TAXONOMY AND NOTATION OF SPATIALIZATION
Emile Ellberger Germán Toro Pérez Linda Cavaliero Institute for Computer Music Institute for Computer Music Institute for Computer Music and and Sound Technology and Sound Technology Sound Technology Zurich University of the Arts Zurich University of the Arts Zurich University of the Arts [email protected] [email protected] [email protected]
Johannes Schütt Giorgio Zoia Basile Zimmermann Institute for Computer Music Institute for Computer Music Faculty of Humanities and Sound Technology and Sound Technology University of Geneva Zurich University of the Arts Zurich University of the Arts [email protected] [email protected] [email protected]
ABSTRACT that include implementation of spatialization have been limited in their capacity to propose adequate notation The SSMN Spatial Taxonomy and its symbols libraries, possibilities to convey spatial information within musical which are the corner stone of the Spatialization Symbolic scores. In spite of the availability of a variety of strategies Music Notation (SSMN) project, emanates from research and tools for spatialization within the context of electroa- into composers’ attitudes in this domain. It was con- coustic music composition, decisions about position and ceived as the basis for the development of dedicated movement of sound in space, or the general question of notation and rendering tools within the SSMN project. virtual space quality remain often a secondary formal The taxonomy is a systematic representation of all rele- issue; in many cases these decisions are left to a post- vant features necessary to specify sound spatiality: shape production stage instead of being fully integrated and acoustic quality of the space, structure, position and throughout the composition process. This situation can movement of sound sources. It is based on single de- marginalize spatialization to an ornamental aspect that scriptors that can be combined in order to define complex can be adapted or reduced without affecting musical spatial configurations. Descriptors can be transformed substance. locally and globally and can be the object of structural On the other hand, performers engaged in the and behavioral operations. The SSMN Spatial Taxonomy interpretation of music involving electroacoustic proposes a corresponding graphic symbolic representa- spatialization (and other kinds of signal processing) find tion of descriptors, operations and other functional ele- mostly a reduced graphic representation of the ongoing ments facilitating the communication of creative ideas to processes in the score [1]. According to the experience of performers and technical assistants. the authors during several years of performance practice This paper focuses on the presentation of the taxonomy the notation of electroacoustic events prioritize mostly and the symbols. Additionally it describes the workflow cue numbers and synchronization events. This limits the proposed for using symbols inside a notation software possibility of a more intimate interaction within the prototype developed within the project. Finally, further performance situation. In addition, the lack of a aspects concerning the actual and future developments of spatialized acoustic feedback while studying prevents SSMN are mentioned. performers from preparing a piece taking into account sound motion. This issue becomes especially relevant when considering the usual restrictions of rehearsal time 1. INTRODUCTION in performance spaces. The need for a graphical representation of spatialization The field of sound representation has undergone continu- within the context of sound diffusion of electroacoustic al development throughout the history of creative arts. music in concert has been also addressed with arguments The issue of sound motion representation, as concerns pro [2] and contra [3]. Nevertheless a generic and composers, has however hardly been studied. Composers practical way to accurately notate spatialization has not have been continually obliged to reinvent strategies for been formulated yet. Even meticulous spatial notation as communicating their ideas of spatial movement of sound, in Stockhausen’s Oktophonie [4] using sequences of performers’ displacement, and description of the perfor- channel numbers instead of symbols –as in the mance space. In fact, even today’s musical software tools introductory notes to the score– is difficult to read for performers. Finally, when audio engineers collaborate with com- Copyright: © 2016 First author et al. This is an open-access article posers preparing compositions within a multi-channel dis- tributed under the terms of the Creative