ABSTRACT
The extent to which we can represent musical information in computers, in a way that matches our mental images, is one of the most intriguing and compelling questions in music research. Leman (1993) distinguishes three kinds of musical representation:
Acoustical, i.e. based on the physical properties of sound (e.g., sonograms, spectrograms).
Subsymbolic, i.e. based on the known behaviour of the human hearing system and how the human brain processes auditory information (e.g., auditory models).
Symbolic, i.e. based on the manipulation of symbols (e.g., Common Music Notation).
