kojs, serafin & chafe agus

8/2/2019 Kojs, Serafin & Chafe Agus

1/6

Aphysical model is a computer simulation ofa sonic object (whether it be a musical instrument, an envi-ronmental phenomenon or an everyday object) based onunderstanding and implementation of the sound produc-tion mechanism. Physical modeling synthesis is an excellentvehicle for conceptualizing and incorporating the reality ofphysical sounding objects such as musical instruments intocomputer-mediated artistic production [1,2]. What differen-tiates this technique from other syntheses is the fact that phys-ical modeling simulates the mechanics of sound production,while other techniques (e.g. additive, subtractive and FM syn-theses) focus on modeling the acoustical properties of the sig-nal as heard and recognized by the listener [3]. Derivation ofthe models from physics ensures that their perceptual identi-ties and behavior are retained under a variety of conditions.

The purpose of this article is to examine how cyberinstru-ments of the physical modeling type were used in music com-positions [4]. In each section we focus on a particular kind ofcyberinstrument. We recognize three categories of cyberin-struments: extended, hybrid and abstract cyberinstruments.The extended cyberinstruments are simulations of existingphysical instruments. Besides pure replication, they enableaugmentation of the instruments parameters beyond the lim-

itations of their physical origins. Hybrid cyberinstruments aretypically combinations of the properties of two or more exist-ing instruments, such as the blotar[5] and ublotar[6], whichcombine the properties of flute and guitar. Abstract cyberin-struments are structures inspired by physical laws, yet withoutequivalence in the physical world. Table 1 displays the basictopology of cyberinstruments created by means of physicalmodeling.

Certain cyberinstrument types are often associated witha particular method of physical modeling. While extendedcyberinstruments are frequently modeled by means ofdigi-tal waveguide synthesis, simulation ofhybridcyberinstruments

is efficiently realized using modalsynthesis. Abstractcyberinstrumentdesigns are facilitated using a mass-spring-damperalgorithmic approach.There are also a number of excep-tions, which suggest that all classesof cyberinstruments can be realizedby techniques other than those as-signed in our categorization.

The following sections treat eachkind of cyberinstrument in turn, de-scribing modeling approaches andpresenting compositional examples. We also provide a sam-pling of compositions written with cyberinstruments. Theseare compositions known and accessible to the authorsthecatalog is not to be considered comprehensive.

COMPOSING WITH EXTENDEDCYBERINSTRUMENTSExtended cyberinstruments enable parametrical expansion ofexisting music instruments beyond the limitations of the phys-ical world. While manipulating extended cyberinstruments,

the composers have generated novel timbres and conceivedan augmented sonic reality.

OriginsMost probably, the first cyberinstrument was the physicalmodel of a vocal tract proposed by J. Kelly and C. Lochbaum[7]. Max Mathews used this cyberinstrument in the composi-tion Bicycle Built for Two(1960). To create a futuristic effect andcelebrate emerging modeling possibilities, Stanley Kubrick in-cluded the song, performed by the dying computer HAL, atthe end of his film 2001: A Space Odyssey(1968).

Hiller and Ruiz AlgorithmThe behavior of vibrating objects such as strings can be mod-eled as a mass-spring system archetype. This behavior is ex-pressed mathematically through difference equations, whichin turn describe the behavior of digital filters needed for thesynthesis. Solving the difference equation for strings pluckedand struck at different places allowed Hiller and Ruiz to simu-late the vibrating string with a series of masses and springs [8].

In Corda di Metallo(The Metal String, 1997), for string quar-tet and electronics, Michelangelo Lupone composed with amodel of the bowed string designed by Marco Palumbi andLorenzo Seno. The model is based on Hiller and Ruizs algo-

2007 ISAST LEONARDO MUSIC JOURNAL, Vol. 17, pp. 6166, 2007 61

Cyberinstruments viaPhysical Modeling Synthesis:Compositional Applications

Juraj Kojs, Stefania Serafinand Chris Chafe

Juraj Kojs (composer, researcher, performer), McIntire Department of Music,211 Old Cabell Hall, University of Virginia, Charlottesville, VA 22903, U.S.A.

E-mail: . Web site: .

Stefania Serafin (researcher, educator), Medialogy, Aalborg UniversityCopenhagen, Lautrupvang 15, 2750 Ballerup, Denmark. E-mail: .Web site: .

Chris Chafe (composer, researcher, educator), Center for Computer Research in Musicand Acoustics (CCRMA), Department of Music , Stanford University, Stanford, CA 94305,U.S.A. E-mail: . Web site: .

An on-line database of compositions w ritten with cyberinstruments by physicalmodeling synthesis along with other information concerning the topic can be accessedat .

A B S T R A C T

This paper details composi-tional approaches in music for

cyberinstruments by means of

physical modeling synthesis.

