How Does Music Expression Depend on Structure?

In what way does music expression depend on musical structure? The question may be clarified by examining how leading performers segment musical phrases, as well as how they agree on selection and emphasis of local events (accents). Accents are local events that attract a listener’s attention and may be immanent (grouping, metrical, melodic, harmonic) or performed (variations in timing, dynamics, and articulation) (Parncutt, 2003). First, I will discuss a new computational model of accent salience in Western tonal music that follows a top-down approach (Bisesi, Friberg & Parncutt, under submission; Friberg & Bisesi, 2014; Parncutt, Bisesi & Friberg, 2013; Bisesi & Parncutt, 2011). By combining the accent theory of Parncutt (2003) with the performance rendering system “Director Musices” of Friberg, Sundberg and Bresin (2006), we estimated the positions and saliences of metrical, melodic and harmonic accents separately, and compared our predictions with the results from two different experiments involving musicians and music theorists, respectively. We are also extending DM in a new direction, which allows us to relate expressive features of a performance not only to global or intermediate structural properties (phrasing), but also accounting for local events (accents) (Friberg & Bisesi, 2014). Second, our model is being investigated in two different ways: perceptually – by comparing predicted versus perceived accent saliences in eminent performances for a selection of Chopin Preludes (Bisesi, MacRitchie & Parncutt, 2012), and physically – by measuring variations in timing and dynamics as performed by eminent musicians (Bisesi & Cabras, in preparation). Third, 16 different performances of the Preludes are clustered in terms on phrasing profiles and accents, and each cluster considered as an interpretative style (Bisesi & Cabras, in preparation). Fourth, I will present a new project that is currently running at the KTH, Stockholm (Bisesi, Fribeg, Addessi & Baroni, in preparation). In this project, our model is being revised by following a bottom-up approach, and a new model of immanent accent salience will be derived by means of machine learning from experimental data, rather than from purely theoretical principles, using also post-tonal music. All the preliminary stages of melody selection, coding and recording, computer interface designing, and preliminary data collection will be discussed in the presentation.

Keywords: computational music analysis – music performance – musical accents – Director Musices – MIR

Erica Bisesi was born in Gorizia (Italy). She completed a Ph.D. in Mathematics and Physics at Udine University in 2007. She has taught acoustics and psychoacoustics at the Udine Conservatory from 2004 to 2006, and psychoacoustics and music cognition at the University of Graz from 2014 to 2015. Her career as a systematic musicologist began in 2007, first at the Centre for Systematic Musicology at the University of Graz, and then in several projects on the psychology of music, psychoacoustics, expressive music performance, music theory and analysis and music information retrieval, collaborating with the Conservatorio della Svizzera Italiana in Lugano, the Kunstuniversität in Graz, the Conservatories of Como and Novara, the Universities of Jyväskylä, Bologna, Padova, Udine and Montréal. In October 2009, Erica was awarded a Lise Meitner postdoctoral fellowship for a two-year project entitled Measuring and modeling expression in piano performance by FWF Austria. In December 2011, FWF funded her three-year Stand-Alone project Expression, emotion and imagery in music performance. Erica received her first music education at the age of five, and completed a M.A. Degree in Piano Performance at Trieste Conservatory in 1996. Over the following ten years, she studied with the conductor F. Mander, the pianists B. Canino in Milan, A. Delle Vigne in Salzburg, Florence and Rome, V. Krpan in Zagreb, A. Kravtchenko in Rovereto, and A. Woyke in Graz. She is now active as a soloist and in chamber music ensembles. Erica is currently a senior postdoctoral researcher at the KTH in Stockholm, and lecturer on psychoacoustics, music cognition, music expression and emotion at the Comenius University in Bratislava. Her research lies mainly in the area of computational musicology, music performance, expression and emotion, and music theory and analysis.