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

Melody extraction from recordings of slovenian folk songs

Anže Mikec (2013) Melody extraction from recordings of slovenian folk songs. EngD thesis.

Download (973Kb)


    The goal of this bachelor's thesis is to analyse existing algorithms and methods for musical feature extraction, implement the most appropriate algorithm and test it against the given test data. Implemented algorithm could then be used for simple melody transcription of sung Slovenian folk songs, if the results would have shown to be promising. After literature review, YIN algorithm was chosen for pitch detection. Segmentation was performed by a library of implemented algorithms called MIRToolbox and estimated values were post-processed with a simple probabilistic model, based on the KernScores library. The final product was fitted with a user interface that enables users to choose a recording they wish to analyse and displays the results when the algorithm is finished. A user can then listen to the recording and simultaneously track its position on a graph.

    Item Type: Thesis (EngD thesis)
    Keywords: automatic music transcription, YIN algorithm, MIRToolbox, autocorrelation, musical features extraction, pitch detection
    Number of Pages: 46
    Language of Content: Slovenian
    Mentor / Comentors:
    Name and SurnameIDFunction
    doc. dr. Matija Marolt271Mentor
    Link to COBISS: http://www.cobiss.si/scripts/cobiss?command=search&base=50070&select=(ID=10097748)
    Institution: University of Ljubljana
    Department: Faculty of Computer and Information Science
    Item ID: 2131
    Date Deposited: 06 Sep 2013 15:57
    Last Modified: 16 Sep 2013 09:31
    URI: http://eprints.fri.uni-lj.si/id/eprint/2131

    Actions (login required)

    View Item