Although the focus is on compo-

sitions written with the models

simulated by the digital wave-

guides, modal synthesis and

mass-spring-damper algorithms,

music written with other model-ing techniques is also reviewed.

LEOMJ17_pp061-066.ps - 10/29/2007 9:00 AM


2/6

rithm. Lupone controlled the followingbowed string parameters: bow position,velocity, force, string dampening, lengthand material density [9]. Corda di Metallopresents a communication between thephysical and cyber strings. Such oscilla-tion is reflected in the formal design ofthe piece. Sections performed by cyberstrings follow sections performed by thestring quartet. Either the cyber string re-

tains the parametrical behavior of thephysical string instrument or its registersand gesture envelopes stretch to unrealdimensions. Lupone blended the bowedstring model sounds with the vocal sam-ples in his later composition Canto diMadre(The Mothers Song, 1998).

The McIntyre, Schumacherand Woodhouse AlgorithmSelf-sustained oscillators can be alsomodeled with the McIntyre, Schumacherand Woodhouse algorithm. This method-ology centers on detailed examination oftime-domain behavior of sounds. Cou-pling nonlinear exciters with linear res-onators facilitates simulation of suchsystems as woodwinds, bowed strings andpipes [10].

Chris Chafe used this technique tomodel the bowed cello [11]. Chafedesigned a cello-like synthesizer thatcombines a number of bowed string in-struments. The synthesizer facilitatescontrol over five performance parame-ters: string length, bow velocity, force,contact position (affecting loudness andtone quality) and string dampening.

In Transect(1999), Chafe used the cy-ber cello to extend the sonic capabilitiesof the physical instrument. Further ex-

panding the cello identity, Chafe fusedthe extended cyberinstrument with amodel of the vocal tract. Emergentsound, the result of this cross-synthesisprocess, carried the sonic, although notphysical, features of both parents [12].

Karplus-Strong AlgorithmIn 1983, Karplus and Strong proposed aneffective algorithm to model plucked

dolin and deep bass strings the size of theGolden Gate Bridge. Augmentations ofthe strings identity are very subtle. Jaffesmoothly transitioned between the soundof physical strings and extended sonori-ties marked with impossible detuning,performance velocity, articulation, dis-tortion effects and length of resonant de-cay, which spanned from a fraction of asecond to approximately half a minute.

Timbrally coherent, Silicon Valley Break-downmay entice the listener to believethat there in fact physically exists such amulti-string instrument.

Tempo perturbation added a com-plexity to the performance, such that itwould be impossible for real perform-ers to reproduce the piece. This im-possibility technique implies a charac-teristically digital augmentation, clearlyparalleled with Colon Nancarrows acous-tic works.

Digital WaveguidesJulius Orion Smith III proposed model-ing with digital waveguides as a novel ap-proach to physical modeling [1618].Digital waveguide synthesis focuses onmodeling the medium in which the wavespropagate. A pair of digital delay linessimulates sound waves traversing the res-onating medium in opposite directions.Interaction of the traveling waves causesresonances and interferences related tothe dimensions of the medium [19].Strings and tubes may be considered suchmedia. If the previous implementationsconsidered the ideal string, in which nolosses occur, digital waveguides take intoaccount these losses by means of low-pass filters in the system. Digital wave-

guides offer a finer level of accuracy inmodeling certain vibrating objects. Fur-ther, compactness of this synthesis facili-tates efficient real-time implementationof the models. In general, systems withquasi-harmonic spectra, such as vibratingstrings and air columns, are suitable forefficient modeling by one-dimensionalwaveguide techniques, while inharmoni-cally behaving sonic objects are more ef-ficiently modeled with 2D waveguidemeshes, modal synthesis techniques ora combination of 1D waveguide andmodal techniques called banded wave-guides.

One-dimensional digital waveguideshave been used to model vocal tracts [20],bowed strings [21], woodwind instru-ments [22,23], piano [24], singing cor-rugated tube [25] and other instruments.Objects with a few inharmonic modeswere modeled using banded waveguides.Examples of such models include res-onating percussion bars [26], musical

string and drum synthesis [13]. First, awavetable is fil led w ith random values.Subsequently, the values are read andsent out to a modifier (e.g. a low-pass fil-ter). The algorithm contains a loop thatis completed when the data is fed back tothe system and re-read after a certain de-lay. This process is continuously repeatedat audio speeds. The resulting sound re-sembles the timbre of plucked strings.

Jaffe and Smith expanded the Karplus-Strong algorithm by adding all-pass fil-ters to the loop and improving the filtersproposed by Karplus and Strong. Theirrefinements included improvement oftuning, better control for tone decay timeand loudness, spectral shaping of the ini-tial plucked sound, variation of toneloudness in relation to its bandwidth,variation of the character and number ofattacks, glissando and slur, simulation ofthe sympathetic string vibrations, simu-lation of a stiff str ing and simulation of amoving pick [14].