Commons Attribution environment, they have to overcome the difficulties of License 3.0 Unported, which permits unrestricted use, distribution, and interpreting placement of sound in space as imagined by reproduction in any medium, provided the original author and source are credited. the composers, who typically invent a personal system of focuses mainly on spectral structure [21] or develops a graphical or textual explanations. perceptual approach to the analysis of acoustic scenes The aim of SSNM is thus to open a new approach of [22]; Vandergorne’s spatial categories and figures are substantial integration of spatial relationships and spatial specifically concerned with sound diffusion [8]; UST processes in musical thinking as well as in composition, (Unités Sémiotiques Temporelles) [23] are obviously rehearsal and performance practice. For this purpose focused on temporal meaning. In our opinion, a taxono- SSMN has defined a typology of spatial movements and my for notation of spatialization should ideally be univer- designed a library of symbols to represent them. In order sal, generic and based on low-level structural features to enable its use in creative processes, an open source that can be represented through symbols. The terminolo- software tool that integrates this library within a common gy should emanate directly from musical practice and be western musical notation context is being developed, as much as possible self-explaining. The work presented allowing editing and acoustic feedback through a render- in the next section was developed under these premises. ing engine. Composers are thus able to use and edit sym- bols describing spatialization in a notation program and 3. TAXONOMY immediately hear the results. Performers are given full information on spatialization in the score and can hear the 3.1 Preliminary considerations results from the beginning of the studying process. The basic units of the SSMN spatial taxonomy are called 2. SCOPE OF RESEARCH descriptors. There are two kinds of descriptors: room descriptors and descriptors of sound sources. Sound During the preparatory stages of the SSMN Spatial Tax- sources can be physical root sound (RS) like instruments onomy, research has focused on the following: and voices or projected audio signals (PA) like micro- • Musical scores containing verbal or graphical annota- phone signals, audio files and streamed audio. tions of spatial indications, focusing on spatialization Descriptors can be simple or compound and are as- and extended notation in contemporary music since sumed to be perceptually relevant, but definitive percep- 1950 [5] tion depends on the interaction between the actual sound • Artistic performance practice wherein spatial placement and the actual spatial configuration. Although descriptors and/or mobility of live performers is relevant to musical are primarily defined in structural (geometrical, mathe- compositions as well as composers' means of express- matical, acoustical) terms, they have been conceived in ing placement and/or motion in space [6], [7] view of musical practice. • Spatialization in electroacoustic media [8] Simple descriptors are the basic atoms of the SSMN • Extended musical notation [9] spatial taxonomy. They are able to denote all single pri- • 2D and 3D symbolic notation [1] mary features relevant to sound spatiality and can be • Typologies of spatial qualities of sound [10] represented as symbols. Compound descriptors are arrays • Spatial perception of sound placement, sound motion of simple descriptors. They are used to represent more and physical space [11] complex spatial configurations and processes (e.g. pat- • Semiotics and epistemology of notation [12] terns, figures, motives, etc.) and can also be represented • Relevant programming languages, audio and graphic as symbols. design software including Csound, PD, Iannix, Super- Descriptors can have several properties that are finally Collider, Max/MSP, Illustrator, WFS [13] [14], Ambi- defined through names, numeric parameters and flags. sonics, IRCAM’s OpenMusic [15] & Antescofo [16], For instance, the descriptor "Position of loudspeakers" inScore [17] has the entry "labels" to name specific items, the parame- • Spatial notation in other fields, especially dance, aero- ters "position" given as Cartesian or spherical coordinates nautics, geographical migration patterns, and theater and "inclination" (yaw) defined as angle and the flag staging [18] "interior" or "exterior" according to their position inside In a nutshell, the specifications for the SSMN research or outside the room. For reasons of clarity, parameter units as well as some parameters and