Biography and Background of the Composer

Richard Karpen is a composer of computer and electroacoustic music, as well as of symphonic and chamber works for a wide variety of ensembles. His works are widely performed in the United States and internationally. He is also a contributor to a textbook on Csound and Computer Music (The Csound Book, editor: Richard Boulanger). His articles have been published in periodicals including Computer Music Journal, Perspectives of New Music, ARRAY, and Musicus. As of May, 2002, Karpen lives in Seattle, where he teaches at the University of Washington. He currently directs the Center for Advanced Research Technology in the Arts and Humanities (CARTAH), the School of Music Computer Center, and DXARTS (Center for Digital Arts and Experimental Media); he is the founder of both.

Karpen was born in 1957 in New York City, where he lived throughout his post-high school education. He was at the Paris Conservatory of Music in 1976 and also attended the Conservatory of Music at Brooklyn College from 1983–84. He holds a BA (1983) from the City University of New York, where he graduated summa cum laude, as well as a DMA (1989) and MA (1986) in composition from Stanford University. From 1985–89, he worked at the Center for Computer Research in Music and Acoustics as a teaching and research assistant at Stanford. He studied composition with Charles Dodge, Gheorghe Costinescu, and Morton Subotnick in New York.

Karpen was a Fulbright Fellow in Padua, Italy, and has been the recipient of numerous awards and grants for his work. These include awards from National Endowment of the Arts, the ASCAP Foundation, the Bourges Festival in France, the American New Music Consortium, the Seattle Arts Commission, the NEWCOMP Contest, Luigi Russolo Contest in Italy, the National Flute Association, a Leverhulme Visiting Fellowship to Scotland, and an invitation to work at IRCAM through Stanford University's Prix de Paris.

Major international festivals that have featured performances of his works include the Gaudeamus International Music Week in Amsterdam, the Warsaw Autumn Festival in Poland, the Sidney Spring Festival in Australia, and the International Computer Music Conferences. He has appeared as a lecturer, adjudicator, composer-in-residence and presenter at many of the leading institutes for Computer Music internationally (most recently the Institute of Sonology, The Hague; CCRMA, Stanford; CNMAT, Berkeley). His music has been released on disc by Le Chant du Monde/Cultures Electroniques, Wergo/Computer Music Currents, Centaur, and Neuma.

Karpen is acknowledged as one of the leading international figures in Computer Music for both his pioneering compositions and his work in developing computer applications for music composition and sound design. [72] 

Overview

In 1986, Richard Karpen was commissioned by Australian flutist Laura Chislett to write a piece for flute and computer-generated sounds. In 1987, the piece was completed and premiered by Ms. Chislett. The result was a work that contained complex, richly layered computer sounds, which alternately were synchronized with and in contrast to a virtuosic, highly varied and exciting flute part.

Karpen said that part of what he enjoyed about the flute in this piece was the "agility, the ability to change ranges quickly, and the sound" of the flute. [73]  The piece has a duration of 13'05", and challenges the flutist in that it contains the use of many extended techniques and demands a high level of technical ability as well as organizational and conceptual maturity. The tape part is comprised of a variety of sounds, ranging from clearly pitched sounds, to evolving pitched sounds, to sounds that are less distinguishable in terms of pitch, being very rich, complex, and constantly shifting in their composition. The piece can be divided into three main sections, which are roughly equal thirds of the piece. The typical compositional tools are exploited fairly thoroughly; that is, Karpen utilizes dynamics, speed, color, range, interval, pitch, and rhythm, each to a full extent. Additionally, in both the flute and computer parts, he extends the range of conventional practice. In the tape part, the manner of pitch generation, as well as the composition and evolution of the sounds is original and creative. For the flutist, the use of extended techniques as well as extremely fast passage-work, organization of material, and tight synchronization presents a challenge and an extension of the conventional limits of flute playing.

Computer Sounds

The computer-generated part for Exchange was realized at the Center for Computer Research in Music and Acoustics (CCRMA) at Stanford University. When composing electroacoustic music, Karpen has said that he always begins with the computer part, and has a main outline, sometimes even an almost final version, of the entire piece before beginning the instrumental part. This is true also of Exchange. The material is occasioned by exploring the sounds, generated by the computer, and letting the music grow out of the sonic material, rather than by imposing musical ideas onto the sounds. This is in part to avoid influencing the composition of the tape part with the technical qualities of the instrument, which could be limiting. In each of his pieces, Karpen tries to "explore new territory", and finds this easiest when working solely with the sonic possibilities of the computer, rather than being limited by his knowledge and familiarity with the sounds of the acoustic instrument.

