The text is a complement to two music installations, composed with the intent that their existence only be discoverable by an audience and not dependent upon an a priori expectation or understanding of music. The first chapter outlines a theoretical framework for discoverable composition incorporating ideas about the integration of art and life, music as dialectic between behavior and environment, and perforative context as a fluid and variable system. The second chapter presents historical examples of music/art installations and performance practices, which are relevant to the idea of audience discoverability. Finally, the third chapter discusses the conception, implementation, and evaluation of the music installations Only if you're there (I'll meet you there) and I think I know you (you think I don't). Written thesis can be found here: http://eamusic.dartmouth.edu/~sean/Homepage/Thesis/DiscoverableComposition.pdf

Transgressive Interactive Art is artwork which both violates the boundaries of its audience and functions via their participation. Transgression, interactivity and art are defined and analyzed. Four transgressive, interactive pieces produced by Charles DeTar in the course of his master's studies at Dartmouth College are discussed. Other examples of twentieth century transgressive interactive art are discussed in relation to DeTar's works. Written thesis and documentation of some of the artworks can be found here: http://tirl.org/installations/transgressive-interactive-art/

This text is a complement to the musical composition "The Spell of Speech," a piece for actress, electro-acoustic sounds and live electronics composed in 2004 as part of the requirements to conclude the Master's degree in Electro-Acoustic Music at Dartmouth College. The core of this research is the relationship between speech and music composition. The first chapter presents a theoretical framework comprising evolutionary musicology and music cognition. The second chapter contains a historical survey of electro-acoustic music in which human voice plays a fundamental role. Finally, the third chapter analyzes the composition of The Spell of Speech, presenting the main techniques used and the musical ideas involved. Written thesis and concert recording can be found here: http://eamusic.dartmouth.edu/~bruno/comp_tsos.html

This paper documents the research and development of three software programs, all of which were based on a single application design. The design was intended to provide a framework for exploring the use of algorithmic composition techniques within the context of a film score. In practice, changes in the intended audience of the application and the musical goals of specific projects led to different implementations, each of which in some way was in keeping with the spirit of the original design. The reasoning behind each implementation is discussed in detail.

The goal of this thesis is to study and compose electro-acoustic music from the perspective of manipulating spectra. We first discuss in historical context tech- nological aspects of numerous electro-acoustic compositions whose main focus is a control of timbre through spectral manipulation. We next discuss my own software (a collection of Max/MSP external objects) which employs frequency domain spectral processing of sound. This software is based on the analy- sis/synthesis techniques implemented in the phase vocoder and it extends those capabilities to more unconventional signal processing for experimental sound design. Finally my electro-acoustic compositional work focused on spectral pro- cessing is discussed.

This work employs physical modeling techniques to synthesize the sound of a Central Javanese rebab, a bowed chordophone used in the performance of gamelan music. The rebab model employs a bowed string physical model and a two dimensional mesh model for the rebab body. The model components are adjusted to the physical specifications of the rebab based on the physical and acoustic characteristics of the instrument. The synthesized rebab is compared to an actual rebab to measure the degree of success of the model. The thesis text and matlab code for the model can be found at http://eamusic.dartmouth.edu/~stefan/RebabModel.

A working definition of sonification, together with its history, purpose, usefulness and relationship to musical composition are presented. Assessment of the potential of sonification is made based on music cognition and brain science research. Guidelines for musical sonification design are suggested, as informed by analogy with graphical design, music cognition and musical composition. Several case studies are provided, both of musical compositions which draw heavily on the sonification of data as a working procedure, and of sonifications of comple numerical processes, historical meteorological data and financial data which are designed to provide insight into these data relationships, but are not necessarily musical compositions. As an exercise, musical compositions were prepared from all of these sonification projects.

