ePrints.FRI - University of Ljubljana, Faculty of Computer and Information Science

Song synthesizer based on the musical features

Lazar Milinković (2012) Song synthesizer based on the musical features. EngD thesis.

[img]
Preview
PDF
Download (692Kb)

    Abstract

    The diploma thesis presents the implementation of a program which synthesizes solo compositions on the basis of predefined attributes. The program calculates their values on a particular track, which it later attempts to emulate with the synthesis of a track with the same attributes. Attribute values are calculated using Matlab library MIRToolbox. Average amplitude, tempo, key, brightness and dissonance are used in synthesis, with the last three attributes defining a set of all possible tones in a track. Synthesis of a piano and a clarinet is implemented. Two more attributes are used for an instrument to be chosen automatically. Based on two methods of measuring, a partial success of the synthesis is observed. Presented are some ideas for improvements in the implementation, which would result in a greater similarity between compositions.

    Item Type: Thesis (EngD thesis)
    Keywords: Attributes, phase vocoder, ADSR envelope, dissonance, hierarchical clustering, instrument synthesis, brightness, tempo, MIRToolbox
    Number of Pages: 40
    Language of Content: Slovenian
    Mentor / Comentors:
    Name and SurnameIDFunction
    doc. dr. Zoran Bosnić3826Mentor
    Link to COBISS: http://www.cobiss.si/scripts/cobiss?command=search&base=50070&select=(ID=00009471316)
    Institution: University of Ljubljana
    Department: Faculty of Computer and Information Science
    Item ID: 1881
    Date Deposited: 26 Sep 2012 13:13
    Last Modified: 26 Oct 2012 08:42
    URI: http://eprints.fri.uni-lj.si/id/eprint/1881

    Actions (login required)

    View Item