Performaster

Increasingly more new piano models are applying MIDI recording for their users. However, there are few accessible tools that compare performances to provide every day users with a better understanding of what a sophisticated performance entails. To fill this gap, I developed Performaster, a performance analysis framework designed to compare MIDI recordings across performers’ decision making within their interpretations. This system allows users to analyze their expressive choices relative to reference performances. This offers researchers a structured means of quantifying the interpretative decisions. In turn, this can be correlated to perceived empathetic responses. Through this project, I learned how much computational models can support expressive understanding, thus leading me closer to answering my fundamental question.