When: Monday, June 16th, 2014 @ 5:00 PM – 6:00 PM
Where: Room 4.31/4.33, Informatics Forum, Crichton Street, University of Edinburgh
Measuring and Modelling Intonation and Temperament
Dr. Simon Dixon (QMU, London)
Pitch is perhaps the most important characteristic of musical sounds, being essential for determining both harmony and melody. Western music theory provides a framework for understanding pitch relationships in terms of intervals, scales and chords, expressed as a first approximation in indivisible units of semitones. Common music notation reflects this world-view. At the same time, it has been recognised since the time of Pythagoras that it is not possible for all theoretically consonant intervals to be perfectly “in tune”, and this has led to many theoretical and practical approaches to intonation, the realisation of pitch, in music performance.
In this seminar, I present two investigations of intonation in recorded music at two extremes of musical practice: a fixed-pitch instrument, the harpsichord, where the tuner determines the intonation of each pitch before the performance, and a variable-pitch instrument, the human voice, which can adjust the pitch of each note to the musical context and also vary the pitch over the note’s duration. In each case we have developed software tools for (semi-)automatic analysis of the pitch content of solo audio recordings.
We analysed a collection of solo harpsichord CDs, estimated the inharmonicity and temperament of the harpsichord for each movement, and compared the measured temperaments with those given in the CD sleeve notes. The observed differences illustrate the tension between temperament as a theoretical construct and as a practical issue for professional performers and tuners. We conclude that “ground truth” is not always scientific truth, and that content-based analysis has an important role in the study of historical performance practice.
The second part of the seminar presents a study on intonation and intonation drift in unaccompanied human singing and proposes a simple intonation memory model that accounts for many of the effects observed. Singing experiments were conducted with 24 singers of varying ability. Over the duration of the recordings, approximately 50 seconds, a median absolute intonation drift of 11 cents was observed, which was smaller than the median note error (19 cents), but was significant in 22% of recordings. Drift magnitude did not correlate with other measures of singing accuracy or singing experience. Neither a static intonation memory model nor a memoryless interval-based intonation model can account for the accuracy and drift behaviour observed. The proposed causal model provides a better fit.
(This was joint work with Matthias Mauch, Dan Tidhar and Emmanouil Benetos.)
Simon Dixon is a Reader at Queen Mary University of London, where he leads research on Music Informatics in the Centre for Digital Music. He has a PhD in Computer Science (Sydney) and LMusA diploma in Classical Guitar. His research interests include high-level music signal analysis and the representation of musical knowledge (particularly rhythm and harmony), and has published over 100 refereed papers in the area of Music Informatics. He is currently President of the International Society for Music Information Retrieval.