How Do People Assess Computer Generated Expressive Music Performances?

Music performance has being studied since long time and several computational systems were developed for generating expressive music performances. These models are generally evaluated by comparing their predictions with actual performances, both from a quantitative and a subjective point of view, often focusing on very specific aspects of the model. However little is known about how listeners evaluate the generated performances and which are the factors influencing their judgement and appreciation. In this paper we present two experiments, conducted during two dedicated workshops, to start understanding how the audience judges the entire performances. In particular we analyzed possible different preferences and expectations of the listeners and influencing factors, such as cognitive styles.