Project

Automated Analysis of Musical Structure for Music Segmentation

This project examines the psycho-acoustic bases of the perception of musical structure by human listeners. Computational models will be built to mimic basic musical perception, such as parsing music into phrases or sections (i.e., recurrent structural analysis), identifying the main themes or hooks of a musical piece (i.e., music summarization), and detecting the most �informative� parts of music for making certain judgments (i.e., detection of musical salience), upon taking complex acoustic signals as input. It will inquire scientifically into the nature of the music-listening process, and offer a practical solution to difficult problems in computer-based multimedia.

This project examines the psycho-acoustic bases of the perception of musical structure by human listeners. Computational models will be built to mimic basic musical perception, such as parsing music into phrases or sections (i.e., recurrent structural analysis), identifying the main themes or hooks of a musical piece (i.e., music summarization), and detecting the most �informative� parts of music for making certain judgments (i.e., detection of musical salience), upon taking complex acoustic signals as input. It will inquire scientifically into the nature of the music-listening process, and offer a practical solution to difficult problems in computer-based multimedia.