Review by Choice Review
Rowe (NYU) explores the development of approximately 25 musical processes with which computers can interact more intelligently with live musicians. These processes, which can be used to build a computer music program's foundation of musical knowledge, are built on three types of algorithms: symbolic algorithms, based on the representation of objects (e.g., triads, key signatures) and their manipulation; sub-symbolic algorithms, which learn behaviors from prior inputs (e.g., quantification of rhythmic pulse); and algorithms that learn to detect musically relevant segments and patterns. Rowe describes numerous compositional techniques and specific interactive systems that composers use. An accompanying CD-ROM includes several working versions of examples described, but this reviewer had trouble accessing these on any but a Mac OS9; many of the examples require the freeware program called MidiShare (Rowe provides information about obtaining this program). The CD-ROM's video and audio clips are playable on either Windows or Mac platform, but the index provided does not clearly indicate the content of these tracks. This fascinating but highly technical book requires knowledge of C++ and Max. Recommended only for graduate collections and researchers/professionals in composition, music theory, music cognition, or artificial intelligence. T. E. Buehrer Kenyon College
Copyright American Library Association, used with permission.
Review by Choice Review