A CAP for graphic scores Graphic noaion and performance Benny Sluchin EIC-Ircam sluchin@ircam.fr Mikhail Mal Ircam Mikhail.Mal@ircam.fr ABSTRACT Many graphic scores use he pich versus ime presenaion, as a naural exension of he usual noaion. In he general case, i displays discree piches, in a fixed imeline. Neverheless, graphic scores use a lo of coninuous lines, and he verical dimension can be adaped o a paricular performance. In such a ay, he insrumenaion is free, and he acual range of a paricular insrumen can be adaped according o he noaion. The presen aricle is iniiaed by a search o provide he performer ih adequae ools o approach he execuion of such orks. A compuer assised performance approach helps he player in he preparaion process for boh: he ime and he pich approximaions. The simulaion can enhance he performance in approaching he graphical noaion.. 1. INTRODUCTION The saring poin of he presen research as a performance issue; orks of he 20h cenury, aimed a concers and recordings, ere approached seriously. The player searched for ools ha ill help him aain he composer s inenions as proposed by he prined (graphical) scores. Curiously enough, such ools did no exis, and he precision of he performance as quie lacking. Our ork focuses on his caegory of orks and a CAP (Compuer Assised Performance) approach, ha is in fac help in preparing he orks for performance, in he same ay a meronome is. 2. PREVIOUS WORK Since several years no, orks having a generic presenaion, ha is o say, open form, has occupied us. Three examples ere discussed: Domaines (1968) by Pierre Boulez, Duel (1959) and Sraegy (1962) and Linaia Agon by Iannis Xenakis and Concer for Piano and Orchesra (1958) by John Cage ([1], [2], [3]). Oher orks by Cage, par of his lae number pieces ere Copyrigh: 2017 Firs auhor e al. This is an open-access aricle disribued under he erms of he Creaive Commons Aribuion License 3.0 Unpored, hich permis unresriced use, disribuion, and reproducion in any medium, provided he original auhor and source are credied. also accessed [4]. In all hese cases, a compuer inerface ha assis he performer in preparing and performing he orks ere developed. The o main composiions e discuss here, presen performance noaion difficulies. Conversely o previous ork [5] 1, our aim is o develop soluions, in a compuerassised performance sudy, for rombone version, and offer aural guides o aain a performance, as close as possible o he graphical noaions used. 3. TWO EXAMPLES, TWO PROBLEMS 3.1 Alvin Lucier s Panorama (1993) Scored for Trombone and piano, Panorama (1993) [6] is a ork hose shape is raced by he sliding rombone par (Figure 1). "During he course of he ork, he rombonis slides coninuously, oulining he shapes dran by he diagonal lines." [6] Figure 1: Alvin Lucier s Panorama, original manuscrip score, firs 25 seconds The musician may sop sounding his insrumen a any ime bu his slide moves coninuously. The piches are noaed ih grea precision ihin microonal unings. Alvin Lucier uses a half, quarer, hird and sixh noe noaion he also adds a shif noaion, in cycles on some piches (Figure 2) and give exac onse ime (minues and seconds). As quoed by Volker Sraebel [7], Alvin Lucier mapped he panorama of he Siss Alps o he piches of a slide rombone. He orked from a reproducion of a panorama draing by Friz Morach afer a landscape phoo by Hermann Vögeli. 1 Enrico Francioni develops a echnical soluion for one-man Ryoanji performance, in he conrabass, percussion and vocalize version.
