algorithm. The pitch shift when moving
a source might be undesirable for musical reasons and can
be deactivated in the rendering software. WFS is capa-
ble of rendering so called focused sound sources. These
virtual sound sources are located between the loudspeak-
ers and the listener [4]. The effect of focused sources is
a strength of WFS systems, especially for moving audi-
ences, and thus frequently used. Plane wave sources in
WFS are characterized through a stable perceived direc-
tion of the source independently of the listeners position.
For the plane wave, the source the position only influences
the phase of the signal. An additional parameter angle is
set implicitly with the source position so it points to the
Table 1. Accessible properties of rendering systems in the current con-
Position Doppler Focused
WFS × × × ×
HOA × × ×
In the SeamLess system , WFS and HOA can be used
in parallel. Sound material can be routed to the different
rendering engines with a desired gain through a send bus
system, explained in Section 3.1. Although this does not
result in a coherent sound field, sounds can be gradually
faded between the rendering engines or played on WFS
and HOA simultaneously. Besides working with individ-
ual sound sources, encoded Ambisonics content can be di-
rectly decoded and played over the speakers. This can be
artificially created Ambisonics files, as well as recorded
content from Ambisonics microphones.
2.1 Loudspeaker Setups
The setups used in the mentioned projects rely on WFS
panels by Four Audio 1. Each panel provides eight WFS
channels. Each channel is using a column of three tweeters
with a horizontal column distance of 10 cm. Four columns
share one woofer. The recent version of the panels is equip-
ped with a DANTE interface for receiving the WFS chan-
nels from the rendering computers.
2.1.1 TU Studio
The WFS system in the TU Studio has an irregular octag-
onal geometry with 24 panels, as shown in Figure 2. A
21 channel Ambisonics dome with the geometry shown in
Figure 3 uses Neumann KH120A loudspeakers.
Figure 2. Geometry of the WFS system in the TU Studio.
Left-Right / m
Front-Rear / m
Floor-Ceiling / m
Figure 3. Dome with 21 speakers at TU Studio [5].
2.1.2 Humboldt Forum
A dedicated listening room with acoustical treatment was
planned for the Humboldt Forum by the M¨
uller-BBM Hold-
ing 2and installed by Neumann & M¨
uller 3. Figure 4 shows
the top view of the listening room, consisting of two oppos-
ing arcs with two entrances. With an area of approximately
7 × 13 m it can hold up to 20 standing peoples.
Figure 4. Top view of the listening room at Humboldt Forum with WFS
panels (gray) and Ambisonics ceiling speaker (black).
Each arc holds 16 WFS panels, mounted in a continu-
ous ribbon above head height, as shown in Figure 5. This
results in a total of 256 WFS channels, respectively 768
tweeters and 64 woofers. 45 Genelec 8020 speakers, drawn
as black rectangles in the top and side view, are used for
HOA rendering. They are arranged in three levels (1.25 m,
1.90 m 2.5 m) on both arcs. A single ceiling loudspeaker
is mounted in the center of the listening area. Additionally,
four Fohhn Arc AS-10 subwoofers are installed at ground
2.2 Rendering Servers
The processing of the SeamLess system is parallelized on
hardware and software level. Work is distributed to several
machines, each able to run multiple rendering processes
for different loudspeaker sections. In the Humboldt Forum,
Figure 5. Side view of one arc, showing WFS panels (gray) and Am-
bisonics speakers (black).
two computers are used for WFS, each rendering 128 WFS
channels. Another computer is used for control and HOA
Dante PCIe cards are used for audio routing between com-
puters and loudspeakers. The models Digigram LX-Dante
and Four Audio Dante PCIe are both suitable for Linux
systems. Digigram offers a closed source Linux driver for
Ubuntu 20.04 which needs to be updated manually for new
kernel versions.
3.1 System Overview
Professional Linux audio systems offer flexible means for
combining different software components, mainly through
the JACK Audio API 4. All rendering nodes are thus equip-
ped with Ubuntu (Studio) 20.04. Figure 6 shows the signal
flow of a system with a main machine for Ambisonics and
routing and two WFS rendering machines, as used in the
Humboldt Forum.