Machinic Actualization of Musical Immanence is a composed text. A text composed by the writer-composer for the reader-composer - on composition: a text by the composer for the composer on composition. What, then, is composition? Composition is creation. Composition is the coming into being of the new, the bringing into being of the ontologically new by the creator-composer. Composition is actualization. The creation of the new takes place through the reciprocally generative and transformative process through which the intensive potentiality of the virtual is unleashed and transposed into actual composition-creations - possible and real. Composition as actualization is poorly understood. This is because it cannot be explained within the classical form of representation, which is grounded on logic and causality. The problem is that representation presupposes a historical causality that is incompatible with - and cannot be explained within - the intensive potentiality of the virtual, and ultimately produces the illusion of knowing the unknown. The virtual is fully determined, but not predictable. It can only be fixated throughout the creative act of composition-actualization. We propose a new model in which composition-creation is defined as the product of a machine. More specifically, the compositional machine creates through actualization and functions as an intensive, asymmetric relation between virtuality and actuality, potentiality and discharge, content and expression. Our machine will be ontologically defined along dimensions of determinability and predictability, and explained in terms of modes of openness, complexity, randomness, complicatedness, chaos, and linearity. At large, this text is a description of the operational space of the compositional machine. In this sense, its nature is geographical-descriptive as well as philosophical-prescriptive. But first and foremost, its nature is musical and aesthetic. A text succeeds only to the degree that it, always, is unfinished. It succeeds merely to the extent to which it can be used, to which it can serve, as a repository for the creation of more - theory and music. Just as we have used the writings of others - philosophers and composers - and placed our text in the middle of theirs, we hope that others can use our writing in a similar fashion. If nothing else, our text will be used by ourselves; placed in the middle of our work. Certainly it is in an experimental state. Consider it an egg, from which to grow - ideas and music, both for which new theory is needed.

Animals is a collection of musical compositions realized using computer and audio recordings of animal sounds. The accompanying text is in three parts. The first section is an essay on bestiaries. The second section concerns my view of composing. The third section is a discussion on Animals.

The Spectral Modeling Toolbox is a collection of functions for digitally analyzing and synthesizing sound. The techniques in the Toolbox generalize the concepts of other analysis/synthesis systems in an environment created for algorithm design and research. The Spectral Modeling Toolbox is an introduction to analysis/synthesis techniques, a spectral processing tool, and perhaps, the foundation for future sound recognition systems. The thesis and MATLAB source code is available at http://eamusic.dartmouth.edu/~kimo

The goal of this thesis was to implement a tool that, given a digital audio input, can extract and represent rhythm and musical time. The purpose of the tool is to help develop better models of rhythm for real-time computer based performance and composition. This analysis tool, Riddim, uses Independent Subspace Analysis (ISA) and a robust onset detection scheme to separate and detect salient rhythmic and timing information from different sonic sources within the input. This information is then represented in a format that can be used by a variety of algorithms that interpret timing information to infer rhythmic and musical structure. A secondary objective of this work is a "proof of concept" as a non-real-time rhythm analysis system based on ISA. This is a necessary step since ultimately it is desirable to incorporate this functionality in a real-time plug-in for live performance and improvisation. here's a direct link to all the goodies ... http://eamusic.dartmouth.edu/~iroro/thesis/

This paper is a supplement to my thesis composition, Bits and Pieces: A Sonic Installation for the World Wide Web. Bist and Pieces searches the World Wide Web for sound files, downloads them, processes them, and broadcasts the results back over the Web. The paper is divided into three chapters. The first discusses the technical issues involved in the creation of the piece, and the various programs that drive the search, download, and sound production processes. The second chapter reviews historical issues relating to collage and sample-based art, both aural and visual. The third chapter covers the aesthetics of this installation with respect to other current pieces of Web-based art.

The Polyrhythmic Modulation Studies is a set of musical compositions that explores the use of multiple accelerating and decelerating voices. These voices often change tempo simultaneously, in effect, modulating between different polyrhythmic relationships, or polyrhythms. Some of these studies are composed for computer-generated sound, some for live performers, and some for live performer(s) and computer-generated sound. This paper will first look at a few composers who have used tempo in similar ways. The focus will then shift to descriptions of my own Polyrhythmic Modulation Studies.