flags will be omit- project are based on a four-pronged study: (a) spatial ted in this paper. typo-morphology resulting in the SSMN Spatial Taxon- The third part of the taxonomy is dedicated to opera- omy, (b) design of symbols, (c) integration of symbols tions, also called modifiers. They can be used to trans- and trajectory editing in notation software, (d) integration form elements previously defined using single or com- of notation software in a rendering engine. So far, an pound descriptors or to generate new elements. For in- introduction to the project was first presented at stance, the basic structural operation "Scaling" can be ISMC|SMC2014 [19] followed by a poster presentation at used to multiply a given parameter or group of parame- TENOR 2015 [20]. ters by a certain factor, "Repetition" to repeat a com- Concerning the specific issue of a spatial taxonomy the pound trajectory previously made out of single segments contributions referred to above present important ideas as straight lines and curves. Global operations can be but are limited in the sense that they were developed in used to generate relationships between complex unities view of specific aspects and purposes different from like sequences and superpositions of existing compound those of notation. Trochimczyk’s [10] classification of trajectories. Cross-domain interactions can be used to spatial designs comes closer to our needs but is con- rule relationships between spatial audio information and sciously limited to certain instrumental setups; Smalley other media like synchronization with visual or choreo- 3. Position of Microphones graphic sequences. a. Referred to an instrument Finally, behavioral relationships like "co-incidence" or name of instrument "attraction" inspired by social and biological movement b. Referred to the space patterns and observed in other contexts (see 3.3) could label, position, inclination help to envision a new paradigm of sound spatiality based on processes rather than geometrical or visual structures. c. Referred to specific movements (e.g. swinging This aspect is not fully integrated in the taxonomy yet microphone) and suggests a promising research direction. 4. Position of loudspeakers As mentioned above, the SSMN spatial taxonomy is a. Fixed intended to become universal and generic. At the moment label, position, direction, inclination not all descriptors have been defined as symbols and not all symbols have been implemented within the software b. Variable (mechanic or human driven) prototype. i. Line Although the taxonomy describes and classifies sound start/end, speed in a three-dimensional space, some objects and symbols ii. Arc are, for practical reasons (mainly rendering, and depend- start/end, curvature, speed ence on existing standardized formats), represented in two dimensions. iii. Pendular motion length, initial height, direction All projected audio can be subjected to speed, accelera- tion and the Doppler effect. Simple trajectories can be iv. Other (e.g. choreography) followed in two opposite directions. B. Spatial quality of the room After considering the wide number of possible curve types only a small number of them was explicitly ad- 1. Space definition dressed in the taxonomy. A detailed evaluation of their a. Open perceptual relevance remains to be done. b. Closed While sound sources can easily be defined as “points” or “groups”, a concept such as “sound plane” is an ab- c. Virtual straction of visual forms often used by composers but 2. Reverberation difficult to define in purely acoustical terms. We have a. Interaction source-room nevertheless integrated it into the taxonomy. The following section presents the complete work as an energy of first reflections related to direct almost self-explaining, structured list of descriptors and sound, energy of late reverberation, decay time operations. Explanatory comments are provided in brack- of primary reflections ets. Behavioral relationships will be discussed separately. b. Room perception (related to late reflections): decay time, heaviness (change of decay time of 3.2 Descriptors and operations low frequencies), “liveness” (change in decay I. Room descriptors time of high frequencies) A. Disposition II. Descriptors of sound sources 1. Shape of the room (generic shapes) A. Types of sound sources a. Cube 1. Sound points length, width, height a. Physical root sound (RS) b. Hemisphere label diameter, height b. Projected audio signal (PA) c. Church (cross form) label length, width, height 2. Groups d. Other shapes => Definition: A group is a collection of sound points dimensions with common spatial features. A group is defined by a perimeter wherein the single elements can be found. 