Incrementally improving interactive music systems
Robert C. Rowe · Contemporary Music Review · 1996
There are strong musical motivations for current efforts to transfer the human concepts underlying musicianship to computer programs. Particularly for interactive music systems, those meant to perform with human players onstage in contexts that may include improvisation, such concepts are essential. Four research areas are reviewed that promise incremental improvement to interactive music systems as work toward machine musicianship proceeds. These areas are: (1) timbre analysis, (2) pattern processing, (3) machine learning, and (4) musical schemata. As each area is discussed, the musical implications motivating its consideration are reviewed together with some relevant contemporary research projects.