A B S T R A C T S

CREATE symposium on SOUND IN SPACE
18-19 March 2000

University of California
Santa Barbara

Curtis Roads (CREATE, University of California, Santa Barbara) - History of sound spatialisation · The art of sound spatialization has assumed a similar position today as the art of orchestration had in the nineteenth and twentieth centuries. To deploy space is to choreograph sound: positioning sources and animating movement. Immersing sound in reverberation, we bathe listeners in its lush ambience. Sound spatialization has two aspects: the virtual and the physical. In the virtual reality of the studio, composers spatialize sounds by imposing delays, filters, panning, and reverberation­lending the illusion of sounds emerging from imaginary environments. Sometimes these virtual spaces take on characteristics that would be impossible to realize architecturally, such as a continuously changing echo pattern. In the physical world of concert halls, sounds can be projected over a multichannel sound system from a variety of positions: around, above, below, and within the audience. This paper presents an historical review of sound spatialisation for music.

Stephen T. Pope (CREATE, University of California, Santa Barbara) - The state of the art in sound spatialisation · There are several aspects to the field of spatial sound, each of which pose different chalenges and offer different potential applications. Although our understanding of aural perception is still incomplete, we are able to both synthesize and record spatial sound fields, and to render sound such that the fidelity of localization is very high (for a specific listener). There are several well-known and effective techniques for creating the perceptual cues that our brains use to localize sound, but the systems that scale well to large spaces or to many listeners are not the same ones that give the best localizational fidelity. The formal study of spatial sound performance in larger space (e.g., concert halls) is still in its (relative) infancy. Most work in this area has been ad hoc, treating the spatial sound performance situation more as an instrumental performance than as a controlled experiment. This presentation will explore the aspects of aural perception that contribute to the difficulties, and the potential, in the recording and playback of spatial sound, and will survey the current techniques used in this area. The slides for this presentation are available in this PDF file.

Gerard Pape (Centre de Création Musicale «Iannis Xenakis», Paris) - Form, space, time · What is the relation between form, space and time in the musical composition? Form implies architecture, which implies space. Space is that which surrounds us. A house, if it is not in the process of being built or modified, has a fixed space/time relation, for example. A spatial object which changes shape over time is a mobile architecture. Music implies form, space, and time. A musical form, a sound structure, is a mobile architecture which exists in a moving (i.e. relativistic) space/time relation. In the course of a musical composition, the listener traverses, via the composer's macro-formal scheme, several different spatial-temporal regions, experienced subjectively as time-varying, sound structure architectures. A composer is also an architect, in that he composes sound structure architectures, each with its own unique, multi-dimensional, topologically-varying shape. From the listener's perspective, music is a kind of spatial object, for it surrounds the listener, evoking a time varying, mobile spatial perception.

Peter Otto (CRCA, University of California, San Diego)- New MacOS-based sound spatialisation processes and interfaces · Sound spatialisation is the process of creating the perceptual illusion of a continuous and physically realistic surround soundfield, in which source sounds can be virtually placed and arbitrarily moved all around the listener in accordance with control information. This paper describes an approach to computer sound processing and control within multiple speaker playback setups, and the eventual combination of this system with technologies for improving the sound quality of the actual playback space. Computation and control methods are prototyped and evaluated using Max/MSP, with the ultimate goal being the refinement of a realtime spatial system conducive to a variety of beneficial applications, such as audio mixing for live and stored audio applications, VR paradigms, teleconferencing, sound for picture, etc. This research aims to refine current sound processing models, define systems for implementing control functions for sound placement, provide graphic user controls, and in general, provide a model system for next-generation spatial sound. Examples of new Macintosh based spatialization processing and control will be demonstrated as a part of the presentation.

David G. Malham (Music Technology Group, University of York) - Ambisonic systems for the spatialisation of sound · Systems based on Ambisonic technology have been used for the diffusion of electroacoustic music for more than a decade. The systems used in the past fed arrays of loudspeakers from four channels of audio coded according to first order spherical harmonic descriptions of the required soundfield. Newer variants of the approach offer a significant improvement upon the original version as a result of the inclusion of higher order harmonics. This enhances the resolution of the system, considerably increases the area over which it works well and allows the composer far more control over the "sharpness" or "diffuseness" of the sound source. Previous work on these higher order systems had largely been limited to theoretical studies. However, recent advances in multichannel computer audio systems have made it economically viable to use many more than four output channels and this has encouraged the development of packages which allow composers to spatialise sounds using second and higher order systems. Whilst the earlier theoretical studies limited their scope to two dimensional, horizontal plane only systems, this paper deals with the design and implementation of fully three dimensional systems using either a partial or a full set of second order harmonics. System requirements, the advantages and disadvantages of the Ambisonics approach and the implications for the composer will be discussed. http://www.york.ac.uk/inst/mustech/3d_audio/SiS2000.html