2. Placement of performers, objects and audience Position and movement of single elements within the a. Fixed perimeter can be defined in the same way as single label, position points. b. Variable a. Root sound i. Line label, number of sources, position of reference start/end, speed point ii. Arc b. Projected audio start/end, curvature, speed label, number of sources, position of reference iii. Other (e.g. random, choreography) point 3. Planes (PA) 3. Free hand using interface => Definition: a plane is a homogenous sound spread III. Operations (transformation or generation of new out in space. trajectories from preexistent single or compound tra- label, shape jectories) B. Spatial quality of single sources A. Structural operations and modifiers 1. Perceived distance (PA) 1. Operations on single sound sources, groups and presence, brilliance, warmth (equalization) planes (position); on simple or compound trajectories C. Dimension of single sources a. Repetition 2. Scale number of reiterations => PA, perception of «bigger or smaller» than real b. Scaling sound source) factor scale factor c. Shift D. Localization of sound sources value d. Rotation 1. Localization of single points (PA, RS) roll, yaw, pitch position, direction, inclination, aperture (PA) e. Mirror (inversion) 2. Localization of groups mirror flag shape, geometrical center, position of each ele- f. Reverse (crab) ment, within the shape reverse flag 3. Localization of planes g. Palindrome (returns to the starting value) position, direction, inclination palindrome flag E. Simple trajectories of sound points, groups or planes h. Random random parameter 1. Linear i. Signals as modifiers a. Straight start/end i. Sinus frequency, amplitude, phase b. Polyline (open) segments, start/end ii. Triangle frequency, amplitude, phase c. Poly_closed (closed polyline) segments, start/end iii. Square frequency, amplitude, phase 2. Circular iv. Saw a. Circle frequency, amplitude, phase center point, radius, start/end angle, direction v. Other b. Slinky (named after the toy invented 1945 j. Simple or compound trajectories as modifiers by Richard James) label start/end center point, radius, start/end angle, direction 2. Operations on compound structures c. Spiral a. Sequencing start/end center point, start/end radius, b. Permutation start/end angle, number of rotations c. Interpolation (morphing) 3. Curve 3. Algorithmic defined functions based on externals a. Bézier algo (label) start/end, control points, reiterations b. Bézier_spline B. Global operations start/end, control points, reiterations 1. Global scaling (space, time) c. Béziergon (closed Bézier curve) a. Linear start/end, control points, reiterations b. Non-linear d. Bernoulli (lemniscate) start/end, control points, reiterations 2. Sequence (Horizontal) e. Other (e.g. Lissajoux, etc.) a. Loop F. Compound trajectories b. Cross c. Tight 1. Compound using simple trajectories d. Pause 2. Generic polygons (selection of basic shapes) 3. Superposition (Vertical) 6. Divergence, synchronous or delayed: movement 1. Synchronous start away from the same location. See also «breakup», p. 9 b. Delay 7. Attraction. See also: «pursuit», p. 10 c. Synchronous end 8. Repulsion. See also: «evasion», p. 10 C. Cross-domain interaction 1. Scaling (time) 4. SYMBOLS 2. Synchronous start 4.1 Early SSMN Spatial Taxonomy and Symbolic 3. Delay representation research 4. Synchronous end Initial decisions about symbol design concerned the ap- proach to symbolic representation. As the taxonomy was 3.3 Further Taxonomy directions being developed a provisional set of symbols was defined Since the primary intention of the SSMN project is to based on ongoing comparative studies of 2-D and 3-D provide a working prototype of a software package that graphic representation of spatial motion. Additional per- can be tested by composers, each aspect of the taxonomy tinent authors were Trevor Wishart (1996) [24], Bijan that has been addressed here undergoes verification by Zelli (2001) [25], Larry Austin (2004) [26], Lasse Thore- users. As indicated earlier, an open source score editor sen [27], Bertrand Merlier (2008) [28] and Vincent Ver- (MuseScore) has been targeted for graphic symbols im- faille (2003) [29]. An overall design concept was adopted plementation allowing real-time OSC messages to be with the primary criteria requiring clarity, legibility and transmitted to a rendering engine. The sound projection rapid recognition through reliance on simple visual sym- tool used for these experiments is an ambisonics spatiali- bols such as cube, sphere, radar, perspective, arrows, zation system allowing the