¶¡ÏãÔ°AV

touch'n'go, or toco y me voy, is a fifty-minute tape and text piece. The piece is divided into eleven sections of varying length, going from less than a minute to ten minutes. Each section explores simple, direct ideas expressed in short Spanish and English poetry or prose. Musically, each section uses a limited number of sources which are usually constrained to a single or very few sound classes. ³Ù´Ç³Ü³¦³ó’n’g´Ç is one of the first attempts to apply a consistent ecological approach in the context of a tape composition.

This section will succinctly describe the main concepts which inform ³Ù´Ç³Ü³¦³ó’n’g´Ç, its structural characteristics, the ideas presented in text format, and a few illustrative examples of the processes involved in the development of the sounds. The section concludes with a discussion of why ³Ù´Ç³Ü³¦³ó’n’g´Ç should be labeled as an ecologically-based composition.
 
 

The text

All the text of the piece is laid out in HTML (HyperText Markup Language) format (Keller, 1999). Each section has its self-contained text but all the text forms a network of interconnected ideas physically linked by hypertext anchors. The text consists of excerpts of Martín Fierro (Hernández, 1872), The Garden of the Forking Paths (Borges, 1962), two verses of a poem by Julio Cortázar (1966), and poetry and prose written by myself (Keller, 1999).

To summarize the ideas contained in the piece would be pointless, so we will describe its structural and thematic axis. The piece’s structure (or lack of it) is centered on the idea of the forking path (Borges, 1962). In other words, the piece is designed by the listener / reader at the moment he establishes a path through the material of the piece. The texts written for this piece make use of algorithm-inspired methods, such as progressive expansion (Coin a name) or progressive reduction of material (sCRATch), modular and randomly combined sections (Pandemonio), and a seven-element series (toco y me voy).

As discussed in (Keller, 1998a), ³Ù´Ç³Ü³¦³ó’n’g´Ç makes reference to the key problem of cross-culturalism in Latin American and Argentinean intellectual practices. Jorge Luis Borges and Julio Cortázar have dealt with cross-cultural issues in all their works. Going back to one of the pillars of Argentine culture, the traditional epic poem Martín Fierroby José Hernández addresses the interaction between indian, black and white culture and how it shapes the emergence of a new socio-cultural identity: the gaucho. A paradigmatic example of this social identity is Atahualpa Yupanqui. This philosopher, poet and composer has created a body of work which is assimilated into popular culture not by market dynamics, but by the force of its internal coherence and its direct reference to everyday life. As we have said in (Keller, 1998a) it is clear that if a concept of Argentine identity can be used at all, it is determined by the dynamics between local and immigrant cultures. In this sense, Argentinean identities are not one homogeneous, abstract entity, but several coexisting and ever-changing approaches to the world. This dynamics ensures a broad and flexible understanding of cross-cultural issues, and explains the traditional Latin American and Argentinean skeptical rejection of fundamentalist and pseudo-"universal" theories.
 
 

The music

Although each of the sections of ³Ù´Ç³Ü³¦³ó’n’g´Ç is a complete, self-contained piece, most of them share material and provide sound references to each other. Thus, the piece can also be divided into four sections of ten to eighteen minutes that comfortably fit a concert setting. Besides using common sound sources, most pieces were created with similar sound processing tools.

All the material for ³Ù´Ç³Ü³¦³ó’n’g´Ç was produced with ecological models (Keller & Truax, 1998) implemented in Csound (Vercoe, 1993), CMask (Bartetzky, 1997), and the AC Toolbox (Berg, 1998), and with real-time asynchronous granular synthesis using MacPOD (Keller & Rolfe, 1998). MacPOD was developed by Chris Rolfe following the POD model documented in Truax (1988). MacPOD was mostly used for time-stretching and real-time gestural transformation of material. Its intuitive and responsive interface allows for transformations which are cumbersome or even impossible to do in batch mode. An example of the effectiveness of this sound processing software is provided in the voice material in let me see . . . how can I word it? The exaggerated inflections in Yupanqui’s voice were created by changing the stretch ratio parameter at the exact beginning of consonants and vowels. Thus producing inflections such as ‘muuuuuy, ’ meaning ‘very’ in Spanish, or ‘profffffundo’ which means ‘deep’ and lasts for approximately a minute.

Direct sound references (Shepherd, 1992) and gestures (Smalley, 1993) which are directly related to the meaning of each section are used throughout the piece. sCRATch uses scraping as its basic sound model. spill, spiel, spoil uses spilling and pouring. Coin a name employs splashing and water sounds. Action to be taken in the event of a fire uses the sound of a lit match as its only source material, and different models of fire. Realpolitik shares some elements with Action, and introduces cracking woods, shattered glass, and explosion-like sounds. A waltz in a ball uses only a sample of a rolling marble.

Two structural processes are exploited in ³Ù´Ç³Ü³¦³ó’n’g´Ç: emergent macro-structural properties by interaction of lower-level elements, and isomorphism. Farewell, welfare provides an example of isomorphous processes at a conceptual, algorithmic, and perceptual level. The piece develops the idea of the infinite labyrinth: a straight line traveled by covering half its length every time (Borges, 1962; Kampis, 1991). The algorithm used for producing all sounds in this piece is a variation of the Karplus-Strong (1983) model and uses a single short sample as source material. This sample is ‘walked’ by two pointers at different sample rates producing an ever-rising or ever-falling sound, depending on the parameters fed to the model. The sound is akin to a combination of Shepard tones with self-similar events that occur at ever-expanding time-spans.

The model developed for Vox Populi uses a small sample pool of conch shell sounds. Being harmonic with some noise content, these samples when combined randomly approximate the behavior of choir-like formants (Keller & Rolfe, 1998; Rodet, 1984; Truax, 1992). These formants are the result of the interaction of granular samples at a meso level. So the effect at a macro level is qualitatively different from the source sound characteristics.