This thesis describes and demonstrates new techniques using the correlation dimension to represent, analyze, and resynthesize audio signals. Most signal analysis techniques focus on the time and/or frequency domains. These techniques are not well-suited for representing what might be called the complexity of a signal. Using methods developed recently in chaos theory, this paper describes a technique for measuring audio signal complexity. The correlation dimension is used in physics to approximate numerically the dimension of a graph produced by experimental data (Grassberger and Procaccia, 1983). Knowing this dimension is important in understanding the nature of the system described by the graph. Data from a periodic system has a correlation dimension of 1, as does random data. Chaotic data has a correlation dimension that is non-integer and greater than 1. From a musical point of view, the correlation dimension of a signal gives information about its periodicity, harmonic complexity, and noisiness. The correlation dimension of an audio signal is useful as a simple and acoustically meaningful metric for timbre. This thesis will serve as a tutorial for computing the correlation dimension of an audio signal and for understanding how to use this value to measure, order, alter, and resynthesize timbres according to their complexity.

This thesis explores some aspects of the relationship between traditional electroacoustic music and underground music culture. Several of my own projects are described which, in addition to standing as works of art, also represent research into the question of the artistic use of fragments of previously composed and commercially released music. The possible intersection between electroacoustic music and underground music communities is established and documented by the project, Institutional Collaborative, undertaken with Terre Thaemlitz. The distribution of experimental electronic music is discussed in relation to the record label I founded, Illegal Art, including the release Deconstructing Beck. Finally, alternative forms of media are explored as a means through which electroacoustic music might reach a larger audience.

This essay is a supplement to my thesis composition for four-channel tape entitled Hydraulis. The essay presents and discusses two issues relevant to the composition: quotation and early music. Chapter One reviews recent scholarship regarding the role played by the quotation of pre-existent music in the creation of new compositions. Chapter Two discusses the so-called early music revival. It also details the use of early music in several key electro-acoustic compositions by other composers. Finally, Chapter Three discusses my own use of early music quotation in several compositions, including Hydraulis.

For a mere moment... is a thematic electro-acoustic composition that may be performed with the aid of staging and projected/displayed video. It traces a person's life from creation to a state of waiting. It borrows heavily from certain Judeo-Christian biblical texts in the form of commentary, spoken phrases of text, and individual words. It is a composition for live performer(s), multitrack tape, and video. The composition relies on mixing sound files in a quadraphonic to hexaphonic space. Pre-production of sound files includes several signal processing techniques, multitrack mixing, score scripting, object oriented programming, and sequencing. The piece was composed for a performance in Brace Commons, Dartmouth College. The piece was realized on May 22, 1997 and included multitrack tape and one performer.

Cookie is a Premiere plug-in that implements an IIR (Infinite-Impulse Response) digital audio filter, with 8 zeroes and 8 poles, both time-varying. The pole and zero locations are input by the user in polar form, and drawn accordingly on a z-plane graphical representation. Cookie is designed to be a useful tool for electroacoustic music composers and sound designers.

The thesis is in two parts: a text-sound composition on tape based on a fragment of a poem by Mikhail Lermontov entitled The Daemon. Part two is an essay about the composition. The tape part is based exclusively on recordings of a female voice, both reading the poem in Russian and singing several short non-verbal melodies. The essay includes an analysis of the text and its relation to the music of The Daemon, as well as a description of sound processing techniques applied to the voice. The overall structure of the piece is based on the structure of the poetic fragment. The Daemon is in three major parts with inner subdivisions.

The following is a supplement to the electro-acoustic composition crow's-eye view no. The first part explores the main elements of this composition: the haegum and its timbre and Yi Sang's text. The second part outlines the synthesis and filtering techniques which were employed. The last part describes the compositional processes and also addresses some of the issues pertaining to the transcultural nature of the composition.

The thesis is the electro-acoustic composition Mississippi Menagerie. The composition explores combinations of acoustic instruments (mandolins and alto flute), with process variants of these instruments, and envelops its audience in a virtual sound space. Through a quadraphonic speaker configuration during performance, the audience is exposed to musical sounds which appear not to come from the speakers themselves but rather from arbitrary (composer specified) positions from within the concert hall. Furthermore, the movement of these sounds within the virtual space possesses an innate physicality through the simulation of distance dependent amplitude attenuation, appropriate Doppler shifting of frequencies, and other natural acoustic phenomena. The spatial, timbral, and melodic relationships between the performers' music and the quadraphonic tape reinforce the programmatic implications of the piece's title. It was given its premiere performance by Alex Ogle and Larry Polansky during the Festival of New Musics held in Spaulding Auditorium at Dartmouth College on April 27, 1996.

