Keynote Speaker
Dr. Johanna Devaney / Brooklyn College and the Graduate Center
2020-10-13 | 10:30 (Live) (America/New_York)
2020-10-13 | 23:00 (Replay) (America/New_York)
This talk will reflect on what we can observe about musical performance in the audio signal and where MIR techniques have succeeded and failed in enhancing our understanding of musical performance. Since its foundation, ISMIR has showcased a range of approaches for studying musical performance. Some of these have been explicit approaches for studying expressive performance while others implicitly analyze performance with other aspects of the musical audio. Building on my own work developing tools for analyzing musical performance, I will consider not only the assumptions that underlie the questions we ask about performance but what we learn and what we miss in our current approaches to summarizing performance-related information from audio signals. I will also reflect on a number of related questions, including what do we gain by summarizing over large corpora versus close reading of a select number of recordings. What do we lose? What can we learn from generative techniques, such as those applied in style transfer? And finally, how can we integrate these disparate approaches in order to better understand the role of performance in our conception of musical style?
Zoom links will be shared on Slack.