When composing using a computer to generate sounds, there is an immense and different universe of sounds, and methods of creating sounds, from that of an acoustic instrument. One of the biggest difference is in the concept of pitch: on acoustic instruments, pitch is most often thought of as being fixed and easily defined as a single note. Different possibilities are available with computer-generated sounds, where a sound can range from a single frequency to an extremely complex sound with no distinguishable pitch. In the end, these differences in the concept of pitch can end up affecting the writing done for acoustic instruments. Karpen has said of his interest in pitch:

"One of the things I have been interested in is the blurring of the distinction between sound, color, pitch and harmony (by harmony I mean simply the sounds occurring simultaneously). This was what I focused on in the making of the computer-generated part of Exchange. I will try to explain simply how it works in that piece. Every 'sound' is made up of a complex of sounds. Sometimes one can hear a discrete, separate sound, but often one hears a continuum of changing material. In every case the basic source of sound is a configuration in which a fundamental frequency is present along with four resonating frequency regions. Two of these four regions are stable; that is, they maintain their ratio relationship with the fundamental to which they are harmonically related. The other two, however (higher in frequency than the stable pair), change their frequency over the duration of the sound (the duration, incidentally, varies between .2 seconds and approximately 60 seconds, depending on the section of the piece)." [74] 

The moving resonators are harmonically related to the fundamental, and the result of the movement is a kind of melody of the upper harmonics over the fundamental. The result of these four regions is that for each sound produced, there is the harmony between the fundamental and the audible harmonics, and a melody produced by the moving resonators, and that the melody and harmony are simultaneous and comprise a single sound, albeit a moving one.

Considering each of these sounds as a single note, the next step is the simultaneous experience of more than one note. In Exchange, Karpen rarely writes for separate notes that have a synchronized attack (fixed chords); rather, the notes are overlapped, both in start and stop time. This technique is one of Karpen's favorite techniques in generating sound by computer. Quite often, the sounds are of the same duration as each other, but their attacks and decays are offset by a certain intervals in time. In this way, many notes will be sounding at a given moment, but the experience of the sound itself is always changing. Sometimes the result is a wash of sound, but a sound that is highly organized and detailed, with parts that are strongly related and not at all arbitrary. In order to avoid the possible pitfall of a shallow or uninteresting tape sound, Karpen uses techniques such as this, as well as favoring complex and dense sounds. The exceptions to these are, at times, percussive sounds.

Some of the sounds heard in the computer part include whip sounds, chords, and dreamy sections. The sounds can be described as pitched, unpitched, short, long, loud, soft, and the textures heard.

Form

Exchange can be formally divided into three large sections. Within and between these sections there is no literal repetition; the unity is achieved by similarity of rhythmic gestures as well as by character and energy of the sounds. The first section lasts from the beginning until 5'39", the second section from 5'39" until 8'26", and the third section from 8'26" until the end, at 13'05". The character between each of these sections is quite distinct; especially interesting is the relationship between the tape and the flute in each. The first two sections are comprised of subsections, and these of phrases and/or gestures. The overall movement through the piece seems to be one in which the flute and tape begin as independent voices, with very little relationship to each other. In the middle section they are tightly woven in an intricate exchange, sometimes in opposition and others not. In the final section the two voices seem to finally come together in calm resolution.

During the first section, the tape part acts almost more as a universe in which the flute is contained. The two do not appear as equal voices. The flute part is far more active and filled with a great deal of variety, while the tape part moves slowly with little concern for the antics of the flute. It is in this section that most of the unusual sounds occur for the flutist, all of the extended techniques aside from the "dirty flute" and the circular breathing occur here. The fast notes in this section have an air of youthful energy, while the scattered smaller gestures depict curiosity and exploration. The first of several sections comes at the beginning, of course. The first sounds we hear are three whips, the third of which cues the flute to begin. The first passage for the flutist is very fast and the notes twist back and forth, leaping through the middle and high registers. The impression is that the tape whipped the flute into action. After the whip sounds, though, the tape becomes more of a background to the flute's activity, and sustains long notes for quite awhile. The flute's gestures become shorter, and contain many percussive sounds, eventually moving to sustained tones at 0'35". This is short-lived, and the flute makes its way back into more energetic and percussive sounds by 0'49". The multiphonics between 1'54" and 2'49" are accompanied by thicker sounds of shifting chords in the tape, reinforcing the richer and more textured sound in the flute. The fast notes are resumed after 0'49", until another subsection begins at 1'28". At first, the connection between flute and tape comes in texture. At 1'27" and 1'28", the flute finishes a phrase and tapers off; the tape increases its texture up to the same point, and then also tapers off. This section contains many multiphonics over the computer's richly complex shifting chords. The texture in the tape gradually increases, with a large surge from 3'00" until 3'15". Another subsection begins here as the tape and flute each begin a new motive. As the section moves on, the flute and tape begin to interact more. Beginning at 3'16", there is more interplay between flute and tape. The tape has legato, overlapping pitches, over which the flute has slurred, flowing, even notes. At 3'24" the tape becomes silent and the flute becomes articulated and short. Then a few seconds later, both parts become legato again. The mood here is flowing, dreamy, and fluid.