Jaffe used the extended Karplus-Strongalgorithm of plucked strings in his com-positions May All Your Children Be Acrobats(1981), Silicon Valley Breakdown(1982),Telegram to the President (1984), Grass(1987) and Racing against Time(2001), aswell as in other pieces. The composer fo-cused on extending the timbral possibil-ities of the string model while preserv-ing its sonic identity. Additionally, Jaffeused the cyberinstrument to simulateperformance modes such as extremelyrapid tempi and register changes impos-sible to accomplish on the physical in-struments.

Jaffe structured May All Your ChildrenBe Acrobats, for computer-generated tape,

eight guitars and voice, as a dialogue be-tween physical and cyber guitar strings.Although continuously stretching thetimbral, pitch and rhythmic possibilitiesof the guitar performance, Jaffe strove toretain the field of recognizable sounds[15].

Silicon Valley Breakdown, for tape, isscored for a symphony of plucked stringcyberinstruments, including piccolo, man-

62 Kojs, Serafin and Chafe, Cyberinstruments

Table 1. Basic topology of cyberinstruments created by means of physical modeling.

Instrument type Modeled object Novel timbre results from

Extended Existing instrument or Extending the instrumentssonic object parameters

Hybrid Two or more existing Blending multipleinstruments or sonic objects instruments together

Abstract Abstract instruments or sonic Combining physical agentsobjects described by terms such of sound productionas masses, springs and dampers such as masses, springs

and dampers

LEOMJ17_pp061-066.ps - 10/29/2007 9:00 AM


3/6

saw [27], Tibetan bowl [28] and glassharmonica [29]. Modeling highly inhar-monic structures is not usual in wave-guide methodology, yet Serafin, Huangand Smith proposed a banded waveguidemesh to model bowed cymbals [30].

S-Trance-S(2001), S-Morphe-S(2002)and That Which Is Bodiless Is Reflected inBodies(2004) are compositions by Mat-thew Burtner that use waveguide models

designed by Stefania Serafin [31]. In S-Trance-S, for metasaxophone and elec-tronics, the identity of the virtual bowedstring is distorted and mixed with thesonorities of an acoustic saxophone [32].At moments, the string timbre com-pletely dissolves into an abstract electricmist.

The cyber Tibetan bowl is tapped andblown by soprano saxophone in S-Morphe-S. The composer widened theinstruments sonic identity while reshap-ing its natural gesture envelope. Thecomposition begins with sustained bowlsonorities and proceeds to a saxophoneand bowl duet, in which the bowl aug-ments the sonic space as the resonatingcavity for the saxophone signal. Blendingof the physical and cyber sonorit ies is fol-lowed by the instruments registral sepa-ration, in which the saxophone partfloats above the resonating bowl.

In That Which Is Bodiless Is Reflected inBodies, the composer focused on the ex-ploration of beatingthe characteristicTibetan bowl sonic quality. The extendedcyberinstrument enabled the composerto generate beating in a multitude ofspectral variations and rhythms. Thecomposition investigates the notions ofdistortion, polymetrical pulse and tex-

tural transformation in the cyber bowl.The composer augments the sonic spacethrough subtle detuning, pitch bendingand registral expansion of its 8-channelspatially distributed body.

Chris Chafe explored the sonorities ofthe cyber bugle and Perry Cooks hose-player waveguide brass model [33] in animprovisation-based compositionEl Zorro(1991). With Greg Niemeyer, Chafe alsodesigned Oxygen Flute(2001), an interac-tive real-time computer music environ-ment. The work utilizes digital waveguidemodels of four 9,000-year-old Chinesebamboo flutes from the Jiahu archeo-logical sites [34]. While the flutes tim-bral qualities remain preserved, Oxygen

Fluteaugments the notion of an instru-ment and performance. The visitorsbreathing directly excites the extendedcyberinstruments. As the v iewers enterthe space, the carbon dioxide levels inthe greenhouse increase. The sensors de-tect the level changes and send the in-

thermore, a number of digital waveguideSynthesis ToolKit (STK) implementa-tions can be heard on Lanskys Music BoxCD (2006). In Fuses clarinet, mandolinand saxophone models. Other pieces onthe CD, such as Composition Project for Se-niorsand A Guy Walks into a Modal Bar, en-gaged the modal bar designed by modalsynthesis.

Lansky was primarily involved in

stretching the cyberinstruments param-eters to unrealistic dimensions and, bydoing so, producing vast-sounding cyberlandscapes in Still Time. Lansky invites thelistener to actively inspect the cinematic,flowing soundscape and reflect on it.

The listener may observe intertwiningstreams of natural and synthetic identi-ties as they are presented through un-cluttered structures arriving from avariety of proximities.

