Vol. 26 Issue 2 Reviews
The Future of Computer Music Software—A Symposium
Dartmouth College, Hanover, New Hampshire, USA, 19 November 2001

Reviewed by Matthew B. Smith
Hanover, New Hampshire, USA

With the ambitious objective of discussing the future of computer music software, Dartmouth College hosted a monumental symposium featuring some of the most visible innovators in the field. Max Matthews, Gareth Loy, Barry Vercoe, David Zicarelli, Miller Puckette, and James McCartney were all invited to present both their past accomplishments and also the current direction of their work.

Eric Lyon, professor at Dartmouth College, welcomed the audience and gave a brief introduction of the participants. Quoting Mr. Loy’s assertion in 1989 that so many computer music systems rapidly reach a point of extinction, Mr. Lyon countered that today’s symposium would explore the few exceptions that have survived longer and whose concepts are still fundamental.

Mr. Matthews, unquestionably the first to embark on the creation of computer music software, spoke first. Opening with the disclaimer, “I’m going to be talking about the past,” he described the innovations he implemented between 1957-1964 at Bell Labs. Moving quickly from a single voice consisting of a triangular waveform in Music I, wavetable oscillators, unit generators, the notion of a score with p-fields, and block diagrams were introduced as Mr. Matthews progressed toward Music V. One critical insight he pointed out was his intent to shift the burden of timbre design onto the composer. As with many of his innovations, this idea is part of the foundation Mr. Matthews provided for the development of computer music software.

Along with F. Richard Moore, Mr. Loy was one of the primary authors of the software developed at the University of California at San Diego as part of the CARL (Computer Audio Research Laboratory) distribution in the 1980s. The goals of this distribution were to coordinate the community work of developing audio tools for the Unix operating system. This provided a context for the exchange of free and open source code that was reasonably portable between different Unix systems. As Mr. Loy described it, the authors of the CARL distribution were “generalists of the most general sort,” so that the user could determine what happened between a “general purpose computer” and a “general purpose loudspeaker.”

While the contributions of the CARL distribution and Mr. Matthews’ work have had an undeniable impact on the current state of computer music, these presentations focused mainly on the past and failed to address the main topic of the symposium: the future of computer music software. It could be implied, though, from these historical presentations that the next generation of music software will be heavily dependent on its predecessors and perhaps may only be an enhancement of the paradigms already established.

Mr. Vercoe continued the pattern of presenting a linear history of his accomplishments beginning with his adaptation of Mr. Matthews’ work into Music 360. This and future permutations were necessary as computer hardware evolved and new capabilities were added. This resulted in Csound. Mr. Vercoe also presented some of his research in which computers interact with performers, with the eventual implementation of this into Real-Time Csound, NetSound, and the industry supported Extended Csound. Regarding the future of music, Denon karaoke machines will be utilizing many of the Extended Csound features for both sound generation and having the machine “follow” the performer. Also, Mr. Vercoe’s most recent project, MPEG-4, utilizes NetSound techniques for reducing file size significantly by synthesizing only after being transmitted over the Internet.

The creator of Max and pd, Mr. Puckette introduced his software as a blank page that imposes no musical style. Addressing the economics of software, he espoused the positive aspects of both free and commercial software: free software benefiting from its distance from commercial pressures and the constraints of the marketplace, and commercial software being more visible and easily available. Going back into history, Mr. Puckette praised Mr. Matthews’ perennial wave-table oscillator and then demonstrated how simple it was to call up such an oscillator in pd. He also emphasized the importance of being able to archive work so that “musical practice is not hopelessly irreversibly imbedded in musical software.” Embracing the proliferation of inexpensive hardware and Linux, Mr. Puckette discussed the possibility of empowering the rest of the world to make their own computer music while maintaining diversity.

Mr. Zicarelli’s work in Max and MSP is so dependent on Mr. Puckette’s ideas that at first glance it may seem redundant to have had both of them speak. Nevertheless, he started by mentioning some of the visual design concerns, and attributed the popularity of the software to the ability to externally load objects. Mr. Zicarelli described the ease with which objects can be linked together as “dynamic binding” and attributed such architecture as a bias toward individual idiosyncrasy rather than experimentation. With that in mind, Max/MSP aims to facilitate creative work through system feedback, clear documentation, and a small number of basic concepts that can be applied throughout the entire program. The implementation of JavaScript into Max now allows users to write Max externals without programming in C. It also allows for the patch to be dynamic and build new parts in response to the user interface.

Describing SuperCollider as a “high level language with a real time synthesis engine,” Mr. McCartney quickly outlined the various versions of the program. These included the current developments of SuperCollider 3 and SuperCollider Server. The most recent programs split into different machines or in the latter case even separate applications. Programmatic patch building, spawning events, trees of algorithms, and various objects and classes were briefly described. Mr. McCartney observed that in designing an architecture certain decisions cut off other possibilities. SuperCollider Server partially explores these other areas while sacrificing some of the benefits of the earlier versions. Running as Macintosh OS X command line programs, the language and synthesis engines communicate via the Open Sound Control network protocol. Overall, Mr. McCartney’s presentation was more technically detailed than the others as he seems to be perpetually modifying and re-creating the SuperCollider environment.

Following these presentations, a panel discussion took place with the six speakers. Beginning the discussion, Mr. Lyon, serving as moderator, brought up the subject of “cannibalism” or the inclination of different software creators to incorporate ideas from others. Mr. Zicarelli suggested that other software (such as Csound) could be easily incorporated into MSP as an object. He also mentioned the possibility of MSP creating plug-ins for use in other software. Mr. McCartney indicated that rather than combining things, dividing up components into a modular system might be more effective.

When the question of what would happen if the software creators had to start over from scratch arose, Mr. Puckette answered that his previous experience and memory would make him incapable of coming up with anything new or revolutionary. This insight does bring up the question of whether these panelists really have anything to say about the future of computer software other than the possible revisions of their own.

With the panel discussion lasting two hours, many issues and questions were raised. While this meeting struggled to provide more than a fleeting glance into the future, these creators are obviously still searching for new methods and solutions to the perplexing problems that arise from music software creation.