A playback computer sends raw audio to all rendering
machines and aligned OSC control data for spatialization
to the main Linux computer. An OSC routing software
converts the incoming generic OSC messages from the play-
back machine and sends the results to all instances of SC
Mix and to cWonder, the WFS control software.
SC Mix is responsible for Ambisonics encoding and acts
as a mixing and distribution component between playback
sources and the different rendering units. Send gains to the
different rendering units can be controlled for each sound
source. Minimal versions of SC Mix are thus running on
the WFS servers for applying the send gains to the raw
audio signals from the playback machine.
This system makes it possible to use different render-
ing systems simultaneously with a single control interface
from the user’s perspective. SC Mix sets the send gains to
the rendering systems according to the OSC Router. The
main machine’s instance additionally prepares a signal for
the subwoofers, encodes an Ambisonics signal from the
individual sources and calculates an encoded Ambisonics
reverb signal. Separate HOA decoders are used for the
mixed signal and the reverb. The tWonder WFS render-
ing instances on the WFS nodes are not directly controlled
from the OSC Router but from cWonder.
3.2 WFS Rendering
WONDER [6] is a software project for real time wave field
synthesis (WFS). It has been designed for running one of
the world’s largest WFS systems in the lecture hall H 104
Linux: MAIN / HOA
Linux: WFS1
MAC/PC: Content Playback
Linux: WFS2
3rd Ord. Decoder
tWonder (1-16)
DAW / Player
OSC Route
Raw Audio
tWonder (17-32)
SC Mix (Send Gains)
SC Mix (Send Gains)
WFS input
WFS input
Send Gains
3rd Ord. Encoder
SC Buses
SC Mix
1st Ord. Decoder
1st Ord. Reverb
Rev. Gains
HOA Speakers
Control (OSC)
Figure 6. Signal flow with audio and control connections for a setup with two WFS rendering machines.
at TU Berlin, with over 800 channels and an audience ca-
pacity of over 600 people. Due to the separation of soft-
ware modules for control and rendering, WONDER can be
used on distributed systems and can thus be scaled to any
number of channels. WONDER has also been used as the
standard tool in the TU Studio and features components
not only for WFS rendering, but also a score player and
additional tools.
Equipped with the same WFS panels as the TU Studio,
the I2AudioLab at HAW Hamburg 5created WONDER Lite,
a streamlined version from the original WONDER reposi-
tory. It includes only the WFS rendering components cWon-
der,tWonder,xWonder and libwonder, minimizing depen-
dencies and easing maintenance as well as further devel-
opment. Based on this work, additional maintenance and
modernizing work was done for this project. Especially the
robustness of the startup process was improved. With the
introduction of systemd services for the startup of the indi-
vidual components, the previously used startup scripts be-
come obsolete. As the WFS rendering is controlled through
plugins and the OSC Router, the original GUI xWonder is
not used.
Other WFS rendering solutions tested within this project
and used at an earlier stage include the SoundScape Ren-
derer (SSR) [7] and the PanoramixApp [8]. Each of these
alternatives has specific advantages. However, WONDER
offers the best options for running on a distributed server
system as needed.
3.3 Ambisonics Rendering
Ambisonics encoding is realized with the SC-HOA plug-
ins [9] in the SC Mix instance on the main rendering ma-
chine. For each source, azimuth, elevation and distance can
be controlled via OSC commands and all resulting Am-
bisonics signals are summed inside SuperCollider. This
concept has been used in previous projects for distributed
spatial audio performances [10].
Decoding of the Ambisonics signals is based on the Am-
bisonics Decoder Toolbox (ADT) [11], a versatile tool for
generating HOA decoders. Making use of Matlab/Octave
scripts and the FAUST [12] compiler, it is possible to build
decoders for various targets. Initially, the proposed system
used decoders in the form of SuperCollider UGens. These
have been replaced with standalone JACK clients, which
are fed with the encoded Ambisonics signal from the Su-
perCollider encoder stage.
For SC Mix and the HOA encoder, a recent headless ver-
sion of SuperCollider, as well as the sc3-plugins are built
and installed on all machines. As the SuperCollider com-
ponents are controlled solely via OSC messages and are
running in the background, graphic dependencies like X11
and Qt not only represent unnecessary overhead but tend to
decrease the audio performance in the absence of a graphic
3.4 Reverb
Artificial reverberation can be added from the DAW pro-
jects, by sending custom reverb channels to virtual sound
sources on the WFS or the Ambisonics system. This was
the standard procedure in most projects on the WFS sys-
tems in the TU Studio and the auditory. In addition, a
parametric Ambisonics reverb is integrated in the proposed
software solution. It is based on the Zita6reverb imple-
mentations in Faust. A SuperCollider UGen is generated
with the Faust compiler for creating a first order encoded
reverb signal of all sources. This signal is passed to an
external first order decoder and subsequently routed to the
Ambisonics loudspeakers. Using a parametric reverb al-
lows to change the reverberation characteristics within or
linuxaudio/zita-rev1- doc/quickguide.html
between projects, without loading additional impulse re-
3.5 OSC Message Routing
As central interface and connection point between all soft-
ware modules serves an OSC router and processor writ-
ten in Python. It is responsible for translating the incom-
ing OSC messages from DAW plugins or other sources to
the right format and sending it to the rendering modules.
The different software modules it communicates with can
be preconfigured or registered as clients during runtime.
It differentiates between UI, rendering and data clients.
The rendering clients are basically the targets of the OSC-
Messages coming in from either UI- or data clients. Data
clients send previously recorded or created movement data
while UI-clients enable interactive input by the user. Ev-
ery rendering client is registered with an address, a posi-
tion format and an update interval. The position of a sound
source can be set in any format, such as Cartesian, polar or
spherical, and will be converted to the right format when
sending it to the client. The maximum send rate for a single
source is limited by the update interval and can also be set
individually for every rendering client. Hence the WON-
DER software, which has its own position interpolation
methods, can be fed with a lower position data rate than
the Ambisonics renderer which works better with a high
position data rate input. While UI and rendering clients
are technical treated the same and are informed about ev-
ery change in the source data, the data clients are only get-
ting changes coming from UI clients to avoid a loop of
OSC messages. For easier handling, inputs of UI clients
can block the input of data clients for a short amount of
time, since some OSC automation plugins are constantly
sending their state.
3.6 Production and Playback
Since the rendering system is controlled in real time with
OSC messages, it is not bound to specific software for pro-
duction and playback, as long as audio and related control
data are streamed synchronously. However, for reasons of
accessibility, a DAW based approach with plugins for OSC
automation is proposed. Reaper was chosen as the main
DAW for content production and playback. One reason for
this is the high flexibility in channel routing when work-
ing with multi-channel content. Audio files with up to 64
channels can be created and processed in Reaper, allow-
ing tracks with up to 64 channels and the same number of
outputs. This makes it also ideal for Ambisonics content
that may have a high channel count depending on its order.
Further, Reaper allows embedding plugins and automation
trajectories into single audio items inside a project. The
final distribution format is a rendered multichannel audio
file, including the embedded automation data, which can
be used for the content playback and arrangement. By us-
ing the Reaper scripts, the rendering of audio output and
the export of automation data can be fully automated.
The approach of separating rendering software from the
DAW relies on plugins capable of sending OSC commands,
based on automation trajectories. Each audio track in a
project is equipped with one instance of such plugin, al-
lowing to automate its source position and other attributes.
A working solution is provided by IRCAM, namely OS-
Car 7, the successor of the ToscA plugin. OSCar has been
designed for controlling the PanoramixApp from a DAW.
Alternatively, the free software plugin osccontrol-light 8
has been tested. Albeit currently less feature-rich, the plu-
gin code is open source and can be adjusted to meet the re-
quirements. Both solutions are general purpose tools and
can be configured to send and receive the relevant OSC
messages through configuration files. Finally, a dedicated
plugin has been developed for this project and is included
in the software repository. It is application-specific and
needs no additional configuration, besides the target IP ad-
dress and port.
3.7 Subsystem Configuration and Integration
The startup of WONDER was originally managed with
shell scripts. Those scripts relied on a specific order of
execution. If the startup failed at some point, all systems
had to be restarted. The scripts were responsible for the
startup of the entire system, including JACK. For this they
used additional configuration files. The configuration is
now stored in the system-wide directory /etc/wonder.
This also includes the speaker locations for the WFS ren-
derer as well as the remaining startup script configuration.
All other configuration files, including JACK configuration
and connections, are located in /etc/seamless.
systemd services offer a more robust startup process in
comparison to a cascade of shell scripts. All services can
be started and stopped independently. They can define re-
quirements, such as network and audio interfaces, and can
be configured to start on system boot. There are system
services now for JACK, tWonder,cWonder, the Ambison-
ics decoders, SC Mix, the OSC Router and aj-snapshot.
Running JACK as a system service resulted in a conflict
with Ubuntu Studio’s autojack when logging in. To pre-
vent this, the ubuntustudio-control package needs to be re-
Improving the startup process significantly increases the
robustness and reliability of WONDER. A failed tWonder
instance is now able to reconnect to cWonder. The aj-
snapshot service is used to manage all JACK connections.
JACK connections can be stored and reloaded as snapshots
for different use cases. The service runs as a daemon and
connects new jack clients, if applicable.
4.1 Combining WFS and HOA
When using Ambisonics and WFS in a combined system,
their individual strengths can be used for specific effects.
Since Ambisonics comes with inherent means of record-
ing, it is well suited for capturing actual soundscapes on
site with dedicated Ambisonics microphones. In the case
of the listening room at Humboldt Forum this seems to be
a recurring concept, since artists often provide field record-
ings from specific sites.
WFS, on the other hand, is well suited for creating highly
locatable sound events from isolated recordings, as achie-
ved by close up miking. Especially for a moving audience,
as in the listening room, this draws the attention to such fo-
cused sound sources and creates the illusion of virtual ob-
jects in the listening space. An advantage of WFS comes
from the fact that human hearing capabilities have the high-
est resolution on the horizontal plane while sources com-
ing from above and below the ear-level can not be localised
with such a high precision. Having the WFS-system with
its very high spatial resolution on the horizontal plane, this
relationship can be used. By connecting the elevation prop-
erty of a source directly to the gain send to the different
systems, the combined use of both systems can be simpli-
fied from the user’s perspective. The send-based approach
allows the continuous fading of sound sources between the
different rendering systems. Since they share the same spa-
tial parameters, the position of all virtual sound sources are
identical, apart from differences between 2D and 3D ren-
4.2 User Control
In previous productions with spatialization systems it has
been found that the demands on user interfaces strongly
vary among composers and other users. Although the in-
cluded plugin comes with a simple graphical spatialization
interface, custom solutions are often desired. Since the
OSC router accepts various different message formats, it
is open for custom interface and control approaches and
also provides appropriate feedback for graphical user in-
terfaces. Individual solutions for the different projects can
thus be implemented easily and might lead to a repertoire
of different control possibilities. Nevertheless, the design
of an intuitive user interface, fitting most use cases, is a
continuing aspect of the project.
The proposed software system allows the seamless integra-
tion of different spatial rendering approaches and multiple
loudspeaker systems. Although the long term operation is
yet to be evaluated, the first application in a production for
the Humboldt Forum delivered a proof of concept, allow-
ing a composer to realize his ideas on the spatial arrange-
ment and evolution with the help of the engineers.