SonoMorphis(1998) is an interactive 3Dinstallation designed by Bernd Linter-mann (graphics) and Torsten Belschner(sound). The sonic part utilizes wave-guide models of pipe and string (with 16xYamaha VL70-m). Both graphics and mu-sic mutate while interacting with the sys-tem. The visuals are based on extractednatural patterns. Both visuals and soundsare transformed to unrealistic ones as theuser interacts with the system [46]. Visualmaterials, shapes and spatial positioningof the objects are mapped to the timbreand pitch of the extended cyberinstru-ments. Additionally, the visuals dynami-cally control sound spatialization in realtime. Through this cross-modal ap-proach, Belschner achieved sensible aug-mentation of the instrumental identities.

Juan Reyes used digital waveguide

models in Straw-berri(1997, flute andplucked string models), Wadi Musa(2001, clarinet model),ppP(2001, pianomodel), Freddie the Friedlander (2004,bowed string model) and Fuxing(2006,pipa model). InppP, for piano and elec-tronics, Reyes used the virtual piano toextend the sonic possibilities of the phys-ical instrument while simulating the ef-fects such as detuning and retuning ofthe strings, generating extreme pitchfluctuations, expanding the instrumentsregisters and modifying the natural en-velope of the struck-string gestures. Com-plementing relationships between thephysical and extended cyberinstrumentsfuels a sinuous augmentation of the pi-ano sonic space.

Physically InformedStochastic ModelsPerry Cook developed the PhysicallyInformed Stochastic Event Modeling(PhISEM) technique, which is based on

formation to enliven the cyberinstru-ments.

Ping, also created by Chafe and Nie-meyer, is a network environment, whichinvolves a series of the cyber-pluckedstrings parametrically expanding andcontracting. Like Oxygen Flute, Ping(2001)expands the notion of the performancespace. The installations Internet con-nections are constantly reforming, with

multiple physical destinations, and thecommunication between the locationsexcites the cyber strings. The mapping ofthese Internet contacts to the cyberin-struments defines the composers ex-tended sonic space.

Ted Coffey used physically modeledglass harmonicas [35] in Armonica Lul-labies (2004), for stereo tape, Koans,for video and sound (2004) and otherworks. In Armonica Lullabies, the com-poser reached into the virtual sonic spacewhile engaging the extended cyberin-strument in delicate augmentation of thesampled harmonicas timbral properties.

Achim Bornhoeft employed a wave-guide model of a plucked string in Vir-tual String(1997), for tape. The cyberstring was implemented in the graphicaluser interface vstring, which allowed thecomposer to manipulate string tension,stiffness, dampening, excitation type andposition, virtual pickup position and fre-quency response. Sounds of simulatedstrings with unnatural physical measure-ments and behaviors expand Bornhoeftssonic space, yet the space remains in-formed about the sonorities of the orig-inal physical instrument [36].

Juraj Kojs has composed a number ofworks with digital waveguides, such as

Garden of the Dragon(2003), Three Move-ments (2004), Air (2006), Concealed(2006),En Una Noche Oscura(2006), InSecret (2006) and To Where He Waited(2006). Musical instruments, every-day objects and musical toys suggestedactions to which various extended cyber-instruments responded. The cyberin-struments expanded the musical spacewhile timbrally enhancing their physi-cal counterparts, participating in a cre-ation of hybrid analog-digital instru-ments and providing a resonant space forthe performance given on the physicalinstrument. The compositions engagedextended cyberinstruments, such as thesinging tube [37], bowed string [38], fu-jara [39], Tibetan bowl [40], bowed bar[41], flute [42] and 2D mesh [43].

Paul Lansky utilized a waveguidemodel of the slide flute [44] in Still Time(19931994) and Sullivans physicalmodel of a plucked electric guitar string[45] in Things She Carried(1997). Fur-

Kojs, Serafin and Chafe, Cyberinstruments 63

LEOMJ17_pp061-066.ps - 10/29/2007 9:00 AM


4/6

pseudorandom organization of smallsound particles [47]. The algorithm isbased on Newtonian equations that ex-plain motion and collision of pointmasses. Statistical principles of particlecollision in a shell are applied to shakerssuch as maraca, sekere and cabasa. Per-cussive instruments with larger numbersor resonances, such as tambourine andsleigh bells, may be also modeled using

the PhISEM algorithm. Individual par-tials are modeled with digital filters,whose resonant frequencies are replacedby another frequency located in close yetrandom proximity to the principal reso-nance every time a collision occurs.

Juan Reyes scored Wadi Musa(2001)for quenas(Andean flutes), cello, clarinetwaveguide model, and s tochast ic mod-els of maracas. Reyes extended CooksPhISEM model of maraca [48] by cus-tomization of gourd-resonant filter co-efficients, shaking rate and energieswith the Common Lisp Music (CLM)program [49]. The program reads andexecutes the score with pre-specifiedparameters. Derived from the model ofa maraca, a cabasa, guiro, tambourineand wind chimes provide continuouslychanging rhythmic patterning in thecomposition. Reyes was primarily con-cerned with augmenting the sonic prop-erties of the physical instruments.

