Automatic performance and piece assessment

How good is a performance? How difficult will a piece be to perform? For the first time, TROMPA will merge expertise by project partner MDW (e.g., Langner & Goebl, 2003, Flossmann et al., 2010) and methods proposed by UPF and exploited by VL, such as ways to rate singing voice performances, a technology successfully recently integrated in computer games (Mayor et al., 2009).

In terms of research challenges, singing voice evaluation needs to be expanded to full choirs. Multiple simple indices of specific complexities will also be established for piece difficulty assessment, based on the technical (physiological) and musical challenges of the piece. These will be combined into a joint model validated by crowdsourcing. Human judgments will also be used to build a separate model for ranking pieces by human-perceived difficulty.



Flossmann, S., Goebl, W., Grachten, M., Niedermayer, B., & Widmer, G. (2010). “The Magaloff Project: An interim report.” Journal of New Music Research, 39(4), 363–377. doi:10.1080/09298215.2010.523469

Langner, J., and Goebl, W. “Visualizing expressive performance in tempo-loudness space”, Computer Music Journal, 27(4), pp. 69–83, 2003.

Mayor, O., Bonada J., & Loscos A. (2009). "Performance Analysis and Scoring of the Singing Voice"., Proc. AES 35th Int. Conf.: Audio for Games, 2009.