programming machines and people:

8/13/2019 Programming Machines and People:

http://slidepdf.com/reader/full/programming-machines-and-people 1/4

of improvised music are well awareof the stylistic boundaries of theirrespective improvising traditions:There are rules, and the performermust follow them or consciously dis-regard them.

The addition of an interactivecomputer system within an improvisation adds a new layer of musical

input and complexity that is foreignto many improvisers. Consummateimprovisers and improviser/tech-nician pairs such as George Lewis,Kevin Patton and Seth Paynter, orEvan Parker and Joel Ryan, are more than capable of suc-cessfully incorporating the new musical entity. However manyperformers of “new” music do not come from an improvisingtradition, and the addition of the “computer as improviser”

©2010 ISAST LEONARDO MUSIC JOURNAL, Vol. 20, pp. 25–28, 2010 25

Chapman Welch (composer), Shepherd School of Music, Rice University, 6100 MainStreet, Houston, TX 77001, U.S.A. E-mail: <[email protected]>.

A B S T R A C T

Many performers of new

music do not come from an

improvising tradition, and the

addition of live electronics to

works written for these perform-

ers may be intimidating due to

their inexperience with improvis-

ing and/or working with tech-

nology. Although inexperience

may be a problem, it can be

overcome. The author describes

techniques and strategies for

creating rule-based improvisa-

tion environments with liveelectronics.

Programming Machines and People:Techniques for Live Improvisation with Electronics

Chapman Welch

Improvisation is a word loaded with sociocultural,political and musical meaning. In Western classical music theterm is often bypassed and substituted with “indeterminacy”or “chance music” [1], and the musical structures that theseprocesses generate are given names like “open forms” or“mobile forms.” In jazz, improvisation is the norm, althoughtrends within the art form, bebop and free jazz for example,may serve to expand the original meaning. Despite the social weight of the word or the question whether certain impro-

visations should be labeled Eurological or Afrological [2], themajority will agree that in all but the most simplistic improvi-sational situations there is a certain amount of skill requiredfrom the musician. These skills are usually ingrained during amusician’s early musical development, and their cultivation ispart of a constantly evolving musical, cultural and oftentimesacademic continuum that can be referred to as a style, perfor-mance practice or improvising tradition. Seasoned performers

Fig. 1. Moiré, notes to the performer. (© Chapman Welch)



26 Welch, Programming Machines and People

and simply provides a tempo for the in-strumentalists that are bound by it. Theclarinetist is the only performer who isprocessed by the computer and may im-provise with the materials with no regardto the tempo set by the conductor. Theclarinetist also determines the length ofsections, as it is up to him or her to de-cide when the current state is over and when the next section will begin. The

first example from the score is found inFig. 2. The rules for the ensemble duringthis section are as follows:

ConductorThe conductor begins the count beforeor after the clarinetist begins the boxedpatterns. The conductor may give den-sity cues (cue to play more gestures) togroups and individual instruments, butthey may not cue entrances directly. Theclarinetist gives a cue to begin the nextsection, and the conductor must cue thepiano chord slightly before the clarinetbegins.

Flute and Viola Play any of the three boxed gestures onany of the four beats. It is not necessary to

hensive examples of these methods. Theelectronics were created using the Max/MSP programming environment [4].Notes to the performer (Fig. 1) are givento clarify the musical notation. I will usethe term improvisation exclusively to de-scribe all elements of indeterminacy.

M OIRÉ , FOR CLARINET

SOLOIST, COMPUTER AND ENSEMBLE

Throughout much of the work, the clari-netist improvises by overblowing throughthe harmonic series, inserting shortoutbursts of multiphonic violence andbending notes freely around estimatedpitches. Though the clarinetist is im-provising with the duration and timbreof these materials, the order is set andthe form is not malleable. However, sec-tions I have termed harmonic/melodic states are interspersed throughout the work. Within these states, the soloist and theensemble are given rules with which to

improvise using harmonic and melodicmaterials provided in the score. Althoughthere is a conductor, during these pas-sages, the conductor does not cue events

may be intimidating due to their inexpe-rience improvising and/or working withtechnology. Although inexperience is aproblem, it can be overcome. Improviserand interactive performer Elizabeth Mc-Nutt provides a possible strategy:

When the terms of a piece are clearlyunderstood by the performer, there is acorresponding increase in interpretiveengagement and refinement. With live

processing, for example, it is useful fora performer to understand the results ofher actions on the processed sound out-put, so she can navigate these elementsas part of her larger job of interpretingthe music [3].

