This paper aims to survey the affordances and limitations of a growing body of audio scores which employ sound recordings as the primary means of communication between composer and performer. Here we follow James J. Gibson’s (1979, p.127) notion of affordances as the potential actions made possible by an object or environment to a given individual—a concept that implies a mutually influencing, transactional relation between actor and object. A material format alone does not wholly determine the action possibilities it affords; composers and performers (i.e.
the actors in this context), as well as audio scores (i.e. the object), are themselves situated and dynamically shaped within wider networks and histories of cultural practice.
These practices mediate and constrain potential relations: for instance, the act of deploying a sound recording as if it were a score suggests a translation of prior scoring practices across media; equally, functioning as a score is just one of the many potential use cases afforded by sound recordings. Nevertheless, at this particular intersection of cultural practice and material format, we contend that audio scores representing information and instructions in sound afford some distinct and different possibilities to composers and performers when compared to scores which deploy some form of symbolic visual representation of sound or sound-producing movement. Via the contrasting works we have taken as case studies, we intend to offer a preliminary sketch of some of the affordances enacted thus far in this dynamic, in process contemporary field. Within this field, we identify two primary sub-categories associated with the temporal relations composed between performer and audio score: reactive and rehearsed.
On the one hand, performers primarily react to the audio score during performance; on the other, the audio score shapes the performers’ interpretations in rehearsal, well before public performance. Louis d’Heudieres’ Laughter Studies 1-3 for two vocalising performers (2015-16) are examples from our reactive category that translate a practice associated with visual scores into a different format to enact distinct affordances. In d’Heudieres’ series, each performer listens to a different audio score on headphones and alternates between vocally imitating and verbally describing what they hear to the audience. The audio scores feature a collage-like succession of many and diverse samples, which are edited and processed to varying degrees. The performers are asked not to overly familiarize themselves with their audio scores beforehand; rather, they should spontaneously react to the often unpredictable changes and transformations between sounds. This arbitrarily-imposed constraint is a kind of translation of sight-reading practices, though it differs in some fundamental respects.
Static visual scores tend to represent sounds in a metaphorical spatial configuration (Cook, 2013, p.23) allowing performers to anticipate and read ahead of their performed actions at any given moment. In addition, a performer may infer the likelihood of future events in a given visual score based upon characteristic features of the dimensions prioritized by the composer—harmonic formulas, instrumental register and techniques, rhythmic complexity, etc.—these all serve to situate the score with reference to similar examples from similar historical periods or genres. Given their past experience, performers may thereby select out an array of unlikely occurrences to facilitate their sight-reading. By contrast, d’Heudieres’ Laughter Studies lock the performers into a relation with the audio score that is always reactive—the format ensures there is little way of hearing or inferring ahead. d’Heudieres couples this limitation with an affordance of sound reproduction technologies: the format mediates and renders commensurable a vast array of sound sources—recorded, synthesized and processed etc.
Potentially any sound within the range of human hearing and the frequency response of the microphones, speakers or headphones, and other associated technologies could be reproduced, manipulated and organized in countless ways. As if to take full advantage of this affordance to composers, many of the sounds in the Laughter Studies audio scores are crowdsourced: culled from freesound.org, a collaborative online database of Creative Commons Licensed sounds. The resulting high degree of unpredictability and diversity in the sounds deployed by d’Heudieres at an often fast rate of change affords varied modes of performer listening, imitation and description: highly affective sounds, such as laughter or crying, invite imitations and descriptions emphasizing semantic associations; by contrast, more abstract sounds or processing invite greater attention to acoustic qualities.However, in contrast to many visual scores featuring transcription, d’Heudieres does not explicitly transcribe and prioritize certain dimensions of his chosen sounds beforehand and on behalf of the performers based on some generalized conception of the capabilities of a human voice; d’Heudieres gives no guidance to performers on how or what they should do to accomplish their tasks, and few concessions are made to idiomatic vocal conventions in the choice of sounds—many are simply impossible to accurately imitate.
Rather, the particular intersections of translated sight-reading practices and audio score format we have enumerated afford d’Heudieres the possibility of recomposing the interpretive encounter between performers and score—the vocalists react to the audio scores, transcribing and mediating freely in their own, individual way during public performance. The ease or difficulty of the task over time for those specific performers are dimensions of compositional exploration, traceable via the variable fidelity of their imitations and descriptions, as well as reinforced by their stumbles and hesitations—sonic and gestural. Overall, the performers’ personas convey no pretensions toward accomplished mastery of their materials, inspired interpretation, or faithful reproduction. Rather the affordances enacted in the relations between the performers and this reactive audio score evoke the register of candid, improvisatory play and games. In a similarly playful vein, not only does the performer in Lara Stanic’s Open Air Bach (2005/2013) react to an audio score during performance, but this audio score reacts dynamically to the performer’s actions as well. Stanic enacts the affordances of the varied assemblage of technologies at her disposal to compose a feedback loop wherein performer, computer, microphones, and loudspeakers mutually influence one another, interacting via live processing of recorded sound.
