Toward the detection of interpretation effects and playing defects

Precise automatic music transcription requires accurate modeling and identification of the spectral content of the audio signal. But music can not be reduced to a succession of notes, and an accurate transcriptor should be able to detect other performance characteristics, such as slow tempo variations or, depending on the instrument detecting some interpretation effects. In a pedagogic way a student could want to improve his level and a good challenge will be to estimate the quality of play of musician. In this paper we present some of the most common playing defects and interpretation effects and we propose a way for detecting them.