The mood abruptly changes again at 3'58", where the tape violently restates the whip sound we heard in the opening. The change here is so drastic and abrupt, it is tempting to call this the beginning of the second section. But, the tape sounds continue the legato dreamy sound, albeit more textured and active, of the previous subsection, and while the flute is indeed faster, the relationship between the flute and tape has not profoundly changed, yet. There is still a great variety of note values, interruptions by rests, and the interjection of multiphonics, and the tape and flute have not shared any material directly. The end of the first section, from about 5'22" to 5'38", contains long and short notes, flutter-tonguing, singing while playing, pizzicato, tongue rams, and tongue clicks. In other words, it restates many of the sounds we've been hearing throughout the first section in a final unifying phrase.

The second section begins at 5'39". This section is characterized by fast unison notes in the tape and computer, of equal and regular lengths, and long phrases, and a departure from the mixed note values, extended techniques, chords, varying textures, and broad sense of time found in most of the first section. This section is consistently very high-energy, and although the dynamics vary somewhat, they tend toward generally loud dynamics. The notes throughout the section move at the same pace, about nine or ten notes per second. Even when the flute is not playing, the computer sustains the repeated rhythm, effecting a constant ostinato rhythm. The exceptions to this are at 6'13", 6'57" through 7'20", and 8'20" through 8'22". At 6'13" and 8'20", the music suddenly comes to a complete silence in both parts; the effect of this silence depends on the constant run of notes preceding the silence. At 6'57", the movement is interrupted by a sudden sustained chord; a harkening back to the first section. This is temporary, however; without developing this idea at all, the fast, repeated notes begin again. From 7'29" until 8'20", neither the flute nor the computer has any moments of rest; this is the longest stream of uninterrupted notes in the piece. From a formal standpoint, this would represent the climax of the piece. This is also the time of the most difficult synchronization between the two voices, and the point at which this is the most crucial.

The third section is completely different from either the first or the second. Beginning at 8'26", immediately following the end of the fast notes, the computer slows back down, reminding us of the timeless sustained chords heard at the beginning of the piece. This time, however, the flute joins the computer in these sustained sounds. The flute overlaps pitches with the computer, much like the way the computer overlapped sounds with itself earlier. The effect here is a layered, evolving sound, in which the flute and tape blend and move similarly. Still, the flute has more motion and some quicker notes, as though it's not quite convinced to let go of its previous material. The dynamics are softer, with occasional bursts of energy, before finally joining the computer at the end. In the last two minutes of the piece, the computer finally has more motion than the flute. The flute is almost background to the computer at the end, just as the computer was background to the flute at the beginning. The piece slowly fades away to nothing, the flute tapering away slightly after the tape.

Pitch / Intervals

A great deal of the pitch material in Exchange was derived from algorithms, and then altered to varying degrees for expressive purposes [ my interview ] . One of the ways the algorithms functioned was as follows: a long pitch set was chosen, somewhere between twenty and forty pitches or so. Then the set would be repeated, this time overlapping itself by a certain number of pitches. This overlapping could happen many times, creating an endless stream of pitches. The result is that pitches that were repeated in the original set would be repeated even more in the each resulting set, gaining more and more emphasis as the set was manipulated. Karpen used this kind of algorithm to generate the pitches for both the flute and the computer parts at times, freely changing pitches as he chose to. The two most noticeable characteristics of the pitch content of the piece are the lack of repeated patterns alongside the clear repetition of certain pitches. There is no place in the piece where a pattern can be identified, yet there is a subconscious sense of unity throughout.

For instance, in the first few phrases, every pitch is used at least once, but no two pitches are associated with each other. There is a high frequency of the note C, above all others. Later, at around 1'04" through 1'14", there are more C#s than any other note. As a performer, the pitch material doesn't hold any patterns that would aid in performance.

Because of the manner in which the pitches were generated, there also aren't many repeated intervals to guide listening or performing the piece. On the other hand, for the very same reason, there are actually some intervals that will occur more frequently than other intervals. Compared to the frequency of recurring intervals in conventional music, though, these are quite subtle. The exception, as already noted, is the occurrence of unisons in the form of sequentially repeated notes. There are times when there is a lack of repeated notes, as well as a lack of repeated intervals or associations between notes, such as from 7'31" to 8'20". This is in contrast to the phrase from 6'44" to 6'58", in which notes are often repeated three or more times sequentially. The contrast is emphasized in the tape part as well. In the latter example, the lower voice of the tape part has two simultaneously sounding notes repeated ceaselessly under the upper tape part and flute. Meanwhile, the upper tape part is in perfect unison with the flute, emphasizing further the repetition of certain pitches. In the former example, though, the lower tape voice, beginning at about 7'35" until about 7'48", has varied pitches that do not match those in the flute or in the other tape part. After this, the lower tape voice again goes back to the repeated bass note.

In short, the pitch and interval materials are quite varied and the most important characteristics to notice are the repeated notes, the unisons between flute and tape, and hearing the difference between sections with repeated notes and those without.