Three external microphones follow the amplitude of the output of three loudspeakers attached to the performer’s body. The amplitudes picked up by each microphone determine the speed and pitch of the live, computer processed playback of each part of an instrumental recording—the third movement of J.S. Bach’s Sonata in E minor for flute, cello and harpsichord BWV 1034. The performer’s goal is to achieve the ‘correct’ playback of the recording through silent movement and somewhat awkward gestures.
The closer the performer moves to the microphones, the greater the amplitude of the signal they pick up, and the more stable and accurate the playback becomes; however, the performer must constantly adjust their distance from each microphone due to the continually changing volumes of the recordings (Cathy van Eck, 2017, p.164-165). Generally static formats and their associated practices, such as conventional symbolic visual scores or d’Heudieres’ audio scores, afford the performer varying types and degrees of interpretative flexibility centered upon a largely stabilized, if relationally defined object. Many of the score’s details and their situated meanings are negotiated before the act of public performance. By contrast, Stanic composes a situation afforded by live electronic sound processing where the score-object itself is highly mutable and fluid during performance. Stanic defines an interactive system, which dynamically responds to and scripts the performer’s movements. However, since the audience also hears these processed, fluctuating sounds, this playback becomes an object of aesthetic interest as well.
Thus far, our case studies have surveyed a reactive engagement between performer and audio score. By contrast, the reed quintet Zugvögel by Carola Bauckholt invites players to interact with and rehearse the audio score well before performance. The audio recordings themselves, which consist of bird calls of various species, are not played in performance; instead, the quintet members are instructed to familiarize themselves with these bird calls and memorize all their nuances in order to reproduce them on their respective instruments. Bauckholt’s use of the audio score format directs this activity towards high-fidelity transcriptions of bird calls rather than reduced abstraction or overt musicalization (the obvious historical precedent is for the latter occurs in the works of Olivier Messiaen) (O’Callaghan, 2012). By using recordings of her sound sources, Bauckholt’s audio score affords of a higher degree of specificity and dimensionality to performers than most visual, symbolic representation of those sources (particularly in regards to spectro-temporal variation in timbre, for example.) In Zugvögel, each performer establishes a one-to-one relationship with the recordings given the implied objective of producing a faithful imitation through whatever means necessary. The mimetic process undertaken is therefore accompanied by the task of parsing each recording’s most salient elements. This parsing bears a further affordance to the performers in that it allows their activation of the specific capabilities and personal knowledge of each instrument.
In contrast to the traditional practice of bird call transcription in a conventionally-notated medium, the recorded format omits much of the symbolic filtering and prioritizations of the composer. It instead defers any such filtering to the instrumentalists of the reed quintet, allowing a more intimate understanding of the instruments’ capabilities to inform a precise rendering. Hence, the performers’ personal knowledge of their instrument’s compatibility with the source material affords a higher degree of fidelity in its reproduction. Through this process, the performer may also come to discover previously hidden action-potentials in relation to their instrument.It should be noted that each recording is visually transcribed by Bauckholt in the score, aiding the performers with precise depictions of pitch, rhythm, timbre, and dynamics. She adds in her prefatory notes, however, that “the notation should only be taken as a guide,” suggesting that the notated transcription holds only a supplementary role in relation to the audio score. This supplementation might be employed in order to address one of the format’s primary limitations: the act of memorization. Because the minute features of each recording must be encoded into memory, the musical information itself is subject to variability and even corruption over time.
Factors influencing the performer’s ability to reproduce the recording may include: the time spent rehearsing, the effort expended in memorizing the recording, the recognizability (or lack thereof) of each sound, the performer’s aptitude or preferences towards certain sounds, and the performer’s assumed ensemble role. Bauckholt’s utilization of recorded materials in conjunction with a traditionally-notated score seems to redress these limitations effectively.Guide for 10 singers by Cassandra Miller offers a similar implementation of the rehearsed audio score in emphasizing the specificity of its recorded materials. Though where Bauckholt instructs the performers to imitate the sound sources as closely as possible, Miller instead opts for a qualitative embodiment of the audio score.
To achieve this, the players are directed to become thoroughly familiarized with a recording of “Guide me, O Thou great Jehovah” as interpreted by American folk singer Maria Muldaur in 1968. The particular recording was selected, according to Miller, because its melody exhibits QUOTE “swoops over a large tessitura that, more than anything, sounds like it feels good to sing” ENDQUOTE (Miller, 2013). As in Zugvögel, the written score to Guide serves only a supplementary function; Miller employs a “quasi-neumatic” open graphic notation, providing only the starting pitch, tempo, and general contour of each vocal line in order to support a flexibility of interpretation.
An insistence on “singability” is prioritized, emphasizing the free-flowing, improvisatory characteristics of Muldaur’s original interpretation. These characteristics need not be translated through a direct imitation of the recording by rendering all its time-based nuances; instead, the audio score affords the qualitative embodiment of Muldaur’s vocal characteristics. In this way, the public performance of the piece depends upon the treatment of the audio score as an embodied entity.