in he flue version, J6 in he double bass version) o a elfh (J8 in he oboe version). Trombone Figure 2: Alvin Lucier s Panorama, microonal noaion 3.2 John Cage s Ryoanji (1983-84) Ryoanji (1983-4) [8] is a series of orks based on he celebraed rock gardens in Kyoo. The conours of he fifeen differen rocks are he basis of he graphic noaion. J. Cage s insrucions are ypical; hey give he poeic amosphere for he performance: Each o pages are a garden of sounds. The glissandi are o be played smoohly and as much as is possible like sound evens in naure raher han sounds in music. ( ) The score is a sill phoograph of mobile circumsances. [8] The solo insrumen pars (here are versions for oboe, flue, rombone voice and double bass) appear all in he same form. The original score presens a frame here coninuous and complex gesures ere dran. For each page, in he lef upper par, e can remark a small save ih he page regiser (Figure 3). On he abscissa, he ime (he lengh of a sysem) is represened on 24 cm. The heigh, he regiser used, is represened on 6 cm and subdivided, according o he regiser, in semiones. # # # # b # # J1 J2 J3 J4 J5 J6 J7 J8 Figure 4: Ryoanji Trombone Gardens Regisers The form of he curves is very precise, bu hard o perform on he insrumens. Boh, he ime facor (horizonal abscissa) and piches (verical) are o be approximaed ih maximal precision. 4. THE PERFORMANCE ISSUE 4.1 Pich discriminaion hreshold According o some auhors ([9], [10]) he pich inerval discriminaion for rained musicians is around a semione enh or 10 cens a 500 Hz (around B4) and increases o loer piches (100 Hz), unil a half-one. To main aspecs presen difficulies in hese pieces. The firs is he difficuly o conrol slo glissando, here he discriminaion of pich change is challenging, as in Panorama. Lucier himself says: Figure 3: Cage s Ryoanji firs rombone version page In Figure 3 he range is precise (from B0 o F#1), and he duraions are o be deermined by he performer. Concerning he duraion, only in he voice version Cage as explici (ca. 2 minues for a "garden", composed of four sysems, engraved on o pages). For he insrumenal versions, musicians generally apply his same duraion. Each "garden" is using a specific range, hich is shon in Figure 4. The verical span (hich is he same for all he pieces) indicaes a variey of inervals going from a semi one (J3 The player is no expeced o be able o hear hese finer unings, bu may perceive hem as beas per seconds agains he piano ones [6]. The second, mainly relaed o Ryoanji, is o evolve in a very lo insrumenal regiser. Ryoanji, rombone version, evolves beeen F-1 (21,83 Hz) and G#0 (51,91 Hz). In his regiser he pich discriminaion is far o enable he fine-uning of he noaed glissando curves. In addiion, he performer physical conrol, lips vibraion and slide posiion, is difficul and uncerain.
4.2 The Glissando Defined as a glide from one pich o anoher, he erm conains much more han he physical change of he frequency. There are musical examples (ehno-musical and conemporary) based exclusively on i. Is musical performance, even on very adaped insrumens (i.e. unfreed boed srings, slide rombone), is far from simple. Tradiional orchesraion books, ha discuss glissandi, miss a pracical approach hich herefore follos. A glissando has o primary feaures: he sar and ending poins, and he duraion. A secondary characer, he form of he sliding is paricularly imporan, can be defined as he speed. A glissando can be linear (having a consan speed) or no linear (having a varying speed). Performing a glissando on a rombone 2 (he insrumen relaed o his research) is srongly relaed o he regiser involved. For middle and upper regisers he execuion is deermined by he slide posiion aken as a saring poin. Naurally, one ill be unable o perform an ascending glissando if he saring slide posiion has he shores slide lengh. Alernae slide posiions for he same noe are hen needed. Anoher problem, concerning he doabiliy of a glissando, is relaed o he fac ha a rombone slide can execue a maximum inerval of an augmened fourh (he span of seven posiions a Bb rombone slide has). In many cases, a glissando has o be iled from o or more simple ones. The slide lengh modificaion needed o perform a glissando may be of some help o measure he speed of he glissando. The modificaion o perform a given inerval changes