Arne Eigenfeldt

The Evolution of Evolutionary Software: Intelligent Rhythm Generation in Kinetic Engine

This paper presents an evolutionary music system that generates complex rhythmic polyphony in performance. A population of rhythms is derived from analysis of source material, using a first order Markov chain derived from subdivision transitions. The population evolves in performance, and each generation is analysed to provide rules for subsequent generations.

Evolving structures for electronic dance music

We present GESMI (Generative Electronica Statistical Modeling Instrument), a software system that generates Electronic Dance Music (EDM) using evolutionary methods. While using machine learning, GESMI rests on a corpus analysed and transcribed by domain experts. We describe a method for generating the overall form of a piece and individual parts, including specific patterns sequences, using evolutionary algorithms. Lastly, we describe how the user can use contextually-relevant target features to query the generated database of strong individual patterns. As our main focus is upon artistic results, our methods themselves use an iterative, somewhat evolutionary, design process based upon our reaction to results.

Emergent Rhythms through Multi-agency in Max/MSP

This paper presents a multi-agents architecture created in Max/MSP that generates polyphonic rhythmic patterns which continuously evolve and develop in a musically intelligent manner. Agent-based software offers a new method for real-time composition that allows for complex interactions between individual voices while requiring very little user interaction or supervision. The system described, Kinetic Engineis an environment in which networked computers, using individual software agents, emulate drummers improvising within a percussion ensemble. Player agents assume roles and personalities within the ensemble, and communicate with one another to create complex rhythmic interactions. The software has been premiered in a recent work, Drum Circle, which is briefly described.

Corpus-based recombinant composition using a genetic algorithm

A generative real-time composition system is described that uses a genetic algorithm to create a population of melodic and rhythm phrases that are combined by intelligent musical agents. The initial population is derived from an offline analysis of a corpus; the population undergoes continual breeding using rules derived from the population itself. The system’s role in the generation of musical material for the acoustic composition Other, Previously, for string quartet, is discussed.

Populations of populations: composing with multiple evolutionary algorithms

We present a music composition system in which musical motives are treated as individuals within a population, and that the audible evolution of populations over time are of musical interest. The system additionally uses genetic algorithms to generate high level musical aspects that control how the population is presented, and how it may be combined with other populations. These algorithms feature fitness functions that adapt based upon context: specifically, by using an analysis of the evolving population, the fitness functions adjust their constituent parameters in selecting strong individuals.

Real-time composition as performance ecosystem

This article proposes that real-time composition can be considered a new performance ecosystem. Rather than an extension of electroacoustic instruments that are used within improvisatory environments, real-time composition systems are produced by composers interested in gestural interactions between musical agents, with or without real-time control. They are a subclass of interactive systems, specifically a genre of interactive composition systems that share compositional control between composer and system. Designing the complexity of interactions between agents is a compositional act, and its outcomes are realised during performance-more so than most interactive systems, the new performance ecosystem is compositional in nature.

An Introduction to Musical Metacreation

Musical metacreation (MuMe), also known as musical computational creativity, is a subfield of computational creativity that focuses on endowing machines with the ability to achieve creative musical tasks, such as composition, interpretation, improvisation, accompaniment, mixing, etc. It covers all dimensions of the theory and practice of computational generative music systems, ranging from purely artistic approaches to purely scientific ones, inclusive of discourses relevant to this topic from the humanities. MuMe systems range from purely generative ones to a variety of interactive systems, such as those for computer-assisted composition and computer-assisted sound design. In order to better appreciate the many dimensions of this interdisciplinary domain and see how it overlaps and differs from research in computer music, this introduction provides a general entry point. After defining and introducing the domain, its context, and some of its terminology, we reflect on some challenges and opportunities for the field as a whole.

A sonic eco-system of self-organising musical agents

We present a population of autonomous agents that exist within a sonic eco-system derived from real-time analysis of live audio. In this system, entitled Coming Together: Shoals, agents search for food consisting of CataRT unit analyses, which, when found, are consumed through granulation. Individual agents are initialised with random synthesis parameters, but communicate these parameters to agents in local neighborhoods. Agents form social networks, and converge their parameters within these networks, thereby creating unified grain streams. Separate gestures thus emerge through the self-organisation of the population.

Real-time timbral organisation: Selecting samples based upon similarity1

A comparison is made between two systems of real-time sample selection using timbral proximity that has relevance for live performance. Sound files in large sample libraries are analysed for audio features (amplitude RMS, spectral centroid, spectral flatness, and spectral energy using a Bark auditory modeller), and this data is statistically analysed and stored. Two methods of organisation are described: the first uses fuzzy logic to rate sample similarity, the second uses a self-organising map. The benefits and detriments of each method are described.

Generative Music for Live Performance: Experiences with real-time notation

Notation is the traditional method for composers to specify detailed relationships between musical events. However, the conventions under which the tradition evolved – controlled relationships between two or more human performers – were intended for situations apart from those found in electroacoustic music. Many composers of electroacoustic music have adopted the tradition for mixed media works that use live performers, and new customs have appeared that address issues in coordinating performers with electroacoustic elements. The author presents generative music as one method of avoiding the fixedness of tape music: coupled with real-time notation for live performers, generative music is described as a continuation of research into expressive performance within electroacoustic music by incorporating instrumentalists rather than synthetic output. Real-time score generation is described as a final goal of a generative system, and two recent works are presented as examples of the difficulties of real-time notation.