Please use this identifier to cite or link to this item: https://hdl.handle.net/10356/44291
Title: Music generation (using markov method)
Authors: Soh, Poh Kuan.
Keywords: DRNTU::Engineering::Computer science and engineering
Issue Date: 2011
Abstract: Beyond solving daily logical problems, this project seeks to employ Artificial Intelligence in music composition, attempting to venture into the deeper regions of understanding of both the cognitive AI and the creative human mind. The report describes the design of an automated music composition system called “Markov Music Generator”, which composes music melodies based on a finite set of existing music phrases designed by a music expert. Using “Markov Chains” as a machine-learning method, the system creates original melodies of the ‘Contemporary Pop’ genre by manipulating the pitch and rhythmic components of music. Modeling after today’s Pop music’s structure, the system is trained by inputs that are classified according to their respective chords. The outputs are then produced by the retrieval of previously trained probabilistic information from Markov’s State Transition Matrices and relevant tables, and generated according to the pre-written chords-sequence, in a successive manner. Within this report, results are presented to show that employing 1st and 2nd Order Markov Chains for the pitch component is able to generate new melodic sequences that are pleasant and diverse, and also able to match the style of the inputs. A separate test setup shows the employing only 1st Order Markov Chains produces better results than using both 1st and 2nd Order ones. Strengths and limitations will also be reviewed. Finally, the report concludes with the discussion of future works that may improve and increase the creativity capacity of the current “Markov Music Generator”.
URI: http://hdl.handle.net/10356/44291
Rights: Nanyang Technological University
Fulltext Permission: restricted
Fulltext Availability: With Fulltext
Appears in Collections:SCSE Student Reports (FYP/IA/PA/PI)

Files in This Item:
File Description SizeFormat 
SCE10.pdf
  Restricted Access
18.87 MBAdobe PDFView/Open

Page view(s) 50

375
Updated on Apr 18, 2021

Download(s) 50

29
Updated on Apr 18, 2021

Google ScholarTM

Check

Items in DR-NTU are protected by copyright, with all rights reserved, unless otherwise indicated.