Use this URL to cite or link to this record in EThOS: https://ethos.bl.uk/OrderDetails.do?uin=uk.bl.ethos.797503
Title: Creating Persian-like music using computational intelligence
Author: Arshi, Sahar
Awarding Body: University of Hull
Current Institution: University of Hull
Date of Award: 2018
Availability of Full Text:
Access from EThOS:
Access from Institution:
Abstract:
Dastgāh are modal systems in traditional Persian music. Each Dastgāh consists of a group of melodies called Gushé, classified in twelve groups about a century ago (Farhat, 1990). Prior to that time, musical pieces were transferred through oral tradition. The traditional music productions revolve around the existing Dastgāh, and Gushe pieces. In this thesis computational intelligence tools are employed in creating novel Dastgāh-like music. There are three types of creativity: combinational, exploratory, and transformational (Boden, 2000). In exploratory creativity, a conceptual space is navigated for discovering new forms. Sometimes the exploration results in transformational creativity. This is due to meaningful alterations happening on one or more of the governing dimensions of an item. In combinational creativity new links are established between items not previously connected. Boden stated that all these types of creativity can be implemented using artificial intelligence. Various tools, and techniques are employed, in the research reported in this thesis, for generating Dastgāh-like music. Evolutionary algorithms are responsible for navigating the space of sequences of musical motives. Aesthetical critics are employed for constraining the search space in exploratory (and hopefully transformational) type of creativity. Boltzmann machine models are applied for assimilating some of the mechanisms involved in combinational creativity. The creative processes involved are guided by aesthetical critics, some of which are derived from a traditional Persian music database. In this project, Cellular Automata (CA) are the main pattern generators employed to produce raw creative materials. Various methodologies are suggested for extracting features from CA progressions and mapping them to musical space, and input to audio synthesizers. The evaluation of the results of this thesis are assisted by publishing surveys which targeted both public and professional audiences. The generated audio samples are evaluated regarding their Dastgāh-likeness, and the level of creativity of the systems involved.
Supervisor: Davis, Darryl N. Sponsor: Not available
Qualification Name: Thesis (Ph.D.) Qualification Level: Doctoral
EThOS ID: uk.bl.ethos.797503  DOI: Not available
Keywords: Computer science
Share: