Score Analyzer: Automatically Determining Scores Difficulty Level for Instrumental e-Learning

Nowadays, huge sheet music collections exist on the Web, allowing people to access public domain scores for free. However, beginners may be lost in finding a score appropriate to their instrument level, and should often rely on themselves to start out on the chosen piece. In this instrumental e-Learning context, we propose a Score Analyzer prototype in order to automatically extract the difficulty level of a MusicXML piece and suggest advice thanks to a Musical Sign Base (MSB). To do so, we first review methods related to score performance information retrieval. We then identify seven criteria to characterize technical instrumental difficulties and propose methods to extract them from a MusicXML score. The relevance of these criteria is then evaluated through a Principal Components Analysis and compared to human estimations. Lastly we discuss the integration of this work to @MUSE, a collaborative score annotation platform based on multimedia contents indexation.