according o he rombone posiion. For example, an inerval of a half one needs a greaer slide lengh modificaion in loer posiions han in higher ones. The slide is no he only ay o conrol he pich. Paricularly in he lo regiser, he lips can operae imporan frequency changes. These changes are hough accompanied by a loss of he paricular imbre. As he lips are vibraing in a differen frequency from ha corresponding o he insrumen lengh. When performing a glissando, he saring pich is ofen jus a floaing momen. The pich has o sar moving immediaely, hough he ear on perceive a change iniially. The same holds for he ending poin of a glissando. Anoher ay o measure a glissando, lies on he noion of differenial beas. Having a reference pich, close in frequency o he moving one, ill creae his phenomenon. Conrolling he relaionship beeen o close frequencies, based on he number of differenial beas in ime, is an approximae and quie effecive mehod. 2 I seems ha some of hese poins may be generalized o sring insrumens. 5. COMPUTER SIMULATION As poined previously he performance of Alvin s Lucier Panorama and Cage s Ryoanji presens some difficulies mainly relaed o glissandi conrol, pich discriminaion and exreme lo pich conrol. The leading idea using simulaion as o produce an exac and precise guide for a performance sudy. 5.1 Alvin Lucier s Panorama (1993 Before building he compuer simulaion a firs sep as inended. Panorama as ranscribed in a Caresian space, in a break poin funcion presenaion. Each page is having he same duraion of one minue (Figure 5). Figure 5: Graphic score conaining piches This represenaion brough a beer score aareness of boh pich versus ime evoluion. The second aemp, o improve and help performance, as o ranscribe he original manuscrip (Figure 1) in a ime proporional musical noaion score. In he original here is no a clear correlaion beeen ime and graphic space. Each sysem lass for one minue (Figure 6) and he dashed measure bars represen regular 10 seconds ime span. Trombone Piano B # 1' 1'00"? # B # 2' 2'00"? # -3 B # 3' 3'00" #? B 0? 2'06" b 10" # 1'14" # -2 # 1'16" # n 2'18" n # 3'18" # n Figure 6: Alvin Lucier s Panorama, ranscribed score, firs page Even if i could seem minor, i is no. This ranscripion improves he readabiliy over he manuscrip and a good ime versus graphic space proporionaliy helps he performer o beer evaluae he linear ime span. Conrolling he slo pich change over ime is one main difficulies for Panorama. For example, a 1 43, he rombone evolves over 7 seconds ih a pich change from a C4 plus 1 Hz o a C4 minus 2 Hz. This change only can be made (as in Xenakis pieces) by referring o # 25" # # 2'29" # 1'32" +2 3'33" -2 # 2'40" # +1 1'43" n 3'49" -2 1'50" 3'51" #
he differenial beas produced beeen he rombone C4 fundamenal and he piano C3 ocave harmonic. Bu i sill presens a 0.4286 Hz/sec variaion. Anoher momen, from 14 02 o 14 14, presens a 0.1667 Hz/sec variaion on a Bb3. In general, he pich slope is very small, in relaion o he pich discriminaion inerval. In addiion o hese four crucial poins (1 43, 6 12, 9 37 and 14 02 ), he piece presens oher problemaic pich slope variaions. For Example, he smalles slope is o be found from 7 05 o 7 32, here he performer is asked o conrol a descending glissando (from a G4 o a G4 minus a sixh one), over 27 seconds. A 0.59 cens/sec variaion. These poins lead us o provide a simulaion o help he performer ih he uning and he duraion. Even if he pich changes is sill belo he pich discriminaion hearing range. The performer, playing ogeher ih he simulaion, can refer o he beaing as a ay o adjus he pich. From graphical daa inferred in he preceden seps, a hree columns able ih ime (in seconds), MIDI pich and he shif in Hz (Figure 7) as creaed. Figure 7: Firs seven rombone evens (ime in sec, MIDI pich and shif in Herz) Ploing he able daa from Figure 7 e obain a glissando curve displaying he Alps landscape (Figure 8) [7]. Insr 1 k1 oscil1 0, 1.0,p3, 2 asig oscili p4,k1,1 ou asig Figure 9: Csound orchesra for Panorama synhesis simulaion The <f2> score funcion is buil from he able daa (Figure 8). And he specral conen is se via a <f1> able in order o have a 5 parials harmonic sound (Figure 10) Figure 10: Alvin Lucier s Panorama, firs 3minues, 5 harmonic sound sonogram As he aim of he simulaion is o be a guide o a performance