simulation of different multi- colors, size, etc. (see figure 1). channel projections in various formats as well as a binau- ral headphone version. The score editor is dubbed MuseScoreSSMN and sends all OSC spatialization in- formation via a dedicated port to Max-based tools (e.g. the SSMN-Rendering-Engine) [19]. While the prototype is being prepared, tested and doc- umented, further aspects that could be symbolized are being oriented towards questions of behavioral interac- tions between two or more sound sources affecting their spatial movement. A research project at the University of Zurich concerning data mining and visual analysis of movement patterns proposes a taxonomy of movement Figure 1. Example of early symbol design research. patterns [18] that can be investigated using sound sources and can be integrated into the spatial taxonomy. The This basic set was improved in subsequent design following list of behavioral attributes and relationships loops. The following major issues came up during the make reference to this work (page numbers) and are pre- process: sented here as a suggestion for further research: • Defining "symbolic" as opposed to "descriptive" (i.e. A. Behavioral attributes icon versus image) 1. Trend-setter: a sound source establishing move- • Creating a grammar such as the creation of compound ment patterns followed by other sources, p. 10 symbols (e.g. a circle with vibrato-type movement in- 2. Follower, p. 10 cluding acceleration) or determining a set of regrouped 3. Indifferent: autonomous (non-uniform) or random staves/tracks for which a common action is proposed) movement within a behavioral context. See also: • Determining parameters of SSMN symbols «dispersion»: “non-uniform or random motion, • Establishing a timeline with key frames (e.g. a dedicat- opposite to concurrence”, p. 8 ed staff) allowing continuous activity of a symbol being B. Behavioral relationships reported on the timeline representation 1. Imitation: see also «concurrence»: “same values of • Pertinence of the use of a quadrant or grid to improve motion attributes at a certain instant or duration”, legibility (see figure 1, upper row, symbols 8-12) e.g. «synchrony», p. 7 • Creating tools for manuscript input to allow a degree of 2. Coincidence: similar positions, full or lagged, p. 8 freedom for composers to deal with situations where the 3. Opposition: bi- or multi-polar arrangement, e.g. taxonomy would not provide the adequate tool for a spatial splitting, p. 8 specific idea (e.g. the utilization of a rubber-stamp for 4. Constancy: “movement patterns remain the same rapid manual input of composer's trajectory designs). (...) for a particular duration”, p. 8 Several strategies of graphical possibilities had to be 5. Convergence: synchronous or delayed, “move- tested in view of integrating these symbols into the open ment to the same location”. See «encounter», p.9 source score editor MuseScore. 4.2 SSMN Symbol set Sinus Triangle This process led to the actual symbol set consisting of the following categories: Square Saw • Physical performance space characteristics • Initial physical placements of performers, microphones, Random Algo
loudspeakers and objects • Position of sound sources (RS, PA) Table 1. Symbols designed according to the taxonomy • Trajectories / displacement of sound sources, micro- phones, loudspeakers and objects The symbols shown in Table 2 complement those refer- • Operations ring directly to elements of the taxonomy. They specify • Stop/End markers delimiting the time domain of sym- further positions and movements of sources or address bols (see “Working with symbols” below) new elements and functionalities. The first two rows • Inter-application communication resources (OSC, include additional types of movements of performers. MIDI) for interaction with external programming envi- The next two rows introduce stop markers for trajectories ronments and modifiers as well as special markers for defining The last two are not explicitly contained in the taxono- pauses within a trajectory without sound interruption. The my. They came up as a result of operational needs during next row presents symbols defining alternate movements the implementation stage. of points and groups. The following two rows specify 3D Table 1 includes only symbols created according to the positions of points and groups, the next one the position taxonomy. Some of them are already available within of planes. The symbols in the last row allow for the defi- MuseScoreSSMN. Trajectories appear in two variations: nition of inter-application communication and a dedicated single direction and back and forth. SSMN staff respectively.
