Χρήση πρωτότυπων μεθόδων διάδρασης με Η/Υ για τη σύνθεση και την εκτέλεση ηλεκτροακουστικής μουσικής
Date Issued
April 30, 2020
Type
Διδακτορική Διατριβή
Abstract
This thesis explores the interaction between humans and machines during musical performance and its effect on electroacoustic music composition and improvisation. It starts by exploring types of interaction following the instrument paradigm, where the performer interacts with the sound synthesis parameters through an interface in real time. In this context, a series of novel compositions and instruments-interfaces are presented that allow or encourage a mode of performance where embodiment plays an important role. These applications are supported by software that proposes innovative solutions in the direction of: (a) a dynamic gesture mapping model where mapping is itself variable and part of the development of the composition; (b) a complete software package for the organization and management of electroacoustic music concerts, using machine learning as an alternative to standard panning algorithms; (c) a multi-buffer granulator using real- time or sampled sources which features statistical distribution methods for the generation of grains. Finally, the role of self-organizing systems with a degree of understanding of their environment, responding to it autonomously in relevant ways through the creative use of machine learning techniques is explored. An application of an autonomous improvisation system which is trained during live concert, featuring innovative elements in multilateral interaction between human and machine agents, as well as introducing 'personality' elements into the latter, is proposed.
Subjects