Dan Trueman utilized Cooks PhISEMmodels in Lobster Quadrille(1999). Thecomposer controlled the models with theBow-Sensor-Speaker-Array (BoSSA) sys-tem [50]. Cyber bamboo wind chimeswere used in Wind in Hands, Water in Feetfor dancers and electronics, in whichthe instrumental timbres are extended

to simulate the water-like sonorities. Thedancers, equipped with sensors (accel-erometers) on their feet, then activatethe instrument as if interacting with thecyber water. The composer augmentedthe instrumental space of the bamboowhile transforming it into the sonoritiesof the flowing water [51].

Kojs Revelations (2005), for circulartoys, resonant plates and electronics,primarily explored the sonorities of phys-ical toys and cyber percussion instru-ments. Plastic superballs, glass marblesand metal Bocci balls were used to con-trol cyber maracas, guiro and bamboowind chimes. Bouncing, rolling and scrap-ing the circular toys against the resonantplates excited unrealistically shaped cy-ber shakers. To complement the scrapingof a physical rubber ball against hard sur-faces, MAX/MSP implementations of thebowed percussion bar physical model de-veloped by Georg Essl and Perry Cook[52] and a friction-bowed string model

striking, bowing and plucking may con-nect the elements. In some design situa-tions, however, such as in reed and bowconnection, it is difficult to decide on ef-ficient control values. Similarly, debug-ging and control of spectral features ofactual sound remain problematic [58].

Hans Tutschku used hybrid cyberin-struments by means of modal synthesisinEikasia(1999).Eikasia, for 8-channel

tape, was written with MODALYS, a newgeneration application of the previouslydescribed MOSAIC, which was devel-oped at IRCAM. In the composition,Tutschku explored the resonances of rec-tangular and circular plates mixed withthe spectra of low piano tones. Most in-terestingly, the composer created hybridcyberinstruments while dynamically fus-ing multiple plates of different spectralcharacteristics. Parametrical combina-tion and oscillation between the multi-ple parental models resulted in creationof cyber plate-like hybrids transpiringinto the intermediate states [59].

Designing hybrid cyberinstruments bymeans of digital waveguide synthesis,which combines excitat ion of one in-strument with the resonator of another,is an uncommon phenomenon. DanTrueman and Gary Scavone involved hy-brid cyberinstruments in their composi-tions. Trueman used the blotar, a hybridcyberinstrument combining propertiesof flute and Charles Sullivans electricguitar models, in the improvisationalcontexts [60]. Novel timbres emerge withthe parametrical oscillation between theflute and guitar identities, dependingupon the prevalence of electric guitar orflute parameters in the synthesis. Scavone

proposed and compositionally imple-mented two cyber blown strings in hisAir Study I(2002).

ABSTRACTCYBERINSTRUMENTSMass-spring-damper algorithms excel inthe simulation of the abstract cyberin-struments. CORDIS-ANIMA is an audio-visual environment [61] that epitomizesimplementation of modeling by com-bining masses, springs and dampers.These agents can be connected in a lin-ear network. A conditional link mayintroduce nonlinear behavior to the sys-tem. While CORDIS simulates the soundsof music instruments, sonic objects andnatural phenomena such as moving sanddunes [62], ANIMA allows modeling ofthe visual component. GENESIS is a com-positional environment in which theCORDIS principles were recently imple-mented [63].

designed by Stefania Serafin [53] wereemployed in Revelations. The models con-tributed to the timbral and temporalaugmentation of the quickly decayingscraping gestures of the physical plates(particularly the plastic ones). Mixedanalog-digital resonating structures re-sulted from the combination of physicalscraping excitation and reverberation ofextended cyberinstruments.

HYBRID INSTRUMENTSAs mentioned in the introduct ion, al-though digital waveguides and mass-spring-damper algorithms can be used tosimulate hybrid cyberinstruments, modalsynthesis is predisposed for such a design.Conversely, modal synthesis was appliedto extended cyberinstruments (squeak-ing door and singing wine glasses [54]and some percussive cyberinstruments[55]). In his compositionErba near checresci segno nero tu vivi (1999), MauroLanza utilized modal synthesis forms im-plemented in the MODALYS software.Lanza designed a set of percussive ex-tended cyberinstruments and combinedthem with the sound of a soprano voice.

The hybrid cyberinstruments presentan amalgamation of multiple cyberin-struments within one unit. The cyber hy-brids inherit the timbral characteristicsof both parents, the dominance of whichdepends on the parametrical alignment.Intrinsically, the hybrids initiate noveltimbres existing solely in cyberspace.