David Wessel (CNMAT, University of California, Berkeley) and Ville Pulkki (Acoustics Lab, Helsinki U of Technology) - Generic 3-D panning tool for the MAX/MSP computer music programming environment · A generic tool for virtual source positioning has been implemented to MAX/MSP software. The tool can be used to produce similar soundscapes for loudspeaker setups of arbitrary number of loudspeakers in arbitrary directions. The tool is based on a panning method called vector base amplitude panning (VBAP). VBAP is pair-wise or triplet-wise amplitude panning method. The VBAP tool is initialized by setting the directions of current loudspeakers. Virtual source positioning is performed by setting the desired direction of a virtual source, the tool selects a proper subset of loudspeakers and applies the sound signal to them with calculated amplitudes. Some examples of use of panning tool on simple panning and on creation of virtual environments and instruments are described.

Frank Ekeberg (Music, City University, London) - Spatio-structural analysis, towards a framework for describing and analysing space in electroacoustic music · In my research I have developed a framework for describing and understanding space as a structural element in electroacoustic music. In this paper I propose an explanation of my research which reveals that that space can be a fundamental element of and a carrier of meaning in electroacoustic music. In the course of my research I have found it necessary to develop a vocabulary for describing spatial components in music, and have furthermore developed an analytical model of musical space which allows for clear identification of the various spatial components used in the structural composition of electroacoustic works. My research is grounded on listening-based analyses of a number of compositions, and is further supported by Denis Smalley's Spectromorphological theory and by research in the fields of Gestalt psychology, perception and spatial hearing. In my model, individual sound objects are studied in terms of their intrinsic space (magnitude, shape, density), extrinsic space (placement, distance, movement) and spectral space (vertical placement of sounds based on their concentration of spectral energy). Combined these components form a composed space, which is the spatial organisation of individual sounds and events into a musical context. My investigation into composed space focuses on the spatiomorphological behaviour of individual sound objects and virtual spaces as well as their interrelationships in order to reveal structural functions of space in the work.

Pedro Rebelo (Music, University of Edinburgh) - On the coherence of sonic space · Sound, as a physical phenomenon, is dependent on space for its articulation. This paper attempts to define a number of issues related to the creative use of sound in relation to space. Recent technology stimulates an integration of a number of conditions previously confined to disciplines such as music, architecture, and sculpture. The author's creative work in installation art and composition is used as a platform for examining structures which are able to provide modes of interaction between sonic worlds and physical and virtual space. The use of acoustic measurements of spaces has become a central in the author s compositional processes. A work in progress involving the use of visually rendered virtual spaces as sonic modulators is analysed in detail in both its technical and aesthetic implications. Inter-media coherence is drawn from spatial perception in physical and virtual worlds; the architectural and acoustic experience of moving through space

Larry Austin (gaLarry, Denton) - John Cage's Williams Mix, a new realization · This lecture is on my research, analysis, restoration, and new realization of John Cage's octophonic, surround-sound tape composition, Williams Mix (1951-53), duration 4:15, realized from its 192-page score. I will demonsrate the restored, original octophonic version, originally performed recorded on eight monaural tapes and performed on eight monaural tape machines with speakers surrounding the audience. The author has made digital copies from and restored the eight original 30-ips reel-to-reel production masters, transferred them to an eight-track ADAT for accurate synchronous playback. The John Cage Trust graciously provided me with a color-Xerographic copy from The Archive of the Trust of Cage's original 192-page score for the creation of the piece, as well as associated sketches and commentary by Cage on the compositional process involved in his (and his collaborators') original (and only) realization for eight magnetic tapes. My project is conceived as a new realization for octophonic tape of Cage's original score plus the creation of a set of variations based on the Williams Mix compositional concept. The current project is sponsored, in part, by the IIEM, Bourges, France, which has awarded a commission to the composer for a residency to complete the work in the year 2000.

Curtis Roads and Alberto de Campo (CREATE, University of California, Santa Barbara) ­ Demonstration of the Creatovox synthesizer · This is the first public presentation of the Creatovox, a new digital instrument for virtuoso performance based on the synthesis and transformation of sound particles. The particles are of various types: synthetic and sampled grains, glissons, and pulsars. The instrument is controlled by a MIDI keyboard, thumbwheels, joystick, sliders, and foot pedals. Spatialisation of the particles in multiple channels is integral to the design of the Creatovox.

