Tuesday, September 16, 2014

Fwd: Music Informatics Research Group Seminar - Polina Proutskova, 24th September, 12:00

Dear all,

On Wednesday 24th September at 12:00, Polina Proutskova (Goldsmiths,
University of London) will present the seminar "Modelling vocal
production in the world's music cultures".

The talk will take place in room AG04 (College Building) at City
University London, Northampton Square, London EC1V 0HB (campus map:
http://www.city.ac.uk/visit ). More details below:


Title: Modelling vocal production in the world's music cultures

In order to apply computational approaches to analyse, compare and
classify recordings of singing, in particular in large, cross-cultural
collections, dimensions of analysis and criteria for comparison have
to be formalised. This is not an easy task. Singing teachers often use
idiosyncratic terminology for vocal production; medical professionals
mainly assess voice disorders. In voice science, the current state of
knowledge allows for detailed analysis of singing on a short temporal
scale, usually of several vocal-folds vibration cycles; for time
frames of several seconds, as are necessary for stylistic analysis,
the knowledge is still limited. A coherent and comprehensive model of
vocal production is yet to be developed. Even within a single culture,
such as Western music, there is little agreement among professionals
about basic terminology.
Publications in English analysing vocal production in other cultures
are rare. A seminal eth-nomusicological study on cross-cultural
comparison of singing was performed by Alan Lomax and his Cantometrics
team. Vocal production was studied using perceptual descriptors, but
the rating procedure did not take the subjectivity of perceptual
descriptors sufficiently into account.
I'll present an exploratory mixed methods study that investigates
whether vocal physiology can be used for a more objective description
of vocal production in a cross-cultural context. Its results so far
have been rather surprising and raise many questions for various
disciplines, in particular for MIR. We'll discuss its possible impact
on further research on singing as well as on computational approaches
to recordings of singing; and elaborate on whether MIR could help to
discover the relationship between singing and society.

Polina Proutskova is a singing practitioner and researcher: she
performs as a singer in a number of musical traditions, teaches vocal
technique, conducts field research in Russian villages. She is a PhD
candidate at Goldsmiths, her work spanning across the areas of
ethnomusicology, voice science and music informatics. Prior to her PhD
Polina worked with ethnomusicological archives on preservation and
interoperability and coordinated a UNESCO report on musical diversity;
her professional experience includes software development in
bioinformatics and formalising a data management system. Her
educational background is mathematics and computational linguistics at
the universities in St. Petersburg (Russia) and Heidelberg (Germany).
More info: http://goldsmiths.academia.edu/PolinaProutskova


Best wishes,

Emmanouil Benetos, Research Fellow
Department of Computer Science
School of Mathematics, Computer Science and Engineering
City University London
Northampton Square, London EC1V 0HB, UK
Tel: +44 (0)20 7040 4154
e-mail: emmanouil.benetos.1@city.ac.uk