In the following examples, I will illus-trate a few of the methods that I haveused for controlled improvisation withinteractive systems. These methods fo-cus on creating rule-based improvisa-tional systems that are both easy enoughthat performers may learn to use themquickly and nuanced enough to allow fora number of interpretations. Although Ihave used these techniques in both past

and upcoming works, I will only discussmy work Moiré, for clarinet soloist, com-puter and ensemble (2008--2009) as itprovides the most salient and compre-

Fig. 2. Excerpt from Moiré. (© Chapman Welch)



Welch, Programming Machines and People 27

cur on a conductor’s downbeat, they arequantized or synced to the conductor’stempo. The result of this is analogousto a sampler instrument whose events

are triggered by certain statistical prob-abilities. The combination of this tempo-rally rigid “virtual instrument” with thefreely improvising clarinet/computerduo yields two unique musical and tem-poral strata that coexist simultaneously. Although the result of this relationship iscomplex and unpredictable, the simplic-ity of the rules allows the soloist and theensemble to easily start improvising withthe materials and with each other.

The second harmonic/melodic stateis similar to the previously discussedexample, but with a few exceptions. Inthis section, the conductor supplies the

rhythmic grid for only the flute, viola andpercussionist, while the pianist, harpistand clarinetist are free from tempo. Theclarinetist again decides when the stateends by signaling the conductor with achange in their melodic material. How-ever, after this signal, the conductor cuesthe ensemble’s transitional material andmoves the piece forward to the next sec-

does not give entrances. The conductoralso cues the end of the page.

During this first state, the computerreacts in a variety of ways, depending on

which materials the clarinetist choosesto play. For example, when the clarinet-ist performs the first set of solid boxedmaterials, this input is routed into a re- verb module in which the amplitude ofthe input controls the intensity of the ef-fect. In contrast, the second set of solidboxed materials triggers short bursts ofthe soloist’s input to be fed into anothereffects system that yields a dramatic,sputtering response from the computer. As this interaction is not notated in thescore, the soloist must learn this behaviorduring rehearsals with the system. How-ever, since the computer’s reactions mir-

ror the musical choices that the soloistmakes, the performer is quickly able tolearn both the cause and the range of thecomputer’s responses.

Although the computer is not process-ing the ensemble, the rules guiding theirimprovisations are inspired by commoncomputer music practices. For instance,since the ensemble’s entrances must oc-

synchronize which gestures are played. Adesignated leader cues the entrances forthe beginning of gestures. The number(density) of entrances is up to the des-

ignated leader. The conductor may givedensity cues (cue to play more gestures)but does not give entrances. The conduc-tor also cues the end of the page. Theend should be as abrupt as possible andshould stop even if the current phrase isincomplete.

PercussionPlay the rolled figure or any of the boxedpitches (singly in any register) on any ofthe four beats. The number (density) ofentrances is up to the performer. Theconductor may give density cues (cueto play more gestures) but does not give

entrances. The conductor also cues theend of the page.

Harp and PianoPlay any of the four boxed gestureson any of the four beats. The number(density) of entrances are up to the per-former. The conductor may give densitycues (cue to play more gestures), but

Fig. 3. Excerpt from Moiré. (© Chapman Welch)



28 Welch, Programming Machines and People

during an improvisation are stored in adatabase. This data is then analyzed inreal time and used to tailor the effectssystem in a type of feedback loop. Whenthe performer is satisfied with the stateof the effects system, they may save thesestates as presets and reload them in fu-ture performances.

As with every improvisation, each per-formance of Moiré is slightly different,

even when performed by the same en-semble. With each ensemble’s renditionof the work, new insight is gained intoboth the system of rules and the musicalmaterials themselves. For me as a com-poser, giving the performer control overthe form and execution of the work isboth liberating and unnerving, as I amultimately relying on the intuition of theindividual players. Still, for an interac-tive musician, this is not a new feeling,as the computer algorithms that respondto a performer’s input often react in un-foreseen ways. However, unlike with thecomputer, these glitches in the instru-mentalists’ “programming” are a resultof their unique musical sensibilities. Asthese sensibilities are the culmination ofa lifetime of music making, they cannotbe understood or predicted by any algo-rithm that I can program. Rather thanseeing this as a weakness in my system, I welcome the unknown, as unpredictabil-ity is at the heart of any improvisation.

References

1. A. Braxton, Tri-Axium Writings, Volume 1 (Dart-mouth: Synthesis/Frog Peak, 1985).

2. G. Lewis, “Improvised Music after 1950: Afrologi-cal and Eurological Perspectives,” Black Music Re- search Journal 16, No. 1, 91--122 (Spring 1996).

3. E. McNutt, “Performing Electroacoustic Music: A Wider View of Interactivity,” Organized Sound 8, No.3 (2003) p. 298.

4. See <cycling74.com/products/maxmspjitter/>.

Received 1 January 2010.