Rhythm

When discussing rhythm in this piece, it seems more appropriate to describe how the music moves through time rather than to compare minute, finite rhythmic gestures. The rhythms serve to articulate form and time, as well as to define much of the relationship between the flute and the computer. These rhythms are not particularly complicated, but are varied greatly. The piece has little perceptible pulse, even though the method of measuring time (and rhythm) is by dividing the piece into one-second pieces. Each of the three sections of the piece uses rhythm quite distinctly, and each moves through time with different energy and pacing. The first section has the greatest variety of note values and perceptible speeds, containing very long sustained notes, very quickly moving notes, and unequal rhythms throughout. The second section, on the other hand, contains notes of almost entirely the same note value; the notes are mostly nine or ten to the second. There are very few rests, resulting in long streams of equal notes. The third section is again quite different from both of the previous sections; this time, the notes are primarily sustained, with occasional interruptions by quicker notes. The music moves slowly, and the many tied notes and unequal note values disguise any possible feeling of pulse, emphasis, or predictability.

The first section of the piece contains music of many different characters. Some of the gestures are crisp and energetic (such as the flute's opening gesture), some are more dreamy and legato (such as at 3'18"), some slightly bizarre and mysterious (5'22" through 5'31"), and some focusing on the sounds themselves, rather than the shape of the gestures (1'31" through 3'00"). The opening sound of the computer immediately defines the character of the opening of the piece, and indeed of much of the piece itself. The whip sound is sharp, energetic, aggressive, and crisp. The ensuing flute line takes up this mode with a string of 32^nd notes, mixed with some 16^th notes and rests. Quickly, different note values and rests become part of the mix, including triplets, sextuplets, and so on. There is the clear avoidance of repeating the same rhythmic gestures, especially in succession, and the use of many ties and compound rhythms. Also, there are often notes placed in the middle of the one-second division, with rests before and after it. In this way, Karpen avoids imposing a feeling of pulse on the piece. However, it is possible, in this piece, to feel a pulse, because the notes are divided and beamed according to the one-second increments indicated. Because of the variety of gestures, the ties, and the placement of notes within the one-second increment, a pulse is rarely felt. Better put, the rhythms do not seem to bear a strong relationship to the pulse, and the one-second increments serve only to place the notes in relationship to each other. Clearly, the way that the music moves through time in this piece is going to be much different than with a piece for which the pulse is the primary rhythmic driving device. The music in Exchange moves in a freer, possibly more linear way. Because of this same lack of pulse, the experience of the motion of time becomes focused on the larger gestures instead, and more attention is brought toward large changes of gesture and noteworthy sonic events. In this first section, many different events occur, each one drawing attention toward itself and away from the previous material. Quick energetic passages give way to sustained notes and slow-moving figures, and then back to different high-energy gestures. In this way, we experience the ebb and flow of the piece, become drawn into its life cycle. Between 1'04" and 1'28", there are many 32^nd notes and 32^nd note nonuplets in succession, articulated at times by sharp sounds in the computer landing exactly at the beginning of the groupings (1'21", 1'23", and 1'25"). This regularity and the high speed of the notes generate quite a bit of energy, and propel the music forward. Before the gesture becomes too regular, though, the computer moves to a sustained sound (1'25") quickly followed by the flute (1'28"). The ensuing section is very luxurious and slow (1'28" to 3'00"); the third section of the piece bears a great resemblance to this section. The focus is on the swelling and relaxing of the dynamics, the exploration of the different registers, and especially of the sounds produced in both the flute and computer.

From this section we move into one of much more regularity and motion, from 3'16" to 3'58". This is the closest relationship yet between the flute and computer; both are moving at about the same pace, in a low register, very sustained, and flowing. The rhythm is comprised mostly of sextuplets, and even becomes predictable. The computer notes are layered over each other to create a blurred effect, which causes an even more sustained and relaxed quality. There are four phrases of this gesture, each separated from the others by silence in both the flute and computer. This is abruptly interrupted by the familiar whip sound, which launches the flute into high-speed action again at 3'58". The material here is much like that which comprises most of the second section of the piece, except that here it is still irregular and the computer continues the layered, sustained sounds of the previous phrases. The flute explores multiphonics, singing and playing, flutter-tonguing, tongue clicks, and also bent pitches during this section. Gradually, the computer part becomes less sustained, the notes become shorter and do not overlap as much (4'59" through 5'12"). Eventually, accents are added to the computer sounds, and both the flute and computer sound short, crisp and articulated. This again is suddenly changed at 5'22", where both parts become dreamy, layered and sustained. This short phrase is cut off abruptly at 5'32", and both parts become aggressive again, increasing the energy and aggressiveness to bring us into the second section of the piece.

