11:20 - 11:40
Andrea Cogliati, Zhiyao Duan
University of Rochester
Piano Music Transcription in Music Notation
Automatic music transcription (AMT) is the process of automatically inferring a high-level symbolic representation, such as music notation or piano-roll, from a music performance. Despite over four decades of active research, AMT systems are still far from matching humans in terms of transcription accuracy. Moreover, most AMT systems output a parametric representation of the musical performance, i.e., a representation expressed in physical terms, such as seconds for note onsets and durations, and hertz or MIDI numbers for pitches. However, traditional music notation is generally more desirable for many applications connecting humans and machines, such as computational musicological analysis and music tutoring systems. Our research focuses on improving the accuracy of AMT systems for piano performances, and bridging the gap between parametric transcriptions and music notation.
Design downloaded from free website templates.