Modal synthesis is based on the propo-sition that any sounding object can bedeconstructed into a set of vibrating sub-structures, such as bridge and body in the

case of the violin [56]. After excitation,each substructure produces well-definedmodes of vibration. Each mode is repre-sented by its modal data, consisting offrequency, dampening coefficients andshape variables. Such modal data mayrepresent structural elements such as vi-olin bridge, body, string, acoustic bell ortimpani membrane. The resulting simu-lation sums up the elements of all modesinvolved in the synthesis. As opposed toa mass-spring approach, modal synthesisallows flexibility in reorganizing the sub-structures of the instrument in order tomodify its physical and thus sonic char-acteristics. MOSAIC is a virtual work-bench designed by Jean-Marie Adrienand Joseph Morrison, which allows theuser to assemble modal substructure ob-jects into musical instruments [57]. Thus,the model is constituted as a collectionof mechanical and acoustic resonantstructures that vibrate and interact undervarious excitation conditions. Adhering,


LEOMJ17_pp061-066.ps - 10/29/2007 9:00 AM


5/6

Tutschku used GENESIS in hisResorption-Coupure(2000). In Resorption-Coupure, a bow-like object, friction objectsand hammers excite abstractly definedresonating structures. Tutschku con-structed the piece by conscious alterna-tion between the perceptual recognitionand nonidentification of sonic objects.This process is reflected in the use of theabstract cyberinstruments (somewhat re-

sembling string-like metal objects) andphysical-world sounds (whispering andbreathing).

Claude Cadoz exemplified creativework with CORDIS inpico. .TERA(2002)for quadraphonic tape. In this composi-tion, Cadoz restricted himself exclusivelyto implementation of physical mod-els. Cadoz suggested that instruments,performers and conductor, as well asresonant performance space, can be sim-ulated by means of physical modeling[64]. Cadoz constructedpico. .TERAas anet of hierarchical relationships betweenthe involved agents. The agents interactand exchange function during the com-position. For example, a cyberinstrumentmay play a model of an instrumentalistwhen appropriate interaction is initiated.The abstract cyberinstruments are de-signed as elastically connected bridges,soundboards and other parts. These cy-berinstruments are inherently responsiveto modifications, thus facilitating myriadabstract sonic identities.

A number of composers, such asHans Peter Stubbe, Ludger Bruemmer,Giuseppe Gavazza, Periklis Douvitsasand Frederic Curien, have worked atthe Association pour la Creation et laRecherche sur le Outils dExpression

(ACROE) center to compose music withGENESIS. While some of them use theabstract cyberinstruments in concertworks (Stubbe), others have incorpo-rated the cyberinstruments in multime-dia (Bruemmer, Gavazza) and theatricalcontexts (Douvitsas). GENESIS is an ex-citing and sophisticated compositionalenvironment that will increase in popu-larity with real-time implementation anddistribution on common platform com-puters.

CONCLUSIONThe above examples demonstrate waysin which composers have utilized physi-cal models to extend the sonic possibili-ties of existing instruments (extendedcyberinstruments), create merged iden-tities (hybrid cyberinstruments) and gen-erate novel vibrating structures (abstractcyberinstruments).

Composers have increasingly favored

14. D. Jaffe and J.O. Smith III, Extensions of theKarplus-Strong Plucked String Algorithm, ComputerMusic Journal7, No. 2, 5669 (1983).

15. R. Dal Farra, An Interview with David A. Jaffe,Journal of New Music Research 22, No. 4, 335347(1993).

16.J.O. Smith III, Synthesis of Bowed Strings, Pro-ceedings of the International Computer Music Conference,San Francisco (1982).

17.J.O. Smith III, Techniques for Digital Filter De-sign and System Identification with Application to

the Violin, Ph.D. thesis, Stanford University, 1983.18.J.O. Smith III, Physical Modeling Synthesis Up-date, Computer Music Journal20, No. 2, 4456 (1996).

19. C. Roads, A Computer Music Tutorial(Cambridge,MA: MIT Press, 1996).

20. P. Cook, Identification of Control Parameters inan Articulatory Vocal Tract Model, with Applicationsto the Synthesis of Singing, Ph.D. thesis, StanfordUniv., 1991.

21. S. Serafin, The Sound of Friction: Real-TimeModels, Playability and Musical Applications,Ph.D.thesis, Stanford Univ., 2004; Chafe [11]; Smith [17].

22. P. Cook, A Meta-Wind-Instrument PhysicalModel, and Meta-Controller for Real-Time Perfor-mance Control, Proceedings of the International Com-puter Music Conference, San Francisco (1992).

23. G. Scavone, An Acoustic Analysis of Single-Reed

Woodwind Instruments with an Emphasis on Designand Performance Issues and Digital Waveguide Mod-eling Techniques, Ph.D. thesis, Stanford Univ., 1997.

24.J. Bensa, Analysis and Synthesis of Piano SoundsUsing Physical and Signal Models, Ph.D. thesis, Uni-versit de la Mediterranee, 2003.

25. S. Serafin and J. Kojs, Computer Models andCompositional Applications of Plastic CorrugatedTubes, Organised Sound10, No. 1, 6773 (2005).