The second section lasts from 5'39" until 8'26", and is characterized by fast and short notes in both the flute and computer, high energy, notes of even-length, and a high degree of synchronization between the flute and computer. Throughout the section, almost all of the notes are the same length; that is, about a ninth or tenth of a second each. It is important to note that the flute and the computer are exactly synchronized throughout the section, which is significantly different from either the first or third section. At the beginning of this section, both the flute and the computer replace some of the notes with rests. When one voice is resting, the other is not; that is, the flute and computer overlap so that there is a note for every division of every second, resulting in a constant stream of even notes. More rests occur toward the beginning of the section, and eventually less occur, so the phrases become longer and longer. The energy is built up by the increase in the length of the gestures as well as a general move toward the upper register in the flute, until the sudden silence at 6'13", where both flute and tape become silent for slightly more than a second. They both resume at 6'14", this time much lower and with more rests between the notes, resulting in a very rhythmic, almost throbbing gesture. There is definitely a sense of pulsating, although it would be difficult to find an exactly regular pulse.

The phrase at 6'44" has probably the strongest sense of pulse in the entire piece. This is because for twelve seconds the divisions are exactly the same; there are nine notes per second. This can be felt at a pulse per second, subdivided into triplets in each pulse. The only break in the incessant fast notes comes at 6'57", where the flute becomes silent and the computer holds a sustained chord. The chord lasts with discernible motion until 7'19", where the running notes are resumed. This time, the computer has a throbbing effect; this is done by layering the notes between three voices in the computer in groups of three. That is, one voice will utter the same note three times, followed by two "notes" of silence. The second voice will do the same, but beginning on the first voice's second note. The third voice will begin on the first voice's third note. Each of the voices is in a different register. The effect is one of continuous notes that swell up and down through the three registers, and that there is a very subtle division of the notes into groups of three. When the flute joins in, at 7'29", this motion is continued, and most of the groupings are into groups of nine, with occasional groups of ten or eight. There is a subtle feeling of pulsation approximately every three notes, which would be equivalent to a metronome speed of 180 beats per minute. But the groupings are occasionally different, breaking up the predictability. Also, the melodic shape of the gestures does not reinforce the groupings; the high, low, and repeated notes happen anywhere within the groups. It is important to note here that Karpen has stated that were he to renotate this section, he would put all of the notes under one beam and leave the organizing to the flutist. This implies that he did not intend there to be a discernible pulse anywhere in the piece. The section begins to come toward its end at 8'09", where the notes gradually become more and more broken up by silences. A "false" ending is reached at 8'20", with a complete silence, but one last statement must be made (from 8'23" to 8'26") before the section is closed.

The beginning of the third section is quite clear, as the constantly running notes that we've been hearing for five minutes suddenly give way to a long sustained chord at 8'26". This last section has a sustained, layered, almost transformational character. It is quite slow throughout, and the changes that do occur evolve slowly and gently. Karpen again uses the device of layering many notes, each with different start and end times. This time, though, the flute joins the computer in this layering effect; the two voices share many of the same pitches and melodic contour, but never start or end the notes at the same moment. This results in a very close relationship between the two instruments, one in which their functions are much less distinct from each other. The computer contains many notes that enter and leave and different times, and in this way creates an unpredictable rhythmic motive. The flute is unable to do this, and instead contains somewhat complex and varied rhythms, along with tied notes, to avoid any sense of regularity or predictability. This isn't done in order to be complicated; it simply is a device to keep the music timeless, liberated and floating, rather than burdened by pulse or predictability. The sense of motion is slow and nonlinear, in contrast to the driving rhythms of the second section of the piece.

As this third section progresses, the note values slowly become longer. At 10'30", the flute is silent for 45 seconds while the computer slowly shifts sounds, preparing for the last phrase. At 11'15" the flute enters with the longest note of the piece, a 50 second long second-register F. This is the only time in the piece where both the flute and computer are playing and the computer has the faster-moving part; it is noteworthy that at the end, more attention would be drawn to the computer sounds. After this long note, the flute utters two shorter notes, and then the final note of the piece is also quite long, this time a 45 second long note. This time the computer changes less than on the previous long note, and eventually the computer sounds fade out, leaving the flute to end the piece for two seconds alone.

Register / Range

Exchange utilizes the full range of the flute with no use of the altissimo register. Most of the piece lies in the middle and high registers, with the occasional visit to the lowest and highest registers. Throughout the piece, leaps are mingled with repetitions on a single note. There is very little stepwise motion, even in the region with the smallest range, and there are no scalar passages. The lowest part occurs from 6' 14" to 6'31" during the second section of the piece, and the highest at 3'58", during the first. The greatest range is covered during the first third of the piece, with many leaps, and these leaps sometimes cover as much as one and a half octaves in a single bound. This is in accordance with the idea that the first section of the piece is one of exploration and experimentation. The smallest range occurs during the fastest section, staying primarily within an octave. It may be surprising that the section with the highest energy and the most climactic music contains the smallest range. It is so very common to use a wide range to support a high-energy gesture, but Karpen instead uses the strength of the driving gesture to effect a climax. It is precisely because the notes are so similar in speed and range that the music is propelled so effectively. Had the leaps been larger or the rhythms more disjunct, the energy of this section would have been broken up and dissipated. The narrow focus serves to concentrate the material. The last section of the piece has only a slightly larger range than the fast section, with the highest note reaching a C#. The increase is range in the last section serves to open up the sound, to support the feeling of liberation and timelessness that brings the piece to a close. The lack of stepwise motion also serves this end, keeping the spaces open and without linear direction.

