Development of a model enabling dance and music interaction with the use of machine learning techniques.

Graduate Thesis uoadl:2937174 15 Read counter

Unit:
Department of Music Studies
Library of the School of Philosophy
Deposit date:
2021-04-01
Year:
2021
Author:
Bakogiannis Konstantinos
Supervisors info:
Αναστασία Γεωργάκη, Καθ., ΤΜΣ, ΕΚΠΑ
Χριστίνα Αναγνωστοπούλου, Αν. Καθ., ΤΜΣ, ΕΚΠΑ
Αρετή Ανδρεοπούλου, Επ. Καθ., ΤΜΣ, ΕΚΠΑ
Original Title:
Ανάπτυξη μοντέλου διαδραστικού χορού και μουσικής με χρήση τεχνικών μηχανικής μάθησης
Languages:
Greek
Translated title:
Development of a model enabling dance and music interaction with the use of machine learning techniques.
Summary:
Technology can retransform the dialogue between dance and music, providing new creative perspectives. In this work, we present the design and development of an interactive model which composes in real-time automated, structurally related to dance, music. A sensor provides the model with data of human motion (input), which are used to gather information regarding motion parameters (e.g., relative distance, speed, direction etc.). The model, trained with a real dataset of human motions and by machine learning techniques, becomes capable of recognizing these parameters. Then, the motion parameters are mapped to sonic parameters (e.g., pitch, duration, audio samples, filters etc.), aiming to create an interactive dance performance. For the implementation, we used appropriate hardware (camera-based Kinect sensor, pc and speakers) and software (TouchDesigner, Wekinator, Max Msp, Max for Live / Ableton Live).
Main subject category:
Fine arts - Entertainment
Keywords:
Algorithmic music, automated music generation, choreomusicology, interactive dance, interactive machine learning, interactive technology, music composition for dance
Index:
Yes
Number of index pages:
1
Contains images:
Yes
Number of references:
42
Number of pages:
68
Bakogiannis.pdf (3 MB) Open in new window