sudy, he use of a 5 parials harmonic sound is inended o help he performer o une and o follo more precisely he glissando. Beaing beeen higher parials are more percepible han beaing beeen loer parials. Anoher simulaion performance guide as buil using a sample of Eb4 ordinario rombone sound (from he SOL sound, Ircam daa base). This 4 seconds real sound as sreched unil o reach 16 40 using Audiosculp ( Ircam) 6. The sreched sound as hen coninuously ransposed using he daa from Figure 7, ih he SuperVP Kernel ( Ircam) 7. As he mean pich from all he performance is 63.31 MIDI pich and he median 63.24 MIDI, e choose Eb4 o minimize he ransposiions of he original sound, keeping as possible is concree characer. A Max ( Cycling74) 8 pach as buil hen, o allo he performer o brose in he simulaion o use he ransformed sound as a guide o sudy (Figure 11). Figure 8: Alvin Lucier s Panorama glissando curve (Time x MIDI pich) The synhesis simulaion as done ih Csound 3 in he OpenMusic environmen 4. The Csound orchesra as simples as possible (Figure 9) jus a single oscillaor (<oscili>) here he frequency (k1), is driven by a <f2> score funcion 5. 3 hp://.csounds.com 4 Wih o exernal Libraries, OM2Csound and Chroma. See: hp://repmus.ircam.fr/openmusic/libraries. For OpenMusic informaion: hp://repmus.ircam.fr/openmusic/home 5 This able, <f2>, is indicaed as being he las <oscili> parameer (2). 5.2 John Cage s Ryoanji (1983 84) For Ryoanji, he main idea as, likely o Panorama, o build a simulaion, a guide o be used in he sudy performance phase, bu he basic approach as compleely differen. The glissandi here are no linear bu freely dran by Cage. The firs sep as o find a sraegy o rerieve he pich versus ime informaion from Cage score. Some sraegies ere sudied, like border deecion in he picure file, as a mehodology o idenify he curve. Bu for all of hem he preprocessing sep (cleaning he original score, erasing he frames, ec.) as o much ime consuming. Insead of ha e adoped an old compuer echnic, hand digializaion of ploed daa (as done in 6 hp://anasynh.ircam.fr/home/english/sofare/audiosculp 7 hp://anasynh.ircam.fr/home/english/sofare/supervp 8 hps://cycling74.com/
he 70 s). This echnic involves having a background image and ih anoher ool o idenify poins (coordinaes) in he background image. I as broadly used in sea aves digializaion. Figure 13: From firs o second frame, Cage s Ryoanji, firs rombone garden Figure 11: Compuer inerface o brose Panorama simulaion Using he OpenMusic ( Ircam) environmen, e firsly reconsruced he scanned score o have each garden on one linear page o proceed aferards o hand digializaion. As he original score as scanned from a manuscrip, all he frames, from he same garden, has no exacly he same heigh. The boom and upper border ere no exacly parallel and he semi-one subdivisions ere no aligned across all he garden frames. To say ha he alignmen as no so easy (Figure 12). Figure 14: Digialized curves, inside OpenMusic <bpf> objec Folloing he curves digialized (Figure 15), i as easy o isolae hem and use each one as a frequency daa curve o a Csound synhesizer as used previously for Panorama simulaion. Figure 15: Exraced glissando curves Wih his las figure one can imagine he sone oulines, Cage used for he score generaion [12]. While he regisers are given, he ime scale is up o he performer. We choose o give o each garden a ominue duraion. All he gesures ere synhesized as for Panorama, bu as he fundamenal frequencies ere o lo e used a seven harmonics specra o help he uning (Figure 16). The final simulaion sound file, ould be impored in an inerface (as for Panorama), o allo he performer o brose and use i for preparing his performance. Figure 12: Tree inersecions, from Cage s Ryoanji firs rombone garden To resolve his draback, e chose o align he differen frames focusing on he coninuiy of he line describing he gesure (Figure 13). From he reconsruced garden, e proceeded o hand digializaion, ih he help of he <bpf> OpenMusic objec (Figure 14), rying o ork as precisely he zoom allo us o do. Compare ih curves in Figure 3. Figure 16: Ryoanji, firs rombone garden, firs 16 sec sonogram