Dynamics

The dynamics of Exchange function very much in the same way as the range and register do, and these elements work together to support the musical ideas. Extreme and sudden dynamics are a characteristic of the first section of Exchange. Just as the rhythms in this section are greatly varied and the greatest range is explored, the dynamics support the characteristic of youthful energy, curiosity, and exploration of this section. The second section has a smaller range of dynamics, which works together with the very narrow range to concentrate the energy, rather than dissipate it with too much variety. The dynamics in this section are varied somewhat at first (the performer is directed to play "mp, mf, f; ad lib" [score]), becoming mostly forte and fortissimo deeper into the section. This section gets its power from the unceasing, driving qualities, with very little variety in articulation, rhythm, dynamics, or range. The last section is very smooth and continuous, with some swells up and down. Many different dynamics are used, ranging from pianissimo to forte, but in contrast to the first section, they do not occur suddenly or abruptly; rather, they move smoothly from one to another, and have a sense of breath and gentleness. The piece ends on a note that moves from mezzo piano to forte to mezzo forte and ending on a mezzo piano. It's clear that Karpen does not associate dynamic levels with cliche ideas of climax or energy. Rather, the dynamics are used artfully for conceptual emphasis rather than a purely sensual experience.

Extended Techniques

Exchange has many interesting sounds and techniques, and some of them will challenge the flutist to either try a new technique or become more proficient at an already-familiar technique. The extended techniques include tongue clicks with the embouchure hole open, tongue clicks with the embouchure hole closed, key clicks with embouchure hole closed, extremely strongly tongued articulation, tongue stop with the embouchure hole closed, singing and playing simultaneously, multiphonics, circular breathing, harmonics, bent pitches, pizzicato, and dirty flute.

The performance notes contain a key describing each of these as well as providing the symbol used for each in the music. A tongue click is just that; a clicking noise made by the tongue, much like the noise sometimes associated with clucking. When clicking with the embouchure hole open, having the mouth open wide aids in the resonance of the sound. When the clicking is done with the hole closed, the entire hole is sealed by the flutist's mouth, and the clicking sound resonates through the flute. Tongue clicks are found at 0'20" and 0'31". The "extremely strong tongued articulation" is done by blowing very forcefully into the flute with a "too" articulation, in an almost percussive manner. This technique is used from 0'49" through 1'00". A tongue stop is a fun technique in which the flutist blows into the flute while completely sealing the embouchure hole, and then suddenly stops the sound by forcefully thrusting the tongue into the hole. This abruptly stops the sound with a kind of a thudding or popping result. This is found at 0'18" and 0'26".

Singing and playing is a more familiar technique in which the flutist plays the note as written and sings at the same time. The ratio of flute to voice can vary quite a bit, but generally equal amounts of each sound is very effective. In Exchange, there are occurrences in which the flute and voice are on the same pitch (1'00" through 1'07") and others where they exact separate pitches (0'43" to 0'48").

The multiphonics in this piece range from stable ones to fairly unstable ones. All of the fingerings are provided above the note, and all of the expected pitches are notated. The multiphonics begin at 1'11", with a series of quickly changing methods for playing a C# as part of a multiphonic. The series begins with a normal C# (third register) followed by four different multiphonics using the note C#. The third one is described as "diffuse", and the fourth as "brilliant" [score]. These move very quickly (two per second) and it is not easy to effect a full sound in that amount of time.

I would suggest practicing each multiphonic separately thoroughly, becoming extremely comfortable with the fingerings as well as physically memorizing the necessary embouchure. When the multiphonic seems more dependable, it may be useful to play a series of these, attempting to do two per second successfully. This should be done with each of the multiphonics separately. When they are all thoroughly learned, the next step would be to practice two consecutive multiphonics by alternating them, gradually becoming comfortable with the speed of the finger and embouchure changes. Only then would it be advisable to try the succession of four multiphonics, also including the normal C#'s at the beginning and end of the group.

The other multiphonics occur from 1'54" through 2'49", 4'55" through 4'57", and 5'32" . Some of these are easier than others; for example, the one at 1'54" includes three pitches and is somewhat delicate, while the following two at 2'02" through 2'07" are more stable and can be played quite loudly.

Harmonics are not at all a new technique, yet warrant a mention here because they are not completely conventional. Harmonics are produced by overblowing a given fingering in order to achieve a harmonic pitch of that fingering; usually the fundamental fingering is of the note an octave or an octave and a fifth below. The harmonics in Exchange occur at 0'35" through 0'40", 1'59" through 2'01", and 2'52" through 2'56". The most challenging use of harmonics is at 2'52", where there is a series of five harmonics, the first two being a twelfth above the fundamental, the third an octave, the fourth two octaves, and the fifth a twelfth again. It takes a bit of practice to navigate the different embouchures and air speeds of these quickly changing harmonics, but the end effect is a very nice variety of colors.