A similar phenomenon can be observed in the ‘structured rain’ material used in least, but not last. This sound was generated by a three-stage process. First, several different drop sounds were produced by convolution. Then these drops were organized as constrained random meso-level events. Finally, the events were combined in a slowly evolving macro-structure, which distributed the meso-events in a slightly irregular rhythmic pattern. The result is a sound that resembles wind and rain with metallic reverberations in the background. least, but not last also makes use of two contrasting spaces - metaphorically speaking, the space of the living and the space of the dead. The first space was produced by mixing a recording of a big, open building space. The other space was realized by applying convolution to the grain samples.

As can be inferred from our theoretical discussion, cultural references are not absent from ³Ù´Ç³Ü³¦³ó’n’g´Ç. The most obvious and probably most meaningful one is the use of the voice of the great poet and philosopher Atahualpa Yupanqui (1980). Although his voice is kept recognizable, we have taken care of always processing it to heighten its presence and to give it a deeper timbral quality. Sound transformations are tightly related to the meaning of the text and its metaphoric implications. The first verses can be translated thus: during those times such things happen that do not happen anymore. The atmosphere of story-telling set by the verses is reinforced by an outdoor background sound which frames all sound events in the first section. As previously mentioned, the word ‘deep’ in the verses: if the river is wide and deep, that one who swims well gets across, is time-stretched to last approximately a minute. The same verses are used to produce a hybrid sound by means of Linear Predictive Coding. This technique lets us filter Yupanqui’s voice through patterns produced by synthesized bubbles. The effect is akin to whispering inside water.

Pandemonium incorporates cultural references in its sample materials and in its algorithmic models. Brazilian percussive instruments, such as cuica and pandeiro, were transformed by ecological models to provide a continuum of instrumental and environmental sounds. The meso time-patterns employed closely approximate traditional Brazilian rhythms without losing an overall ‘careless’ and ‘aimless’ feel. The sound palette is further enhanced by introducing hybrid sounds such as the breaking-glass cuica, or the broken pandeiro. These sounds were generated by applying breaking time-patterns to instrumental sampled grains.
 
 

The format

³Ù´Ç³Ü³¦³ó’n’g´Ç proposes a system that requires a flexible and modular format. The ideology underlying the piece is expressed in its thematic axis: the gaucho or the traveler, in its structure: a path established by the listener / reader, and in its sound material: direct references to a social and environmental context. The ideas developed in the text are related to (and match one-to-one) the sections of the tape piece. Nevertheless, they are given to the reader as a network of short excerpts that are freely combined. This same process is mirrored in the music.

Once I had established a rough plan for ³Ù´Ç³Ü³¦³ó’n’g´Ç, a pressing question had to be answered: what format could comprise sound and text in a modular, readily accessible layout? Both Enhanced CD and CD-ROM seemed to fulfill these requirements. I chose to present two versions of the piece: a data CD-ROM for the Macintosh platform and a commercially released Enhanced CD, playable on any CD player, Macintosh or IBM-PC computer. The text material of the piece, including the Csound code, is freely available on my website (Keller, 1999). I only refer to the data CD-ROM in this documentation.

The random-access structure of the data CD-ROM allows the user to freely peruse the text in his browser and simultaneously listen to any section of the music. This has two implications on the way the piece is developed and perceived. First, the material cannot be structured in a linear, one-dimensional fashion because in any case this layout can be overridden by an active user. Second, the piece requires a thoughtful involvement from the user: clicking randomly and arbitrarily cutting sections becomes boring and uninteresting. (Of course, following the spirit of the piece we establish no rules or restrictions whatsoever). Given this context, the user / listener / reader is asked either to leave his passive role and engage in an active, conscious, creative process, or not to engage at all.
 
 

After presenting the basic methods and concepts behind ecological modeling and composition and after providing an overview of the issues addressed by ³Ù´Ç³Ü³¦³ó’n’g´Ç, it seems wise to restate the elements that characterize ³Ù´Ç³Ü³¦³ó’n’g´Ç as an ecologically-inspired piece.

The material used in ³Ù´Ç³Ü³¦³ó’n’g´Ç makes direct reference to everyday and environmental events. This material also incorporates cultural elements that place the piece within a very specific and distinct cultural context: Latin America. Most sound processing is done through models that apply ecologically feasible transformations on the material. Transformations that produce abstract material are used sparingly. They are usually introduced by means of simple compositional processes, e.g., expansion of the number of elements, time-stretching, hybridization, resonant filtering with uniform spectral characteristics. Sound spaces are kept consistent and stable, i.e., there are no spatial changes without appropriate cues and transformations in the related sound events.

The sound unit used throughout the piece is the event. In other words, there is not a single sound that has a suddenly cut decay, or an attack without transients. Most sounds were produced by the interaction of exciting systems - granular algorithms - with resonant systems - generic physical models. At a meso level, all algorithms follow ecologically feasible patterns, that is, they produce events that occur or may occur in our terrestrial environment. At a macro level, the piece explores the emergent properties resulting from the interaction of complex elements at a micro and meso level. The format chosen for the piece requires the active participation of a conscious user. In other words, the listener / reader engages in a process of mutual-determination (Varela et al., 1989; Keller, 1998a) with the material of the piece. Thus, the general model of environment-individual interaction is scaled down to the artwork-individual pattern formation process. Indeed, if ³Ù´Ç³Ü³¦³ó’n’g´Ç is not an ecologically-based piece, it is nothing at all.

 

Title

Abstract

Introduction

EcoModels

EcoComp

EcoMethods

CaseStudy

References