In its recent state, the system is still under development
and specific aspects are subject to improvements. Due to
the modularity of the system, alternative rendering soft-
ware can always be tested, compared and included. For in-
stance, the Ambisonics Toolkit (ATK) for SuperCollider 9,
which is part of the SC3-Plugins, was recently upgraded
to also allow Higher Order Ambisonics. Other possible
improvements include parametric reverb with directional
early reflections for an increased plausibility or the use of
high quality Ambisonics room impulse responses.
Based on free and open source software, the proposed
system is an open source project itself and can be accessed
through the related GitHub repository.10 A detailed docu-
mentation with instructions for musicians, as well as setup
instructions for system administrators can be found in the
corresponding GitHub pages. 11 Due to the flexibility, this
solution might thus be applicable in related setups and use
[1] A. Lindau, R. Kopal, A. Wiedmann, and S. Weinzierl,
aumliche Schallfeldsynthese f¨
ur eine musikethnolo-
gische Ausstellung: Erfahrungen aus Produktion und
Rezeption,” in Proceedings of the Inter-Noise 2016:
45th International Congress and Exposition on Noise
Control Engineering: towards a quiter future, 2016.
[2] M. Kronlachner, “Plug-in suite for mastering the pro-
duction and playback in surround sound and ambison-
ics,” Gold-Awarded Contribution to AES Student De-
sign Competition, 2014.
[3] K. L. Hagan, “Textural composition: Aesthetics, tech-
niques, and spatialization for high-density loudspeaker
arrays,” Computer Music Journal, vol. 41, no. 1, pp.
34–45, 2017.
[4] H. Wierstorf, A. Raake, M. Geier, and S. Spors, “Per-
ception of focused sources in wave field synthesis,
Journal of the Audio Engineering Society, vol. 61, no.
1/2, pp. 5–16, 2013.
[5] H. von Coler, “A System for Expressive Spectro-spatial
Sound Synthesis,” Ph.D. dissertation, 2021.
[6] M. A. Baalman and D. Plewe, “WONDER - a Soft-
ware Interface for the Application of Wave Field Syn-
thesis in Electronic Music and Interactive Sound Instal-
lations,” in Proceedings of the International Computer
Music Conference, 2004.
[7] J. Ahrens, M. Geier, and S. Spors, “The SoundScape
Renderer: A unified spatial audio reproduction frame-
work for arbitrary rendering methods,” in Proceedings
of the 124th Audio Engineering Society Convention.
Audio Engineering Society, 2008.
[8] T. Carpentier, “Panoramix: 3d Mixing and Post-
production Workstation,” in Proceedings of the Inter-
national Computer Music Conference (ICMC), 2016.
[9] F. Grond and P. Lecomte, “Higher Order Ambisonics
for SuperCollider,” in Proceedings of the Linux Audio
Conference 2017, 2017.
[10] H. von Coler, N. Tonn¨
att, V. Kather, and C. Chafe,
“SPRAWL: A Network System for Enhanced Inter-
action in Musical Ensembles,” in Proceedings of the
Linux Audio Conference, 2020.
[11] A. Heller and E. Benjamin, “Design and implementa-
tion of filters for Ambisonic decoders,” in Proceedings
of the 1st International Faust Conference (IFC), 2018.
[12] Orlarey, Yann and Fober, Dominique and Letz,
Stephane, “Syntactical and semantical aspects of
Faust,Soft Computing, vol. 8, no. 9, pp. 623–632,
ResearchGate has not been able to resolve any citations for this publication.
Conference Paper
Full-text available
The SPRAWL system is an audio and information routing network, enabling enhanced ways of interaction for musical ensembles. A Linux audio server with a SuperCollider mixing and spatialization system is connected to several access points via Ethernet, using Jack-Trip for audio transmission. The unified access points are based on a Raspberry Pi 4 with a 7 inch touch screen and a USB audio interface , running SuperCollider, Puredata and additional tools. Using custom applications on these units, musicians are able to control the routing and spatialization of the system, allowing personal monitoring , the distribution of click tracks and related auditory information, as well as visual information and cues. The system is designed as a meta-instrument for developing individual configurations and software components for specific compositions and use cases. It has been successfully used in on-site concerts and ported to wide area networks for remote rehearsals and performances.