The bent notes refer to notes that are bent in pitch upward or downward from a given pitch, in this case with no destination pitch being marked. Pitches can usually be bent in two ways: the first is by altering the fingering or sliding the finger gradually off of one of the tone holes, and the second is by changing the angle of the airstream to cause the pitch to go flatter or sharper. In Exchange, a bent pitch occurs at 4'57", and since it bends downwards is most easily executed by directing the airstream downward at a steeper angle so the pitch becomes flatter. The flutist can either roll the flute in towards their body or duck the head (or lips) toward the flute, or both.

Pizzicato refers to an extremely short, crisply tongued note, reminiscent of a pizzicato produced by a stringed instrument. The pizzicato found at 5'34" simply needs to be short, loud, and accented. The "dirty flute" or "breathy" sound technique is simple also, and refers to a note that is played with rough sounds, rather than the conventional clear sound usually desired. The dirty sound contains frequencies including the fundamental as well as many overtones at fairly high volumes; the result is that the tone contains many sounds, much like a multiphonic. Karpen directs the flutist to use a "breathy" sound from 6'14" to 6'36". This is the same technique used in Antas' piece, A River from the Walls, in which she uses the term "dirty flute".

The last extended technique that Karpen calls for in Exchange is circular breathing. While this is a widely recognized technique, it is not a widely learned one. Circular breathing is the act of breathing in and exhaling at the same time, resulting in a continuous, uninterrupted tone on the flute. There are two notes in Exchange that are to be produced using circular breathing, both at the end. The first one is a second register F that lasts for fifty seconds, and the second (the last note of the piece) is a B above the F, and lasts for forty-five seconds. In the score, Karpen states: "If the performer cannot circular breathe, the long tones should be broken into segments (as few as possible). The performer should plan their breathing to make the last segment the longest and as long as possible." A good description of circular breathing and how to achieve it is described in Robert Dick's book The Other Flute. When I first performed this piece, I had not learned to circular breathe, and attempted to do each of the two notes in one breath. In practice, I was regularly successful, but in performance this proved much more difficult to achieve. My preference at the time was to do it without circular breathing, because circular breathing so often causes irregularity of tone. Karpen stated that the irregularity was part of what he enjoyed about this technique, however.

Notation

The notation of this piece is conventional in most regards, except for some unusual beaming and occasional extensions of normal notation. For example, at 0'55", there are many notes included under a beam that branches out into three beams. This notation (used in many 20th Century pieces) denotes acceleration from the beginning to the end of the beamed pitches. All of the pitches are included for the multiphonics, as well as fingerings above the notes.

Synchronization

Synchronizing the tape and flute parts in Exchange involves two different skills for the flutist; one is to know when to play what, and the second is to actually play it extremely accurately. The score contains cues, and the tape part contains many distinctive events, as well as minutes / seconds noted regularly, so knowing where to be is not very difficult. There are times, though, when there is so much attention needed just to keep track of the notes going by that it is difficult to glance down at the time and at the stopwatch to see if things are on track. The section from 5'39" to 6'12" is one of these times, where the notes in the flute and tape parts are going by at about nine or ten notes per second, and it is possible to go too slowly or too quickly. The synchronization just after 6'12" is extremely important, because the flute and tape should come to silence suddenly and together. The other section that is very long and notey, and because of that difficult to synchronize, lasts form 7'29" to 8'19". To be more accurate, the difficult section in terms of synchronization last from 7'29" to 8'09". At 8"09" the tape part becomes suddenly much more distinguishable and the flutist can jump to this spot if he / she is indeed out of synch with the tape. This, of course, is not ideal, but is a good idea, because at the end of the phrase, at 8'19", there is again a silence in both flute and tape at the same moment. If the flutist is out of synch at 8'09", but doesn't jump to 8'09", then the sudden silence will probably not happen; it is much more likely if the flutist does synch up with the tape at 8'09". These two places are the most difficult for crucial timing, and most of the difficulty arises from the long duration of fast notes with little opportunity to check the time.