Hemi- Perf_rotate Perf_free Cube sphere
Church Other Perf_to&from Perf_other
Performer Perf_line Trajectory Modifier_end _end Music Perf_arc stand Pause_start Pause_end Micro- Alternate_ Alternate_ Audience phone point group Loud- Swinging 3D_point (RS) 3D_point (PA) speaker microph. Swinging Choreo- 3D_group 3D_group loudsp. graphy (RS) (PA) Sound Sound Root_plane Root_plane point RS point PA back front Group Group RS PA Communica- SSMN Staff tion OSC Plane Scale
Table 2. Additional symbols Straight Polyline
4.3 Working with symbols Poly_ Circle closed Figure 2 below illustrates the basic workflow within MuseScoreSSMN: (A) selection of a symbol from the Slinky Spiral “SSMN Palette”; (B) placement in the score; (C) defini- Bézier_ tion of parameters in the “Inspector window” correspond- Bézier spline ing to the symbol chosen; (D, E) display of the trajectory or trajectories designed by the user in the interactive Béziergon Bernoulli “Radar window”. This window contains a top and a side view (E). Each circle corresponds to 10 spatial units to be scaled according to the real space. question of symbol activity in the context of digital repre- sentation possibilities. On one hand, having a score in the digital domain allows for much greater latitude in provid- ing continuous information through windowing, with or without animation. On the other hand it might be neces- sary to reduce the displayed information in the printed version of full scores and parts for reasons of clarity. New possibilities appear when imagining interaction through integration of various software applications dedi- cated to facilitating artistic processes. A collaboration between the research teams of “inScore” and “Faust” at GRAME (Lyon) and SSMN has recently been undertaken with the expectation of creating tools to facilitate interac- tion on a local level and in web applications for visual display and audio rendering purposes. Other aspects be- ing currently investigated are SpatDIF compatibility and the integration of SSMN Elements within the MusicXML protocol.
Figure 2. Workflow within MuseScoreSSMN. 5. CONCLUSIONS Results of the SSMN project have already been tested In this example the phrase played by the tenor saxo- with composition students at the Zurich University of the phone will be picked up by a microphone (projected Arts and presented at the Haute École de Musique of audio) and spatialized according to the “Bernoulli” sym- Geneva. This experience has revealed encouraging devel- bol (see Table 1) placed above the staff at the beginning opments, such as increased awareness of spatialization of bar 1. The corresponding “Trajectory_end” symbol possibilities within the composition process and aug- (see Table 2) at the end of the first bar marks the exact mented spatial listening acuity. The main intention of the point in the timeline at which the trajectory ends, thus project is to reflect on the ways we think of and work defining its effective duration. The initial position of the with spatiality in composition and to envision procedures performer (root sound) is defined by the “Performer” that integrate spatiality from the very beginning. The symbol (see Table 1). After playing bar 1 the player is software prototype is intended as a tool that facilitates the asked to move towards a new position defined by a simi- exploration of such procedures. Further tests and experi- lar symbol at the end of bar 2. The trajectory used is ences should help to clarify if similar workflows can defined by the “Perf_line” symbol (see Table 2) at the become practical and open enough to meet the necessities beginning of bar 2. of different composers. Additionally, two dedicated SSMN staffs have been set The taxonomy presented here reflects approaches to to define the spatialization of pre-produced samples. The spatialization based mainly on geometrical and visual movement of the sample named “Textur 1” is defined by concepts such as lines, curves and planes. New organiza- a “Bézier_spline” symbol (B, see also Table 1). The re- tion paradigms can be envisioned by introducing time sulting trajectory is shown in the radar window (D). It based dynamic movement patterns as observed in biolog- contains 4 control points (marked with tangents) and will ical and social