Corey Cheng and Gregory Wakefield (Electrical Engineering and Computer Science, University of Michigan) - Moving sound source synthesis for binaural electroacoustic music using interpolated head-related transfer functions (HRTFs) · This paper describes a recently developed interpolation algorithm for Head-Related Transfer Functions (HRTF's) and its application to the synthesis of moving sounds for some recently composed electro-acoustic music. By exploring the perceptual quality of the algorithm's output for several different types of sounds, we show that the algorithm has strengths and weaknesses congruent to those predicted by psychoacoustic and engineering literature. By interpreting the algorithm's strengths and weaknesses as limitations of a compositional tool, we suggest that the limitations of HRTF-based spatialization algorithms can have a significant bearing on the practical implementation of spatially-based musical ideas. Specifically, we propose some simple guidelines by which HRTF-based spatialization can be used to take advantage of its strengths while avoiding some obvious weaknesses. Specifically, the composer's HRTF's were measured for several hundred spatial locations in an anechoic chamber by Dr. John Middlebrooks at the University of Michigan Kresge Hearing Research Institute. An interpolation method was developed by which the measured HRTF's could be used to synthesize a binaural sound from a monaural sound traversing a simulated, user-defined spatial trajectory. A simple Graphical User Interface (GUI) was written in MATLAB as a front-end to this algorithm to process several sounds, which were then included in a few short pieces of electro-acoustic music. In this paper, we give a brief introduction to HRTF's, describe our method of HRTF interpolation, and describe the GUI. We report our findings on how the algorithm was used to realize some spatially-based compositional ideas both successfully and unsuccessfully. Finally, we invite the audience to listen to some of the results over headphones, both as individual sounds and as final pieces.

Fernando and Jose Ramon Beltran (EEC, University of Zaragoza) - Virtual space reverberation simulation using a multispeaker system · In this work we have developed a virtual space reverberation simulation using a multi-speaker system. In a first step, for a 3D simple-geometry virtual space where the sound source and the listening position are placed, we compute the impulse response with the source image method. Each virtual sound source can be placed in a geometrical point in the walls of the room. We can compute the corresponding panning to locate this sound source with a multi-speaker system. In our first implementation, we have used a four loudspeakers system in the corners of the room. We have done a projection over the X-Y plane to get the geometrical points where each rebound of the early reflection is placed. For each rebound, the associated delay is computed, and its energy is assigned to the corresponding loudspeaker depending on the geometrical position. With this computation we obtain four impulse response -IR- associated with each loudspeaker. We match the IR of each loudspeaker to one reverberation algorithm and we generate the four different output signals. We have used three different reverberation algorithms (Gardner, Dattorro, and Jot) to compare the results. We have added some resonant filters to simulate the low frequency reponse of the room. Limitations arise from the equipment needed to work with more channels. Listening results using four channels can be presented in the symposium.

Yon Vissel (Physics, University of Texas, Austin) - Audio spatialisation for quantum spaces · Noncommutative geometry has emerged in recent years as a compelling framework lending order to diverse problems in mathematics and physics, much in the way that conventional geometry has done throughout this century. This is true even though, as geometers, we still have limited intuition and only indirect experience with such generalized, quantum spaces - most commonly as the phase spaces of quantum mechanics. Whereas the objects of classical differential geometry may be visualized by means of an embedding in a (flat) Euclidean space of an appropriate dimension, no such visualization exists for quantum spaces. The pursuit of alternate representations leads to an unusual problem in audio spatialization - that of embedding sounds in virtual, quantum spaces. The cleanest class of examples, that of the quantum torus, is also one amenable to direct implementation in the digital realm, as presented here. The techniques employed may be understood as quantum modifications of standard signal processing operations.

Gerhard Eckel (Virtual Environments Division, GMD, Sankt Augustin) - Sound sculptures · Sound projection is very limited by the uniformity of loudspeaker radiation patterns. All sounds produced by loudspeakers are projected into acoustic space with the same anonymous radiation pattern. Most natural sound sources are characterized by highly frequency-dependent radiation patterns creating a strong sensation of presence in acoustic space due to the idiosyncrasy of their acoustic signature. Electroacoustically reproduced natural sounds as well as processed or synthetic sounds lack such a signature, which is easily recognized by the auditory perception system. The sound sculptures project is concerned with the design of sound rendering systems capable of creating complex and time-varying radiation patterns with arrays of loudspeakers and special digital signal processing software for control. Such active acoustic antennas allow for an independent control of several virtual acoustic objects rendered at once. Two acoustically rich and perceptually very compelling sound sculptures have been produced in the context of the project so far. The kinetic sound sculpture "Stele" had been installed at GMD´s Birlinghoven Castle in June 1998. A further development of the "Stele", named "Viola Spezzata", has been exposed at the Contemporary Art Museum of the City of Bonn in summer 1999 and was very well received by the audience. In my presentation I will present the two sound sculptures, explain how they have been realized and discuss plans for future installations.

