International Computer Music Conference 2002 (review)

[Editor's note: Selected reviews are posted on the Web at mitpress2.mit.edu/e-journals/Computer-Music-Journal/Documents /reviews/index.html. In some cases, they are either unpublished in the Journal itself or published in an abbreviated form in the Journal.]

The 2002 International Computer Music Conference (ICMC2002) was held in Göteborg, Sweden, jointly organized and hosted by the School of Music and Music Education at Göteborg University and the Chalmers University of Technology. Göteborg is the second largest city in Sweden and is known as the manufacturing center of Volvo automobiles and Ericsson cellular phones. ICMC2002 was overall a successful conference run by a very professional and experienced group. Much respect should be accorded the general coordinator, Magnus Eldénius, and his efficient staff for accepting, organizing, and deploying such a large, difficult project as ICMC2002. Furthermore, Göteborg University provided an excellent location for the event, with all the technology required to successfully host this "hyper-technological" event.

The chosen theme for ICMC2002 was "Voices of Nature." Mr. Eldénius asked in his introduction: "What is the border between nature and art? Nature is present in one way or another in most of today's computer music activities—explicitly as in physical modeling, psychoacoustics, evolutionary algorithms, or implicitly as a foundation for the physical laws that govern most human activities. And, of course, as an artistic inspiration for many composers." This theme is exciting, but unfortunately there were only a few papers and pieces that directly dealt with it.

ICMC2002 didn't have any poster sessions; only papers, demos, and, of course, concerts. Of 193 submitted papers, 121 were accepted, so parallel sessions were necessary to fit everything into five days. (Thus I must make the obligatory statement that I cannot do justice to giving a complete review of the ICMC2002 papers since that would violate fundamental laws of physics!)

Groups and institutions that had three or more papers and demos at ICMC2002 were numerous. These include the Center for Computer Research in Music and Acoustics (CCRMA); Institut de Recherche et Coordination Acoustique/Musique (IRCAM); The Department of Information Engineering (DEI) and the Center of Computational Study (CSC) of the University of Padova, Italy; Massachusetts Institute of Technology (MIT) Media Lab; NTT Communication Science Research Lab; Carnegie Mellon University (CMU); Virginia Center for Computer Music; Sony Computer Science Laboratory in Paris; and the University of California at San Diego (UCSD). These groups cover a broad range of specialties in current computer music research from physical modeling to musical controllers, machine listening, and signal description.

This year's Swets & Zeitlinger Distinguished Paper Award went to François Pachet (Sony Computer Science Laboratory) for his paper entitled "The Continuator: Music Interaction with Style." Mr. Pachet demonstrated his program, which aims to create a stylistically consistent music imitation system that is also musically interactive. The title refers to the system's ability to take a player's musical input and continue in a stylistically similar way. His demonstration at the conference provided a fulfilling and amusing look at this creative solution to making imitative music systems interactive. Several images, movies, and the software are available online (www.csl.sony.fr/pachet).

Machine listening was a popular cognitive topic this year. Roger Dannenberg (CMU) demonstrated a program that implements machine listening which produces a high-level analysis and "explanation" of an input piece of music, such as John Coltrane's Naimaor the Christmas Carol We Three Kings. Kenji Suzuki et al. from Waseda University, Japan, presented their first step in creating an autonomous music performance system: an algorithm with the ability to create its own musical sensibility by listening to several raw audio signals. Other papers tackled onset detection using neural networks, quantifying melodic similarity, and improving score following.

Physical modeling of the Tibetan singling bowl was also a popular topic this year. Georg Essl and Perry Cook of Princeton, and Stefania Serafin et al. (CCRMA) presented results that use banded waveguides. Mr. Essl has expanded his model to incorporate...