contexts. The persistent idea of sound as be repeated once. The corresponding parameters includ- an object, closely related to visual and geometric con- ing start and end positions (x, y) are shown in the inspec- cepts, could be challenged by an understanding of sound tor window (C). The sample named “Textur 2” begins at as a continuously changing field of energy, as the result the fourth beat of bar 1. It was defined as a polyline. Both of interacting information streams. Although the emer- samples have “Trajectory_end” symbols above the corre- gence of new notation paradigms will be supported by an sponding staff. An SSMN staff can be used independently evolving technology that already makes possible the of musical events and become a timeline for other kinds integration of interactive interfaces in performance prac- of information (e.g. choreography notation, film editing). tice, it can be assumed that conceptual thinking in com- The green line indicates a selection including the saxo- position will remain the major source of aesthetic innova- phone staff and the upper SSMN staff. The radar window tion of spatialization in electroacoustic music. (E) shows the superposition of both trajectories. Here the Acknowledgments ends of the trajectories are highlighted with a point. The position of the performer is not displayed in the Radar The SSMN research team is grateful for the assistance window since it is not relevant for rendering. and support offered by the Swiss National Foundation for Scientific Research, the members of Institute for Com- 4.4 Current developments puter Music and Sound Technology at the Zurich Univer- A basic operative feature to be implemented in the near sity of the Arts, the Computer Music division of the future is the possibility of saving movement patterns Haute École de Musique of Geneva, and the GRAME, defined by descriptors and modifiers. Another issue is the Lyon. 6. REFERENCES [16] A. Cont, J. L. Giavitto and F. Jacquemard, “From Authored to Produced Time in Computer-Musician [1] K. Patton, “Morphological notation for interactive Interactions”, CHI 2013 Workshop on Avec le electroacoustic music”, Organised Sound 12(2), pp. Temps! Time, Tempo, and Turns in Human- 123-128, 2007. Computer Interaction, Apr 2013, Paris, France. [2] D. Besson, “La transcription des musiques ACM, 2013.
the Unconventional, LIT Verlag, Münster, 2013. Mathematics and Social Sciences- 45e année, n° 178, pp. 51-55, 2007. [10] M. Trochimczyk, Space and Spatialization in Con- temporary Music: History and Analysis, Ideas and [24] T. Wishart, On Sonic Art, ch. 10. Routledge, Taylor Implementations, Moonrisepress, California. 2011. & Francis Group Ltd. Oxford, 1996. [11] J. Kapralos, Auditory Perception and Spatial (3D) [25] B. Zelli, Reale und virtuelle Räume in der Com- Auditory Systems, York University Press, Canada, putermusik: Theorien, Systeme, Analysen, Tech- 2003. niche Universität Berlin, 2001. [12] J. J. Nattiez, Music and Discourse: Toward a Semi- [26] L. Austin, John Cage’s Williams Mix (1951-3): The ology of Music, Princeton University Press, New Restoration and New Realizations of and Variations Jersey, 1980. on the First Octophonic, Surround-Sound Tape Composition. Edited by Hall and Sallis. Cambridge [13] G. Theile, “Wave Field Synthesis- A Promising University Press, 2004. Spatial Audio Rendering Concept”, in Proceedings of 7th ICDAE (DAF’04), Naples, 2004. [27] L. Thoresen, and A. Hedman, “Spectro- morphological Analysis of Sound Objects: an adap- [14] J. Daniel, Représentation de champs acoustiques, tation of Pierre Schaeffer's typomorphology”, Or- application à la transmission et à la reproduction de ganised Sound 12 (2), pp. 129-141, Cambridge Uni- scènes sonores complexes dans un contexte versity Press, 2007. multimédia. Thèse de doctorat de l’Université, Paris 6., 2000. [28] B. Merlier, Vocabulaire de l’espace en musiques électroacoustiques, Editions Delatour, France, 2008. [15] C. Agon, G. Assayag and J. Bresson (Eds), The OM Composer's Book Vol. 1. Collection Mu- [29] V. Verfaille, “Effets audionumériques adaptatifs: sique/Sciences, Editions Delatour France / IRCAM, théorie, mise en oeuvre et usage en création 2006. musicale numérique”, Doctoral Thesis, l’Université Aix-Marseille II, France, 2003.