John A. Biles

Evolutionary Computer Music

An Introduction to Evolutionary Computing for Musicians.- Evolutionary Computation for Musical Tasks.- Evolution in Digital Audio Technology.- Evolution in Creative Sound Design.- Experiments in Generative Musical Performance with a Genetic Algorithm.- Composing with Genetic Algorithms: GenDash.- Improvizing with Genetic Algorithms: GenJam.- Cellular Automata Music: From Sound Synthesis to Musical Forms.- Swarming and Music.- Computational Evolutionary Musicology.

GenJam: evolution of a jazz improviser

Abstract GenJam is an interactive genetic algorithm that performs jazz improvisation, both independently as a stand-alone soloist and collaboratively as an interactive agent. This chapter describes GenJam’s chromosome representations for melodic ideas, the genetic operators it uses to evolve those ideas, and its use of the evolutionary paradigm in real-time interactive performance. The chapter also discusses pragmatic and artistic issues regarding GenJam’s participation in the author’s Virtual Quintet. 1. Introduction At the time of this writing, GenJam is about six years old, and it has come a long way from its initial conception and implementation. What began as a project intended to generate a couple of papers during a sabbatical year has evolved into a viable performance system that plays gigs two or three times a month. GenJam’s current repertoire comprises over 160 tunes in a variety of jazz, Latin and new age styles. It can perform in 4/4, 3/4, 5/4 and 12/8 time, and it has allowed the author to work as a single trumpet player in a climate that no longer supports hiring live jazz quintets. The Virtual Quintet can compete with single pianists and guitar players, something a trumpet player cannot do alone. More importantly, it allows the Virtual Quintet to compete with the dreaded DJ! At a technical level, GenJam stretches the genetic algorithm paradigm in some interesting ways. It uses two hierarchically interrelated populations to represent its store of melodic ideas for creating improvisations. The goal of its search is to build entire populations of good melodic ideas, not just search for a single individual, which magnifies the problem of premature convergence. Consequently, many of GenJam’s mutation operators promote diversity in its populations. GenJam’s mutation operators are unusual in that they do considerably more than flip the occasional bit. In fact, most of GenJam’s knowledge of improvisational development is embedded in its musically meaningful mutation operators. Finally, GenJam uses its evolutionary machinery for real-time interactive performance by evolving what it hears a human play into what it plays in response. At an artistic level, GenJam’s aspiration is to be received by its audience as a competent straight-ahead jazz improviser. This places GenJam in the mainstream artistically, a place most computer musicians seem to avoid. GenJam makes no attempt to extend the conception of improvisation or make any profound artistic statement on the nature of music. It simply tries to meet the audience’s expectations for what straight-up jazz sounds like. Instead of challenging an audience artistically, then, GenJam tries to challenge an audience technically by achieving an accessible and convincing performance from a computer. This chapter describes how GenJam adapts the evolutionary paradigm to meet this goal.

GenJam in Perspective: A Tentative Taxonomy for GA Music and Art Systems

GenJam is an interactive genetic algorithm (GA) that models a human jazz improviser and performs regularly as the authors sideman on jazz gigs. GenJam learns to improvise full-chorus solos under the guidance of a human mentor and trades fours in real time with a human performer in chase choruses. In this article, the author first briefly describes GenJams architecture, representations, genetic operators and performance characteristics. He then places GenJam in the context of a proposed taxonomy for GA-based music and art systems.

Life with GenJam: interacting with a musical IGA

GenJam, short for genetic jammer, is an interactive genetic algorithm (IGA) that models a jazz improviser and performs as a featured soloist in the authors virtual quintet. GenJam evolves populations of melodic ideas under the guidance of a human mentor, whose feedback provides the environment under which individual ideas either survive to breed or die off. GenJam also uses its genetic algorithm machinery as a real-time melodic development paradigm to evolve phrases played by a human into its improvised responses in chase choruses. The paper provides an overview of the GenJam architecture and focuses on interface issues for three classes of users: mentors who train GenJam individually, audiences who train GenJam collectively, and performers who interact with GenJam in real-time performance situations.

Evolutionary Computation for Musical Tasks

If the preceding chapter was an introduction to evolutionary computation (EC) for musicians, this chapter is intended as an introduction to music as a problem domain for EC researchers. Since we cannot hope to provide even a bare-bones treatise on music appreciation, much less music theory, we assume that the reader is at least somewhat familiar with music, if not as a producer, at least as a consumer. We will start by trying to define some musical terms to work with, including ‘music’ itself, which will lead us to a brief excursion into human–computer interaction as a metaphor for musical performance. We will then conduct an informal task analysis of music to define the tasks musicians perform and survey how EC has been applied to facilitate (or obfuscate, in some cases) the performance of those tasks. We will then summarize the various approaches that have been taken in representation, fitness and genetic operators.

GenJam: An interactive genetic algorithm jazz improviser

GenJam is an interactive genetic algorithm that learns jazz improvisation. It uses two hierarchically related populations to represent melodic ideas at the measure and phrase levels. These populations are evolved using tournament selection, single‐point crossover, musically meaningful mutation, and replacement with a 50% generation gap. Fitness for the individual measures and phrases is derived from real‐time feedback, which is provided by a human mentor while GenJam improvises to the accompaniment of a synthesized rhythm section. GenJam has been used for actual gigs under the billing Al Biles Virtual Quintet, which features the author on trumpet and GenJam on a variety of synthesized instruments, playing a repertoire of over 90 tunes in a variety of jazz, Latin, and new‐age styles. Recent enhancements to GenJam include a pitch‐to‐MIDI capability that allows GenJam to listen to a human soloist, map his four‐bar phrases to the GenJam genetic representation for phrases, apply selected mutation operators to ...

ESPRIT: A signal processing environment with a visual programming interface

ESPRIT (Explorer speech processing system from the Rochester Institute of Technology) is an integrated speech research development environment that runs on the TI Explorer, optionally augmented by the TMS‐320 based Odyssey DSP board. The goal of ESPRIT is to provide speech scientists, linguists, and engineers an intuitive environment in which to collect, process, and display speech signals. ESPRITs module editor allows users who are not programmers to draw data‐flow programs made up of built‐in and user‐defined speech processing algorithms, display functions, and standard utilities. ESPRITs display editor allows users to manipulate the graphical displays that result from running these programs to zoom, scroll, rearrange, take precise measurements, and perform a variety of other operations. While ESPRIT provides standard signal processing algorithms (FFT,LPC) and displays (waveforms, spectrograms, waterfalls, spectral slices), users who develop their own Lisp or TMS 320 programs can easily install them t...