Videosymphony No. 1 is an audio-visual composition that combines electro-acoustic music with natural images and computer graphics. The composition belongs to a genre which I define as audio-visual music. The medium on which the composition exists is video tape. Attached to the composition is an essay containing three chapters and an appendix. In the first chapter, the technological development of audio-visual tools is described. The way in which these developments led to various forms of visual-musical genres is also discussed. In the second chapter, some of the perceptual issues are examined, especially those in the aural and visual domains. Chapter three describes the aesthetic considerations involved in the composition of audio-visual music. The appendix contains the scenario of the composition with the sequence of musical and visual events as they exist on the tape. See abstract and stills

This thesis documents original applications of wavelet theory to digital audio signal processing for electro-acoustic musical purposes. A non-mathematical introduction to wavelets for the musician is given first, along with a brief description of previous work in wavelet theory and audio signal processing. A mathematical summary of multiresolution analysis, the heart of wavelet theory, is then presented. Three original applications of selected theoretical constructs in wavelet theory are then proposed and implemented. Specifically, the concepts of second-generation wavelet shrinkage, exponential decay of wavelet coefficients, and wideband-octave frequency decomposition are used to construct algorithms which denoise, spectrally enhance, and wideband-equalize digital audio, respectively. A secondary goal of this thesis is to provide both the end-user and the C++ programmer with documented, publicly available tools that accomplish these audio-related, wavelet-based tasks. To this end, the MATLAB and GNU C++ environments are used to prototype and implement the three algorithms mentioned above on a Silicon Graphics Indy workstation. Audio examples of the output of these programs are presented along with each algorithm. Appendices document the development of the MATLAB code and C++ libraries which were used for these experiments. C++ classes for audio file (AIFF file format) I/O, interpolating and non-interpolating delay-line processing, real-time digital filtering (simple time-domain convolution), and real-time audio processing are discussed. In addition, a failed but instructional attempt at a C++ class that models a proposed wavelet file format (WAIFF-Wavelet Audio Interchange File Format) is discussed.

The thesis is in two parts: a composition for saxophone, bass clarinet, bassoon, tape and electronics processing entitled Conversations, and an essay. The essay includes aesthetic concerns about Conversations along with a consideration of its sound sources, processing techniques and structure. The second chapter of the essay concerns pieces for clarinet and electronics written by other composers. The appendix of the essay discusses the physics of clarinet multiphonics. The tape part of the composition is based largely on computer processing of recorded sounds of the instruments involved in the live performance.

This thesis documents work in electro-acoustic sound art. The main concern was to build a musical instrument which combined the natural sounds of metal objects with electronic sounds and digital signal processing. The first chapter of this thesis is a discussion of the aesthetic concerns of sound art, and of some sound artists' works responding to those concerns through their use of technology. The second chapter is a description of the author's project in sound art. He decided to create a "sculpted instrument" founded on the principle of the musical saw. Composed of two metal plates connected to a single stand, different parts of the instrument are electronically amplified and processed through a DSP. I call the instrument the Forgery. The third chapter of this thesis is a presentation of the composition for the Forgery. A formal analysis is coupled with a discussion of the issues addressed by the composition. More particular elements of the piece are placed within these issues and within the form of the piece. The final chapter contains some conclusions drawn from the project. Advantages are weighed against disadvantages; future prospects and directions of the instrument and the approach to creativity it represents are considered.

See PAST. A model of timbre perception based on the mechanics and constraints of the auditory system is presented. This work is motivated by the need for timbre perception models that reflect both static and time-varying properties of sound. The applications of such a model include timbre recognition, data compression and perceptual based timbre modification and (re)synthesis in music composition. The preprocessor to teh model, based on the MQ sinusoidal analysis algorithm, simulates the effects of cochlear analysis including auditory masking. The output of the preprocessor is a time and frequency analysis organized into spectral tracks. The model represents properties of tracks as curves. The feature set of the model includes spectral onset asynchrony, spectral onset peak asynchrony, spectral onset peak envelope, spectral persistence, amplitude envelope, mean-frequency envelope, spectral envelope, spectral density, and harmonicity (how well a group of tracks fits a harmonic series). The model is successful in simulating timbre recognition and in creating timbre categories. The software implementation of the model displays feature curves graphically. Timbral modifications can be made by altering these curves or by applying selected features from one timbre to another. Modified timbres can be resynthesized to audition the perceptual results.