Full-text available
Wave Field Synthesis (WFS) allows virtual sound sources to be synthesized that are located between the loudspeaker array and the listener. Such sources are known as focused sources. Due to practical limitations related to real loudspeaker arrays, such as spatial sampling and truncation, there are different artifacts in the synthesized sound field of focused sources. In this paper we present a listening test to identify the perceptual dimensions that are associated with these artifacts. Two main dimensions were found, one describing the amount of perceptual artifacts and the other one describing the localization of the focused source. The influence of the array length on these two dimensions is evaluated further in a second listening test. A binaural model is used to model the perceived location of focused sources found in the second test and to analyze dominant localization cues.
Full-text available
This paper presents some syntactical and semantical aspects of FAUST (Functional AUdio STreams), a programming language for real-time sound processing and synthesis. The programming model of FAUST combines two approaches: functional programming and block-diagrams composition. It is based on a block-diagram algebra. It as a well defined formal semantic and can be compiled into efficient C/C++ code.
This article documents a personal journey of compositional practice that led to the necessity for working with high-density loudspeaker arrays (HDLAs). I work with textural composition, an approach to composing real-time computer music arising from acousmatic and stochastic principles in the form of a sound metaobject. Textural composition depends upon highly mobile sounds without the need for trajectory-based spatialization procedures. In this regard, textural composition is an intermediary aesthetic—between “tape music” and real-time computer music, between sound objects and soundscape, and between point-source and trajectory-based, mimetic spatialization. I begin with the aesthetics of textural composition, including the musical and sonic spaces it needs to inhabit. I then detail the techniques I use to create textures for this purpose. I follow with the spatialization technique I devised that supports the aesthetic requirements. Finally, I finish with an example of an exception to my techniques, one where computational requirements and the HDLA required me to create a textural composition without my real-time strategies.
Räumliche Schallfeldsynthese für eine musikethnologische Ausstellung: Erfahrungen aus Produktion und Rezeption
  • A Lindau
  • R Kopal
  • A Wiedmann
  • S Weinzierl
A. Lindau, R. Kopal, A. Wiedmann, and S. Weinzierl, "Räumliche Schallfeldsynthese für eine musikethnologische Ausstellung: Erfahrungen aus Produktion und Rezeption," in Proceedings of the Inter-Noise 2016: 45th International Congress and Exposition on Noise Control Engineering: towards a quiter future, 2016.
Plug-in suite for mastering the production and playback in surround sound and ambisonics
  • M Kronlachner
M. Kronlachner, "Plug-in suite for mastering the production and playback in surround sound and ambisonics," Gold-Awarded Contribution to AES Student Design Competition, 2014.
A System for Expressive Spectro-spatial Sound Synthesis
  • H Coler
H. von Coler, "A System for Expressive Spectro-spatial Sound Synthesis," Ph.D. dissertation, 2021.
WONDER -a Software Interface for the Application of Wave Field Synthesis in Electronic Music and Interactive Sound Installations
  • M A Baalman
  • D Plewe
M. A. Baalman and D. Plewe, "WONDER -a Software Interface for the Application of Wave Field Synthesis in Electronic Music and Interactive Sound Installations," in Proceedings of the International Computer Music Conference, 2004.
The SoundScape Renderer: A unified spatial audio reproduction framework for arbitrary rendering methods
  • J Ahrens
  • M Geier
  • S Spors
J. Ahrens, M. Geier, and S. Spors, "The SoundScape Renderer: A unified spatial audio reproduction framework for arbitrary rendering methods," in Proceedings of the 124th Audio Engineering Society Convention. Audio Engineering Society, 2008.
Panoramix: 3d Mixing and Postproduction Workstation
  • T Carpentier
T. Carpentier, "Panoramix: 3d Mixing and Postproduction Workstation," in Proceedings of the International Computer Music Conference (ICMC), 2016.