Dan Overholt (MIT Media Lab) - The Flying Violin · The Flying Violin is a spatial piece for solo violin using Max/MSP and SPAT. This talk explains the technology behind the piece and demonstrates it in performance.

Jean-Marc Jot (Emu Creative Technologies) - Presentation of the IRCAM spatialisateur * The Spatialisateur is library of signal processing and control interface modules for real-time spatial processing of sounds, available in Max/MSP or jMax. The library includes the Spat~ object, a configurable and modular real-time spatial processor integrating the localization of sound events with environmental reverberation and distance effects. This object can be easily configured for a chosen 3D positional encoding technique and loudspeaker layout (or headphones). Several Spat~ processors can be associated in parallel in order to process several source signals simultaneously. The reverberation engine complexity can be configured according to the application and the available digital signal processing resources. The desired effect is specified independently from this configuration and is, as much as possible, preserved from a reproduction mode or listening room to another. The control interface uses a perceptual paradygm allowing to control a set of mutually independent spatial attributes for each Spat~ object, including the sound event's direction, distance and orientation, the room reverberation decay characteristics, and the effects of early reflections. More information is available at http://www.ircam.fr/produits/logiciels/log-forum/index-e.html.

James Harley (Music, Moorhead State University) - Heterophonic stereophony and the illusion of motion: a case study · As a composer of both electroacoustic and acoustic music, I have developed an interest in the musical and acoustical possibilities of heterophonic textures. By this term is meant simultaneous streams of similar, but not identical, material, be it melodies, rhythmic patterns, textures, or audio signals of various types. Originally, the aim was to create dynamic musical textures of some "width," or "depth," through the use of multiple voices intentionally out-of-phase or sync. In the acoustic realm, judicious spatial placement of the performers or ensembles will emphasize this aspect of the music, as the broadly similar signals reach the listener from different points. This talk discusses the heterophonic spatial strategies of my pieces Cantico delle creature (1993), for mezzo-soprano, chamber choir, chamber orchestra, and computer-generated sounds, and Voyage (1986) for four-channel tape, realized on the UPIC at CEMAMu. The compositional approach is conceived in order to take advantage of the spatial separation of the signals. The shared cues given by different groups or channels presenting similar, heterophonic material adds a sense of breadth and dynamism to the music that could not be achieved otherwise.

Maria Anna Harley (Thornton School of Music, University of Southern California) - From circles to nets: on the signification of spatial sound imagery in new music · This paper presents a historical and aesthetic analysis of two types of spatial designs used in contemporary music (instrumental, vocal and electroacoustic) - the circle (and, by extension, the sphere) and the net (or web; of a greater spatial complexity and multidimensionality). Circular designs are illustrated with examples from works by Xenakis, Oliveros, Schafer, and Serocki. Here, simplicity and uniformity of musical material are often accompanied by a "mystical" intention of the composer who attempts to influence the mental state of the listeners. Enveloped in sonorities, surrounded by sounds, the audiences are let on a path to contemplation or even trance (resulting from a confluence of musical and spatial features). In the second design, composers reflect the growing awareness of the network-like structure of the physical and human worlds. Selected designs by Boulez, Brant and Ptaszynska are discussed as examples of the heterogeneity of musical material and the diversity of its interconnections. The author draws from concepts of "intimate immensity", "phenomenology of roundness", and "dialectics of space" (Gaston Bachelard's Poetics of Space) and notions of "absolute/abstract space" and "contradictory/differential space" (Henri Lefebvre's The Production of Space).

Alberto de Campo (CREATE, UC Santa Barbara) - Spatial processing in SuperCollider · This talk presents an overview of spatialisation techniques using James McCartney's SuperCollider 2 langage . The focus is primarily on adapting and extending standard reverberation and spatialisation for the special possibilities SuperCollider's integrated environment. The application is a "microreverberation" algorithm that spatialises individual particles in granular synthesis.