6. CONCLUSIONS The o pieces reaed here ake advanage of graphical noaion bu presen essenial difficulies in performance. We propose here a compuer sraegy, simulaion based, o assis he musician in he preparaion and performance. Concerning Panorama, hile e found a ay o bring a possible assisance for he glissando conrol and fineuning, he opimizaion for he glissandi posiion changes sill is o do. We look forard o deal ih i mainly ih a consrain sysem. Concerning Ryoanji, he main problem (glissando conrol and fine-uning) sill unsolved. As poined previously, main pars from his ork sill impracicable, from cogniive and insrumenal reasons. Ryoanji evolves in he exreme loer regiser ha makes very difficul o conrol and lisen o he small microonal aves. For he firs garden (here Cage s regiser is from B1 o F#1) he gesures (according ih our hand digializaion) evolve beeen C1 minus 19 cens (32.35 Hz) and a F#1 minus 14 cens (45.88 Hz). We have o pieces mainly based on glissandi, bu ih o differen performance goals. While in Panorama Lucier expecs for a precise pich and ime performance, in Ryoanji, he graphical noaion has no only a score funcion, bu also an inspiraional and poeic purpose. As saed in he Ryoanji insrucions: The glissandi are o be played smoohly and much as is possible like sound evens in naure raher han sounds in music. [8] Composers, inspired by exra musical facs or conceps can use sonificaion as an inspiraion ay. Translaing graphic or visual aspecs in musical noaion can be source of novely and challenge, bu someimes i can be source of difficuly. The noaion graphic space does no have he same properies as he basic pich versus ime real space. If in he graphic space, e can represen any pich inerval size, i is jus a scale quesion dependence. In hearing orld, humans are bounded by heir cogniive pich hresholds. While in he graphic space, e can navigae in he ime line, jumping from fuure o pas and viceversa, in he real orld he ime appears o be irreversible. These o musical composiions ere recorded based upon he presen research [14]. 7. REFERENCES [1] B. Sluchin, M. Mal, Open form and o combinaorial musical models: The cases of Domaines and Duel, 3rd Inernaional Conference on Mahemaics and Compuaion in Music, Paris, June 15-17, 2011, p. 211-217. [2] M. Mal, B. Sluchin, Théorie des jeux e srucure formelle dans Duel (1959) e Sraégie (1962) de Xenakis, L influence des héories scienifiques sur le renouvellemen des formes musicales dans la musique conemporaine, Symposium, organized by M. Grabócz, CDMC, Paris, 28 november 2014. hps://.youube.com/ach?v=zixiyc7il84&f eaure=youube_gdaa [3] B. Sluchin, Linaia-Agon Toards an Inerpreaion Based on he Theory, Inernaional Symposium Iannis Xenakis, M. Solomos, A. Georgaki, G. Zervos (Ed.), Ahens, May 2005, 1 16. hp://cicm.mshparisnord.org/colloquexenakis/ [4] B. Sluchin, M. Mal, Inerpreaion and Compuer Assisance in John Cage s Concer for Piano and Orchesra (1957-58), SMC - Sound and Music Compuing Conference, 2010, pp. 1 8. [5] E. Francioni, Csound for Cage's Ryoanji, Csound Journal, vol. 18, 2014, 1 20. hp://.csounds.com/journal/issue18/francioni.h ml [6] A. Lucier, Panorama for rombone and piano, manuscrip score, 1993. [7] V. Sraebel, The Sonificaion Meaphor in Insrumenal Music and Sonificaion s Romanic Implicaions, The 16h Inernaional Conference on Audiory Display (ICAD-2010), 2010, pp. 1 8. [8] J. Cage, Ryoanji for Trombone Solo, Henmar Press Inc., 1985. [9] J. H. McDermo, M. V. Keebler, C. Micheyl, A. J. Oxenham, Musical inervals and relaive pich: Frequency resoluion, no inerval resoluion, is special The Journal of he Acousical Sociey of America, vol. 128, no 4, 2010, pp. 1943 1951. [10] F. Bianchi, S. Sanuree, D. Wend, T. Dau, Pich discriminaion in musicians and non-musicians: Effecs of harmonic resolvabiliy and processing effor Journal of he Associaion for Research in Oolaryngology: JARO, vol. 17, 2016, pp. 69 79. [11] J. Cage, Ryoanji: Solos for Oboe, Flue, Conrabass, Voice, Trombone ih Percussion or Orchesral Obbligao (1983-85), PAJ: a Journal of Performance and Ar, vol. 31, no 3, 2009, pp. 58 64. [12] M. Foler, Finding Cage a Ryōan-ji Through a Re-Modelling of Variaions II. Perspecives of Ne Music, vol. 47, no 1, Winer, 2009, pp. 174 191. [13] K. Schöning, Ryoân-ji: oder die Gären der Leere, Neue Zeischrif Für Musik (1991-), vol. 155, no 5, 1994, pp. 24 27. [14] J. Cage, Ryoanji, Benny Sluchin rombone, Sylvain Rappapor percussion, CD produced by u_performance@homail.com, 2015, Paris, France.