Audio compression techniques such as those devised by the Motion Picture Experts Group (MPEG) can significantly reduce the storage and transmission requirements for high quality digital audio. However, compression complicates the manipulation of audio in many applications. Operations such as gain control, equalization, and mixing are common and their implementation is relatively simple for PCM audio streams. Application of these operations to compressed digital audio, though, is not as clear. This paper details algorithms for direct manipulation of digital audio in MPEG compressed form. These operations require orders of magnitude less processing, do not entail recompression of the digital audio, and incur minimal quality loss.

Physical modeling synthesis, based on a model of a physical system, offers many advantages over traditional sound synthesis methods which focus on modeling the spectrum and amplitude envelope of sounds themselves. This thesis offers a general discussion of physical modeling synthesis and describes Physical Modeling Synthesis (PMS), a synthesis application based on a physical model of struck strings, developed for the Macintosh Computer by the author.

With the increasing availability of standard frequency-domain file formats, a new generation of frequency-domain transformation computer software is now surfacing. This document describes the QuickMQ application for the Macintosh computer and its numerous frequency-domain transformations. These transformations include convolution, deconvolution, spectrum mix, granular desynthesis, brightening, harmonic rotation, and a unique implementation of a spectrum-processing language based on the graphics manipulation language Pico by Gerard Holzmann (1988). The QuickMQ Macintosh application allows editing of frequency frame files in a user-friendly movie environment, where the changes in spectrum may be observed over time. QuickMQ allows the user to describe frequency domain translation algorithms to produce new audio transformations based upon original frequency files that alter frequency, time and amplitude. http://music.dartmouth.edu/~skos/it/vsymph

The following documents the thoughts, methods, and implementation of an interactive system used to realize musical compositions. It is based on the notion of combining the wide range of sonic possibilities offered by electronic technology with performances involving acoustic instruments, more specifically the shakuhachi, a Japanese end-blown flute. The main interest was to develop an interactive system that was capable of integrating electronically generated sound with the shakuhachi in a variety of ways. A series of essays is presented to explicate the general paradigm serving as a framework for developing the interactive system. Topics include: a rationale for integrating electronic and acoustic sound in composition, an examination of the term gesture as it is used to describe sonic phenomena in the context of electro-acoustic music, a description of the shakuhachi's sonic qualities, the necessity for incorporation temporal flexibility in the design of the interactive system, the advantage of amplifying the shakuhachi when performing in conjunction with electronic sound, techniques which can be used to integrate acoustic and electronic sound at various levels of perception, and a description of the interactive system itself, both in terms of the hardware devices and software design employed. An ideal realization of the interactive system based on the essays outlines would provide a platform to appropriate the dynamic qualities of the shakuhachi and generate electronic sounds with analogous sonic characteristics. The interactive system implemented here is a subset of this ideal system. One of its main features is to monitor the amplitude dynamics of the shakuhachi performance and use this information to control the spatial movement of the electronic sounds. Future development of the system would prioritize an enhancement of the monitoring system, as well as the implementation of processes that generate sounds which perceptually and necessarily group with the acoustic sound.

This paper describes an application for experimental digital filter design using neural networks. Neural networks have been used in signal processing to model digital filters in applications that include nonlinear adaptive filtering (Tresp et al. 1992, Nerrand et al. 1993), FIR and IIR networks (Back and Tsoi 1991, Wan 1990), and noise filtering (Anderson and Montgomery 1990, Hoyt and Weschsler 1990). In SpleEn, a signal processing learning environment, neural networks are used as an experimental approach to filter design. SpleEn employs feedforward neural networks designed in terms of adaptive filters. The networks are trained using error correction: least-mean square (LMS) with back-propagation. SpleEn learns to generalize the mapping between pairs of user-entered sounds, represented as digital samples. SpleEn is a software application for the Macintosh computer developed using Think C and DSP Designer.

