Guest lecture by Dave Meredith


Title

Discovering translation-invariant patterns in music and other multidimensional datasets

Speaker

  • Dave Meredith
  • Computing Department, City University, London

Time and Place

  • Thursday, November 23, 2000
  • at 12.15
  • Lecture Room A414 (4th floor), Department of Computer Science

Abstract

Many music analysts and music psychologists agree that one of the most important steps in achieving a satisfying interpretation of a musical surface is the identification of important instances of repetition.

Our goal is to build a computational model of expert music cognition and it seems clear that one of the most important components in such a model would be an algorithm that can identify the most important instances of repetition in a musical surface.

Existing algorithms for discovering repeated patterns in musical data are limited in various ways. For example, some can process only monophonic music, others are incapable of finding patterns "with gaps" and so on.

A new algorithm called SIA will be presented that discovers complete sets of translation-invariant patterns ("translational equivalence classes") in multi-dimensional datasets.

SIA takes a multidimensional dataset as input and generates as output a set of "translational equivalence classes" (TECs). Two patterns in a dataset are said to be "translationally equivalent" if and only if one can be obtained from the other by translation alone. Each pattern in a dataset is a member of exactly one TEC and the TEC to which a pattern belongs contains all and only those patterns to which the pattern is translationally equivalent.

SIA generates for each of the largest repeated patterns in a dataset the TEC that contains that pattern. The set of TECs generated by SIA for a musical surface represented as a multidimensional dataset typically contains those instances of repetition that are considered to be musically most interesting.

Brief bio

Dave Meredith is a Research Fellow in the Computing Science Department at City University, London. Since October 1999 he has been working with Geraint Wiggins on an EPSRC funded project whose goal is to develop improved computational models for the processes involved in expert music cognition. In particular, he is interested in modelling the perception of similarity in music.

Dave did his undergraduate degree at Cambridge University where he studied Natural Sciences for one year and then Music for two years. He then defected to "the other place" (Oxford, Music Department) to study for his doctorate. He is still trying to persuade them to give him a PhD in mathematical music theory.


Welcome!


Esko Ukkonen


Anna Pienimäki