Johanna Devaney: Automatic Extraction of Performance Data from Recordings

Johanna Devaney is a Montreal-based composer/theorist. Her compositional interests are both electronic and acoustic, while her theoretical work is focused modeling performance practice. She is currently working on a PhD in the Department of Music Research at the Schulich School of Music, McGill University where she is currently studying with Ichiro Fujinaga.

Recorded audio offers a wealth of information for studying performance practice. This talk will examine the challenges of automatically extracting performance information from audio recordings, with a focus on the singing voice. With this information it is possible to study a number of different performance parameters, including note timings, intonation, vibrato rates, and dynamics. The discussion will focus on a case study of intonation in solo soprano performances of Schubert’s ‘Ave Maria’, where both the note onsets/offsets and fundamental frequency were estimated automatically. 

 This talk represents joint work with Ichiro Fujinaga, Dan Ellis, and Michael Mandel.