26. G. Essl and P. Cook, Measurements and EfficientSimulation of Bowed Bars, Journal of the AcousticalSociety of America108 (2000) pp. 379388.

27. S. Serafin et al., Analysis and Synthesis of Un-usual Friction Driven Musical Instruments, Proceed-ings of the International Computer Music Conference,Gteborg, Sweden (2002).

28. S. Serafin, C. Wilkerson and J.O. Smith III, Mod-eling Bowl Resonators Using Circular Waveguide Net-

works, Proceedings of Conference on Digital Audio Effects(DAFX-02), Hamburg, Germany (2002).

29. G. Essl, Physical Wave Propagation Modeling forReal-Time Synthesis of Natural Sound, Ph.D. thesis,Princeton Univ., 2002; Serafin et al. [27].

30. S. Serafin, P. Huang and J.O. Smith III, TheBanded Digital Waveguide Mesh, Proceedings of theWorkshop on Future Directions of Computer Music(Mosart-01), Barcelona (2001).

31. Serafin [21]; Serafin et al. [28].

32. M. Burtner and S. Serafin, The Exbow MetaSax:Compositional Applications of Bowed String Physi-cal Models Using Instrument Controller Substitu-tion,Journal of New Music Research31, No. 2, 131140(2002).

33. Cook [22].

34. C. Chafe, Oxygen Flute: A Computer Music In-

strument That Grows,Journal of Computer Music Re-search34, No. 3, 219226 (2005).

35. Essl [29].

36.A. Bornhoeft,Virtual String(1997), .

37. Serafin and Kojs [25].

38. Serafin [21].

39. J. Koj s and S. Serafin, The Fujara: A Physical

real-time implementation of cyberin-struments through physical modeling.This trend has been facilitated by emerg-ing software such as Synthesis ToolKit[65], Pure Data [66], and Max/MSP [67]with PeRColatethe free external li-brary of physical models [68] and ChucK[69], to name a few.

Waveguide synthesis seems to repre-sent the prevailing approach in the cur-

rent trends due to its compact andefficient implementation, which makesit particularly suitable for real-time use.Many composers have explored percep-tual ambiguity when extending the tim-bres of an existing instrument whileretaining its identity (extended cyberin-struments). Others have focused on dis-torting such identities by creating cyberhybrids. Abstract cyberinstruments, sug-gested by Borin in 1992, have been re-cently developed and established as sonicentities with strong, however abstract,identity attributes.

References1. G. Borin, G. De Poli and A. Sarti, Algorithms andStructures for Synthesis Using Physical Models,Computer Music Journal16, No. 4, 3042 (1992).

2. J.O. Smith III, Physical Modeling Using DigitalWaveguides,Computer Music Journal16, No. 4, 7491(1992).

3.V. Valimaki and T. Takala, Virtual Musical Instru-mentsNatural Sound Using Physical Models, Or-ganised Sound1, No. 2, 7586 (1996).

4. The present article builds on work previously out-lined in C. Chafe, Case Studies of Physical Modelsin Music Composition, paper read at InternationalCongress on Acoustics, Kyoto, Japan (2004).

5. D. Trueman and L. DuBois, PeRColate (2005),; discusses Version0.9b5.

6.V. Stiefel, D. Trueman and P. Cook, Re-coupling:the uBlotar Synthesis Instrument and the sHowlSpeaker-feedback Controller, Proceedings of the In-ternational Computer Music Conference(2004).

7.J. Kelly and C. Lochbaum, Speech Synthesis,Pro-ceedings of the Fourth International Congress on Acoustics(1962).

8. L. Hiller and P. Ruiz, Synthesizing Sounds by Solv-ing the Wave Equation for Vibrating Objects, Jour-nal of the Audio Engineering Society 19 (1971) pp.463470, 542551.

9. M. Lupone, Corda di Metallo(1997), .

10. M. McIntyre, R. Schumacher and J. Woodhouse,On the Oscillation of the Musical Instruments,Journal of the Acoustic Society of America 74, No. 5,13251345 (1983).

11. C. Chafe, Simulating Performance on a BowedInstrument, in Max Mathews, ed., Current Directionsin Computer Music(Cambridge, MA: MIT Press, 1989).

12.A linear combination of two models, this cannotbe considered as an instance of a dynamically hybridinstrument. Cello and other excitations were pipedinto a vocal tract in place of the glottis.

13. K. Karplus and A. Strong, Digital Synthesis ofPlucked String and Drum Timbres, Computer MusicJournal7, No. 2, 4355 (1983).

Kojs, Serafin and Chafe, Cyberinstruments 65

LEOMJ17_pp061-066.ps - 10/29/2007 9:00 AM


6/6

Model of the Bass Pipe Instrument in an InteractiveComposition,Proceedings of the International ComputerMusic Conference, New Orleans (2006).

