A framework for the real-time analysis of musical events

Download

PDF Download (694Kb)

Description/Abstract

In this thesis I propose a framework for the real-time creation of a harmonic structural model of music. Unlike most uses of computing in musicology which are based on batch processing, the framework uses publish/subscribe messaging techniques
found in business systems to create an interconnected set of collaborating applications within a network that process streamed events of the kind generated during a musical
performance. These applications demonstrate the transformation of data in the form of MIDI commands into information and knowledge in the form of the music’s harmonic
structure represented as a model using semantic web techniques.

With such a framework, collaborative performances over the network become possible with a shared representation of the music being performed accessible to all performers both human and potentially software agents. The framework demonstrates
novel real-time implementations of pitch spelling, chord and key extraction algorithms interacting with semantic web and database technologies in a collaborative manner. It draws on relevant research in information science, musical cognition, semantic web and business messaging technologies to implement a framework and set of software components for the real-time analysis of musical events, the output of which is a
description of the music’s harmonic structure. Finally, it proposes a pattern based approach to querying the generated model which suggests a visual query and navigation paradigm.