The identity of Muldaur as a vocalist is appropriated and filtered through the bodies of each singer; the audio score affords access to the minute sonic details of Muldaur’s registration, phonation, and articulation. In her instructions, Miller references the score’s resemblance to oral tradition; the composer acknowledges this larger cultural practice in Guide, one in which the material is shaped by the corporeal identity of the interpreter over multiple generations and iterations (Fischer-Lichte, 2008, pg.89).
The performative liberty granted to the singers allows this phenomenon to be communicated in an arguably more acute, concrete way than a symbolic transcription. The two primary sub-categories of audio scores we have outlined, reactive and rehearsed, serve well as an initial distinction to guide our survey; however, these categories may be combined, weighted, and hybridized to varying degrees. In Carolyn Chen’s Adagio (2009) three or four performers listen to an audio score on headphones, synchronizing emotive facial expressions with the music—specifically, an excerpt of the second movement from Bruckner’s 7th Symphony as performed by the Münchner Philharmoniker, conducted by Sergiu Celibidache live in 1994.
The distinct affordance this work enacts via its associated entities is the encapsulation of a chain of translations of specific media and socio-cultural phenomena. First, a recording encapsulates Celibidache’s interpretation of Bruckner’s visual score and its associated practices. Then, an excerpt of this specific recording is subsequently appropriated by Chen as an audio score. Bruckner’s visual score also acts as an adjunct to this audio score, annotated with affective descriptors that presumably translate and embed Chen’s own listening practices and relations with the excerpt. The performers rehearse and memorize descriptors such as “elevated,” “yearning,” and “solitary”—suggesting, but not prescribing the exact facial expressions to perform. The audio score affords not only a time-based structure for these performative events (akin to a click track), but a live stimulus which the performers react to, mediating Chen’s affective instructions. This ‘private’ listening practice is then transposed into a concert setting.
The audience hears no sound; they voyeuristically observe the staging of a ‘private’ listening experience they know well, speculating upon the incomplete trace of the aforementioned chain of mediations. At the very least, the performers’ exaggerated facial gestures clue the audience into the ‘romantic’ qualities of the unheard musical referent.Michael Baldwin’s Buzzed (2015) for solo French horn similarly hybridizes the reactive and rehearsed categories of audio scores. Baldwin’s use of multi-tracking affords the possibility of concretely encoding and layering the documented history of his collaborative exchange with the performer in the audio score. Specific sound events trigger the performers’ memories of collaboratively developed, rehearsed, and embodied moments, both physical and sonic. The audio score is arranged in a Digital Audio Workstation—this affords the possibility of altering the speed of playback, muting layers, or looping sections to facilitate the rehearsal of this often complex score (Mercé Bosch-Sanfelix, 2016, passim.
). During public performance, the horn player listens and reacts to the audio score. Baldwin notes that this use of the format is constrained by the performer’s perceptual limits (Mercé Bosch-Sanfelix, 2016, pp.5-6); auditory events within the score may be confused with one another when they share similar acoustic qualities.
Equally, certain sounds produced by the performer may mask or obscure sounds in the audio score. Unlike their visual counterparts, audio scores often represent information and instructions in the same domain as the performed product. As such, many audio scores afford the imitation and embodiment of diverse sound sources with varying degrees of fidelity. Unlike many visual scores, a composers’ conception of a voice or instruments’ capabilities need not determine the fidelity of the performers’ realization—a somewhat standardized, generalized vocal or instrumental approach is deemphasized in favor of an individual musicians’ knowledge of the action-potentials of their specific instrument (as in d’Heudieres’ Laughter Studies, Bauckholt’s Zugvögel, and Miller’s Guide). Distinct affordances are enacted through the intersection of prior cultural practices and the audio score format: in Laughter Studies, sight reading practices are translated into a format that limits performer’s abilities to listen or infer ahead; a mediated oral tradition in Guide affords a qualitative embodiment of the recorded material’s specific characteristics; Chen’s use of Celibidache’s recording in Adagio affords the encapsulation of successive mediations of Bruckner’s score, etc.
In addition, digital manipulation of recorded sound affords a mutable score-object as in Stanic’s Open Air Bach, and during rehearsal in Baldwin’s Buzzed.Our survey of the affordances enacted in recent audio scores point towards further avenues of exploration for creative practice. Chen’s and Balwdin’s pieces constitute an initial step towards possible hybrid approaches to the relations between audio score and performer; cross-modal interactions and mappings could be extended further; within a single piece certain performers might engage with visual scores and others with audio scores enacting distinct affordances; varying degrees of prescription or flexibility in the interpretation of audio scores might be explored; the audio score could be at times mutable via live processing and at others largely fixed; the score itself could be revealed as an object of aesthetic interest over time—it might be alternately diffused over speakers in different spatial configurations including headphones; various modes of performer listening might be invoked in a single score; distinct performative and material limitations may be considered as compositional prompts; finally, these diverse approaches and others might be applied asymmetrically across individual members of a given ensemble or at different times within one piece. In all of these cases—speculative as well as realized—representations of information and instructions in sound afford the possibility of fundamentally recomposing the interpretive encounter between performer and score.