40. Serafin, Huang and Smith [30]; G. Essl et al., Mu-sical Applications of Banded Waveguides, ComputerMusic Journal28, No. 1, 5163(2004).

41. Essl and Cook [26].

42. Cook [22].

43. Trueman and DuBois [5]; S. Van Duyne and J.O.Smith III, Physical Modeling with the 2-D DigitalWaveguide Mesh, Proceedings of the International Com-puter Music Conference, Waseda University, Japan(1993).

44. Cook [22].

45. C. Sullivan, Extending the Karplus-Strong Al-gorithm to Synthesize Electric Guitar Timbres withDistortion and Feedback, Computer Music Journal14,No. 3, 2637 (1990).

46. B. Lintermann and T. Belschner, SonoMorphis(1998), .

47. P. Cook, Physically Inspired Sonic Modeling(PhISM): Synthesis of Percussive Sounds, ComputerMusic Journal21, No. 3, 3849 (1997).

48. Cook [47].

49.J. Reyes, Composing for the Physical Model ofthe Maraca (2001), .

50. D. Trueman, Reinventing Violin (1999), .

51. D. Trueman, personal communication, 2005.

52. Essl and Cook [26].

53. Serafin [21].

54. F. Avanzini, S. Serafin and D. Rocchesso, Inter-active Simulation of Rigid Body Interaction With

national Computer Music Conference, Hong Kong(1996).

67. D. Zicarelli, An Extensible Real-Time Signal Pro-cessing Environment for Max, Proceedings of the In-ternational Computer Music Conference,Ann Arbor, MI(1998); M. Puckette, Max at Seventeen, ComputerMusic Journal26, No. 4, 3143 (winter 2002).

68. Trueman and DuBois [5].

69. G. Wang and P. Cook, ChucK: A Concurrent,On-the-Fly Audio Programming Language, Proceed-ings of the International Computer Music Conference, Sin-

gapore (2003).

Juraj Kojs is a senior Ph.D. student in Com-position and Computer Technologies at theUniversity of Virginias McIntire Departmentof Music. Under the guidance of JudithShatin, he investigates how cyberinstruments,through physical modeling, enable a contin-uum between physical and virtual realities inmusic.

Stefania Serafin is currently associate profes-sor in sound modeling at Aalborg Universityin Copenhagen. Before moving to Denmark,she received a Ph.D. in computer based musictheory and acoustics from Stanford Univer-sity in 2004 and a Masters in acoustics, sig-nal processing and computer science appliedto music from IRCAM in Paris.

Chris Chafe is the Duca Family Professor ofHumanities and Sciences and director ofCCRMA, Stanford University. His doctoratein music composition was completed at Stan-

ford. A current project, SoundWIRE, exploresmusical collaboration and network evalua-tion using high-speed networks (and physicalmodels) for high-quality sound (and distrib-uted synthesis).

Friction-Induced Sound Generation, IEEE Transac-tions on Speech and Audio Processing13, No. 5.2,10731081 (2005).

55. Cook [47].

56.J.M. Adrien, The Missing Link: Modal Synthe-sis, in G. De Poli, A. Picalli and C. Roads, eds., Rep-resentations of Musical Signals(Cambridge, MA: MITPress, 1991).

57.J. Morrison and J.M. Adrien, MOSAIC: A Frame-work for Modal Synthesis, Computer Music Journal17, No. 1, 4556 (spring 1993).

58. Morrison and Adrien [57].

59. H. Tutschku, Portfolio of Compositions, Ph.D.thesis, Univ. of Birmingham, 2003.

60. Trueman [51].

61. C. Cadoz, A. Luciani and J.L. Florenc, CORDIS-ANIMA: Modeling and Simulation System for Soundand Image SynthesisThe General Formalism,Computer Music Journal 17, No. 1, 1929 (spring1993).

62.A. Luciani, N. Castagne and N. Tixier, MetabolicEmergent Auditory Effects by Means of Physical Par-ticle Modeling: The Example of Musical Sand, Pro-ceedings of Conference on Digital Audio Effects, London(2003).

63. N. Castagne and C. Cadoz, GENESIS: A FriendlyMusician-Oriented Environment for MassInterac-tion Physical Modeling, Proceedings of the Interna-tional Computer Music Conference, Gteborg, Sweden(2002).

64. C. Cadoz, The Physical Model as Metaphor forMusical Creation pico..TERA, a Piece Entirely Gen-erated by Physical Model, Proceedings of the Interna-tional Computer Music Conference, Gteborg, Sweden(2002).

65. P. Cook and G. Scavone, The Sy nthesis ToolKit(STK), Proceedings of the International Computer Con-ference, Beijing (1999).

66. M. Puckette, Pure Data, Proceedings of the Inter-


LEOMJ17_pp061-066.ps - 10/29/2007 9:00 AM

kojs, serafin & chafe agus

Documents