Related to this, but subtly different, is the requirement that the flutist be in perfect synchronization with the tape throughout these fast sections — they are not only in rhythmic unison, but the flute and high tape part are also in melodic unison for most of these two fast sections. It is entirely possible that the flutist knows exactly where to be and what it should sound like, and still have difficulty executing perfect synchronization. The challenge is that there are between eight and ten notes per second, containing many leaps and an unpredictable series of notes. It is difficult to get every note correct, it is difficult to play them at exactly the right speed, and difficult to hear which note the computer is on in order to match it. In conventional music, keeping a steady tempo requires some effort, but there is a certain amount of room for error. In Exchange, there is extremely little room for error, whether the error is a wrong note, uneven finger technique, or slightly wrong tempo. Also because of the speed, recovering after an error is also difficult, because it's not easy to decide how many notes ahead to skip to become correct again or to be able to read ahead enough to do so. The solution, it seems, is for the flutist to listen to the recording many times in order to get a sense of the sound of the speed. Also, it will be helpful to practice without the tape but with the metronome, as well as to practice often with the tape and get a sense of whether the flute part is tending to be too slow or too quick. Because the markings indicate seconds, then quarter note = 60 is appropriate. Another tool may be to finger but not sound the notes along with the tape — this would enable the flutist to hear the tape pitches as they pass by, whereas the tape part will be somewhat obscured when the flutist actually plays the pitches.

In general, though, synchronization is not difficult. Because every second of the piece is indicated, then the flutist should not have trouble being within a second of synchronization at any given time. As noted above, the exception is when it's either physically extremely challenging or when comparing the time in the score against the stopwatch is impractical for a few moments. The piece has many moments in which the flute and tape play off each other in a constant, intimate dance; in these places, it's not difficult to synchronize even without the stopwatch because there are so many aural signposts to aid orientation. From 3'16" to 3'58", for example, each phrase in the flute corresponds to a phrase in the computer; the interplay makes synchronization fairly dependable. The last section, too, is not difficult. In this case, the pitch material as well as the silences in the computer orients the flutist. It is in this section that the flute and tape seem to truly come together, and there is a sense of resolution. The ease in synchronization reflects the easing of tension in the piece.

Approach to Practice and Performance

As with all electroacoustic pieces, the performer will probably spend some time practicing with the tape as well as without. With this piece, some of the practice time will also include using a stopwatch, to check the pacing and speed. Special practice time will need to be devoted to some of the extended techniques and to the high-speed and technically difficult passages. The practice methods for the former were discussed in the section concerning extended techniques.

My experience practicing this piece led me do develop some careful and thorough practice techniques, and I will attempt here to describe my process. In approaching most pieces, I would spend a great deal of time reading the score and thoroughly learning the material before ever attempting to play it. By the time I would actually pick up the flute, I would be pretty comfortable with the notes and rhythms, but not necessarily with the tempo. Then I would begin the process of practicing the piece, in a very conventional way, paying attention to the dynamics, articulations, etc., while working on getting the piece closer to the tempo at which I wished to perform it. For some especially difficult extended techniques, such as the less stable multiphonics, I would isolate these and practice them separately, incorporating them to the best of my ability as I practiced each section. This traditional sort of practice worked for me for a majority of the piece, until I reached the extremely fast technical passages.

When I first worked on extremely fast passages, I also practiced these in a conventional way. Let's take 6'44" to 6'57" for example. I knew that eventually I would need to reach a speed where each slash articulated a second of time, and could plan on putting the metronome eventually to a quarter note = 60. In this particular section, all of the beams were in groups of nine notes, so I decided to subdivide into three groups of three notes. The metronome speed could be subdivided into thirds also, giving me a goal tempo of 180 beats per second. I was able, in the beginning, to play the entire passage at about 120 beats per second, so I started there. If I was able to play the passage perfectly three times in a row, then I moved the metronome up a notch to the next setting. Gradually I worked up closer and closer to the desired tempo. Surprisingly, this only worked to a certain extent; I found that once I reached about 166 on the metronome, I simply could go no further without making mistakes. At first, I wondered if this was simply the limit of my capabilities, and would not be able to successfully play this passage up to tempo. In the end, I did find a solution, though; in fact, I was able to play the passage perfectly at far beyond the required tempo. Ultimately, I reached a metronome speed of 208 beats per second for the subdivisions, which would equal just over 69 beats per minute when not subdividing (faster than a second per slash).

What I had discovered was that my manner of playing posed physical limitations that I was unaware of until I had attempted this particular challenge. The solution lay in playing differently, by playing with extreme relaxation in the body, especially the fingers. I found that I had to start far slower than I had originally (in fact, at half of the speed of my original beginning tempo) and this time make sure that I was completely relaxed as well as completely confident before speeding up. It had been tension in my fingers and body, tension I was completely unaware of, that had caused binding and sluggishness in my playing. Once I had recognized this and addressed it, I was able to develop a much more fluid and relaxed technique, and ultimately play the phrase correctly. This, of course, improved my playing not just of this piece, but of everything I attempted thereafter.

While this technique worked for learning to play the notes correctly and up to the right tempo, it did not, unfortunately, ensure that I would be able to synchronize with the computer throughout the piece. I still found a great deal of difficulty synchronizing in the sections between 5'39" and 6'12" and 7'29" and 8'09". The synchronization here takes a great deal of good practice time to get just right, as well as intimate familiarity with the flute and tape parts both, so that asynchronization doesn't even begin to happen. Many problems of this sort can be solved simply by being extremely familiar with the tape part, to the point of recognizing individual pitches, which takes time and attention.