Chapman Welch received his D.M.A. in music composition and electronic music from theUniversity of North Texas, where he workedat the Center for Experimental Music and In- termedia (CEMI) from 2001 through 2006.He has taught at Rice University, the Univer- sity of Houston and the University of NorthTexas. Currently, he serves as the supportspecialist for the Rice Electro-Acoustic MusicLabs (REMLABS). Welch’s music has been

presented at numerous festivals in the UnitedStates and abroad, including the LaTex festival, June in Buffalo, SPARK, Hawaii Inter- national Conference for the Humanities, the

Florida Electro-Acoustic Music Festival, ICMCand SEAMUS conferences.

to and accompanies the soloist with adistant, ghost-like doubling of the frag-mented yet plaintive clarinet melody. Bymeans of pitch-tracking, certain pitchesroute the soloist’s input into a series ofeffects systems, creating a synth-like, pul-sating ostinato. Once again, the clear re-lationship between soloist and computerallows the performer to quickly begin im-provising with the invisible partner.

As in the first state, the ensemble’srules for improvisation are also inspiredby electroacoustic techniques. In thisexample, the conductor again providesthe tempo for the flute, viola and per-cussionist’s virtual sampler. The pianoand harp are free from this tempo andserve to reinforce the harmonic mate-rial found in the computer accompani-ment. Although the computer materialis static and drone-like, the harp and pi-ano emerge from the texture in waves ofundulating energy before disappearingback into the computer harmony. This yields an effect similar to what would becreated if the computer harmony werebeing processed using granular synthesis.Furthermore, due to the similarities be-tween their materials, the harp and pianosound as if they were actually manipulat-ing the computer material. This relation-ship creates a sonic environment where itbecomes difficult to tell who is affecting whom as all the parts merge into a cohe-sive and symbiotic musical unit.

F UTURE W ORK AND CLOSING R EMARKS

The previous discussion provides an over- view of how I have implemented impro-

visation within a concert music setting. Although Moiré represents the culmina-tion of my past experiments, there are anumber of projects that I am completingthat use these techniques in new ways.For instance, I am currently finishing apiece for erhu, yangqin, bamboo fluteand laptop ensemble. As in Moiré , thereis a conductor who provides a rhythmicgrid to which the laptop musicians sync.Instead of triggering samples, the laptopmusicians use these rules to choose boththe timing and the types of computerprocesses that are applied to the improvising musicians’ input. Unlike in Moiré ,in some sections the instrumentalists are

allowed to improvise freely on writtenmelodies with no instructions or rules.This lack of instruction owes to the per-former’s expertise in the rich traditionof Chinese melodic ornamentation ( jiahua ). In addition to this work, there isa version of Moiré planned in which in-formation about a performer’s choices

tion. The second example from the scoreis shown in Fig. 3. The rules for the in-strumentalists during this section are asfollows:

ConductorThe conductor begins the count beforeor after the clarinetist begins their boxedpatterns. The conductor may give den-sity cues (cue to play more gestures) to

groups and individual instruments butmay not cue entrances directly. Theclarinetist gives a cue to begin the nextsection, and the conductor must cue thefinal boxed material for the flute, viola,harp and piano during the extended c-sharp glissandi in the clarinet part.

Flute and Viola Play any of the four boxed gestures onany of the four beats. A designated leadercues the entrances for the beginning ofgestures and also cues which gestures areto be played.

The number (density) of entrancesis up to the designated leader. The con-ductor may give density cues (cue toplay more gestures) but does not giveentrances. The conductor cues the finalboxed material during the extended c-sharp glissandi in the clarinet part.

PercussionPlay any of the phrases as written. Phrasesmay be repeated as many times as desiredand may even establish a rhythmic pulse.The number (density) of entrances is upto the performer. The conductor maygive density cues (cue to play more ges-tures) but does not give entrances. Theconductor cues the next section.

Harp and PianoPlay the boxed figures in order as in-structed (starting from nothing, cre-scendo, and fade to nothing over 5--15seconds with 1--8 seconds of rest). A des-ignated leader cues the entrances for thebeginning of gestures. The conductorcues the final boxed material during theextended c-sharp glissandi in the clarinetpart. This should be played immediatelyeven if the current phrase is incompleteor all boxes have not been played.

The clarinetist plays through the com-plete boxed melody one time beforemaking new melodies using the dashed

boxes as melodic cells for their improvi-sation. Like the “acoustic sampler” fromthe previous section, this melodic processalso has an electroacoustic origin as theclarinetist’s reworking of the melodic ma-terial is not unlike live sampling where in-put is sampled, fragmented and remixedin real time. The computer responds

programming machines and people:

Documents