The McAulay-Quatieri sinusoidal representation of sound extracts frequency spectra as amplitude peaks thereby providing a flexible and perceptually intuitive format for frequency domain transformations. This thesis presents some techniques for sound transformation based on the McAulay-Quatieri representation. A new application for Macintosh computers developed by the author, McAulay-Quatieri Transformer, will be presented. It performs these spectral transformations on data created by the Lemur Macintosh analysis/ synthesis program.

This thesis is intended as a supplement to the electro-acoustic composition Vocali. The introduction provides a brief history of the development of languages changing role within musical composition of the 20th century. Following this, the texts used in Vocali are presented with a discussion of relevant aesthetic issues. The third section describes the recording and sound manipulation processes used in the composition. In the final section, the form of the composition is analyzed on a macroscopic and microscopic scale. Detailed illustrations are provided.

HS is a programming language for computer assisted composition and experimentation that is oriented towards specifying formal descriptions, and artificial models, of generative musical structures. The language uses atomic units that are non-implementation specific. Therefore HS is style and media non-biased. HS uses a limited set of polymorphic objects that act upon lists of data that comprise atomic units, primitive objects and object hierarchies. Both nesting and recursion are applied to these objects giving rise to hierarchical, self-referential structures. The prototype language has been implemented and is running on a NeXT machine. The design of the language and the function of the primitives are a response to some general observations on the uses of Computer Assisted Composition (CAC). The first part of this thesis presents a survey of CAC and shows that there are three main objectives for composition using computers: event specification and editing, procedural specification and artificial modeling of musical structures. A CAC system should represent these objectives in a clear and consistent manner and integrate them so that the user is not forced to work under the boundaries of one particular paradigm. The second section presents the design philosophy and implementation of the HS programming language. HS is an integration of many aspects of CAC presented above into a small symbolic language. The goal is to hide the explicit complexities of musical structures by embedding complex data structures in a more convenient representation. To achieve this goal four main issues were important for the design process. The first was the problem of defining a set of primitive functions that represent building blocks for CAC. The second was the consideration of how these primitive objects behave with respect to their intended generative characteristics, the third issue was how to represent these objects for subsequent manipulation in a symbolic language and the final consideration is how the output of the language interfaces to a compositional working environment. Some examples of Computer Assisted Composition using HS are given and criteria for future extensions are considered.

Mutation functions, derived from Larry Polansky's theory of morphological metrics, are used to mutate the contours of digitally sampled sounds in their frequency domain representation as well as their time domain representation. The theory of the mutation functions is presented and examples of each are shown. An application of this novel method of cross-synthesis has developed on a NeXT computer and a library of the mutation functions on various sounds is cataloged. (See McKinney, ICMC Proceedings, 1992, Montreal).

This thesis describes a set of tools for manipulating sampled signals using wavelet transforms, and outlines potential musical applications of this library. Wavelets improve on existing time frequency methods by representing signals as a sum of a single wavelet basis function at different positions and scales. A wavelet representation can be used to synthesize or transform sounds in a variety of interesting ways. While the use of wavelets in music has been discussed in several papers, there has been no easy way for composers of electro-acoustic music to experiment with wavelet transformations. In addition, there has been a need for a clear, non-technical presentation of wavelet theory, so that composers and other musicians can understand their potential usefulness. The thesis begins by examining relevant background material, including the theory and musical applications of various time and frequency based representations. The theory of wavelets is presented in more detail, along with some of the ways wavelets can be used in a musical context. The software library is then described at several levels of detail, including a brief tutorial and a set of examples. A set of appendices provide additional related information. (See Kussmaul, ICMC Proceedings, 1992, Montreal).

Current wind synthesizer controllers are modeled on traditional wind instruments so that players of traditional wind instruments can play the controllers without altering their technique substantially. While this has obvious commercial advantages, it also means that the new controllers inherit many of the limitations of their acoustic forebears. The author argues that a far more powerful wind synthesizer controller could be built by starting from scratch, designing a controller that is based on musical and ergonomic goals. The design and construction of one such controller is described.