Frank Pecquet (Institut d'Esthétique, University of Paris I) - Space and sound representation · En acoustique on dira que les molécules d'air gravitent dans l'espace pour se propager d'un point à l'autre en fonction des variations de pression exercée dans l'air. L'atmosphère tient lieu de réceptacle des molécules d'air, de lieu si l'on veut. Si la perturbation mécanique est produite par des oscillations comprises approximativement entre 20 et 20000 Hertz par seconde, on entendra un son. Pourvu que ces oscillations soient suffisamment fortes pour que nous puissions les détecter et suffisamment faibles pour ne pas détruire notre mécanisme perceptif. Après une impulsion acoustique - claquement de main par exemple, l'onde, résultant des variations de pression de l'air, se déplace de la source à une certaine distance, dans une certaine direction et à une certaine vitesse - 330 mètres par seconde à 0 degré Celsius. Bien entendu elle se propage différemment selon la matière ; bois, verre ou métal sont autant de matériaux différents permettant à la célérité du son de varier sensiblement. Plus l'impulsion est forte plus l'onde se propage loin, à condition qu'elle ne rencontre pas d'obstacle. La distance parcourue par l'onde est tout à la fois relative à l'intensité du choc, à sa taille qu'au milieu dans lequel elle évolue. En effet si le son se définit substantiellement par rapport à l'air, que nous interprétons ici comme l'espace réceptacle, l'environnement en constitue le contexte acoustique déterminant. Lieu, déplacement, distance sont autant de notions qui caractérisent le son au centre de notre problématique sur l'espace. L'air est au son ce que le son est à la musique, sa substance. Si le son, en tant que réalité acoustique, dans sa génération comme dans sa propagation, ne représente pas de lui-même la musique, il en est le médium indispensable. En ce sens il semble juste de dire qu'il y a déjà, à l'origine, pour qui pense la musique, une pensée de l'espace.

Francois Pachet, Olivier Delerue, Peter Hanappe (Sony Computer Science Lab, Paris) - MusicSpace goes audio · The MusicSpace project aims at providing high-level user control on music spatialization, i.e. the position of sound sources and the position of the listener's avatar. This is done by introducing a constraint system in a graphical user interface representing the sound sources, and connected to a spatializer. The constraint system allows to express various sorts of properties on configuration of sound sources. When the user moves one source - through the interface or via a control language - the constraint system is activated and tries to satisfy the constraints that may have been violated. A first Midi version of the MusicSpace has already been designed and proved very successful. We describe here a second version of MusicSpace which now handles full-fledged multi-track audio files. We report on the design of the system and preliminary experiments.

David Eagle (University of Calgary) - Sound diffusion with the aXiO, AudioBox, and Sound Travels · The aXiO (alternative eXpresssive input Object ) is a new electroacoustic instrument/controller which gives the musician a broad range of expresssion and multi-dimensional control of MIDI synthesizers and samplers. It was designed and built by Brad Cariou at the University of Calgary and conceived to provide digital artists with the intimate control and flexibilty needed to express themselves in various new media. Using a Macintosh computer running the MIDI progam MAX, it is completely user-programmable. The aXiO has a cross-like structure and stands about body height, vaguely resembling a stream-lined robot. It is played with both hands on three distinct playing surfaces: a sophisticated joystick for the left hand, a velocity-sensitive keyboard for the right hand, and, running up the musician's shoulder, is an array of buttons used to change voices or trigger musical sequences, images or video. In performance, the right hand selects and plays the notes or events with aftertouch and pitch bending capabilities while the left hand provides more expression and transformation of the sound. For instance, the left hand joystick could be used to move a sound in space, to change its overall volume, to sustain a sound, to change its colour, and to do it all simultaneously in real time. The aXiO has the potential to transform electroacoustic music from a predominantly studio art to a performance art. Concluding lecture

David G. Malham (Music Technology Group, University of York) - Future possibilities for the spatialisation of sound · Throughout the ages, there have been composers who have been interested in developing a spatial element within their compositions. However, because of the limitations imposed by acoustic instruments, this has tended to be very limited. With the development of electronic media and the emergence of the electroacoustic genre, there are now many more composers exploring this area. Composers of electroacoustic music have long been in the forefront of the development of multichannel systems for the spatialisation of sound but apart from the abortive seventies foray into Quadraphonics it is only recently, with the appearance of various "surround" systems on film soundtracks, that the commercial world has begun to catch up. However, none of the systems currently in use, either commercially, experimentally or in electroacoustic music, are close to being a true match to the capabilities of our ears. This paper examines three related leading edge technologies, Holophonics, Wavefield Synthesis and Ambisonics and asks where they might lead in the 21st Century. From this examination, the Hyper-Dense Transducer Array emerges as a prime candidate for the "ultimate" sound spatialisation system.

Contact:

	JoAnn Kuchera-Morin and Curtis Roads
        CREATE
        Department of Music
        University of California
        Santa Barbara, California 93106
	Telephone (805) 893-8352, Fax (805) 893-7194

        info@create.ucsb.edu