Thesis Open Access

From heuristics-based to data-driven audio melody extraction

Bosch, Juan J.


DataCite XML Export

<?xml version='1.0' encoding='utf-8'?>
<resource xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance" xmlns="http://datacite.org/schema/kernel-4" xsi:schemaLocation="http://datacite.org/schema/kernel-4 http://schema.datacite.org/meta/kernel-4.1/metadata.xsd">
  <identifier identifierType="DOI">10.5281/zenodo.1120334</identifier>
  <creators>
    <creator>
      <creatorName>Bosch, Juan J.</creatorName>
      <givenName>Juan J.</givenName>
      <familyName>Bosch</familyName>
      <nameIdentifier nameIdentifierScheme="ORCID" schemeURI="http://orcid.org/">0000-0003-4221-3517</nameIdentifier>
      <affiliation>Universitat Pompeu Fabra, Barcelona</affiliation>
    </creator>
  </creators>
  <titles>
    <title>From heuristics-based to data-driven audio melody extraction</title>
  </titles>
  <publisher>Zenodo</publisher>
  <publicationYear>2017</publicationYear>
  <subjects>
    <subject>Melody Extraction</subject>
    <subject>Automatic</subject>
    <subject>MIR</subject>
    <subject>Music</subject>
    <subject>Retrieval</subject>
    <subject>Symphonic</subject>
    <subject>Instrument</subject>
    <subject>Agreement</subject>
    <subject>Tonality</subject>
    <subject>Timbre</subject>
    <subject>Stereo</subject>
    <subject>Source-filter</subject>
    <subject>Separation</subject>
    <subject>NMF</subject>
    <subject>Visualisation</subject>
    <subject>Evaluation</subject>
    <subject>Dataset</subject>
    <subject>Contour</subject>
    <subject>Salience</subject>
    <subject>Pitch</subject>
    <subject>Supervised</subject>
  </subjects>
  <contributors>
    <contributor contributorType="Supervisor">
      <contributorName>Gómez, Emilia</contributorName>
      <givenName>Emilia</givenName>
      <familyName>Gómez</familyName>
      <affiliation>Universitat Pompeu Fabra, Barcelona</affiliation>
    </contributor>
  </contributors>
  <dates>
    <date dateType="Issued">2017-06-27</date>
  </dates>
  <language>en</language>
  <resourceType resourceTypeGeneral="Text">Thesis</resourceType>
  <alternateIdentifiers>
    <alternateIdentifier alternateIdentifierType="url">https://zenodo.org/record/1120334</alternateIdentifier>
  </alternateIdentifiers>
  <relatedIdentifiers>
    <relatedIdentifier relatedIdentifierType="URL" relationType="IsIdenticalTo">http://mtg.upf.edu/node/3737</relatedIdentifier>
    <relatedIdentifier relatedIdentifierType="DOI" relationType="IsVersionOf">10.5281/zenodo.1120333</relatedIdentifier>
    <relatedIdentifier relatedIdentifierType="URL" relationType="IsPartOf">https://zenodo.org/communities/mdm-dtic-upf</relatedIdentifier>
    <relatedIdentifier relatedIdentifierType="URL" relationType="IsPartOf">https://zenodo.org/communities/mir</relatedIdentifier>
  </relatedIdentifiers>
  <rightsList>
    <rights rightsURI="https://creativecommons.org/licenses/by-nc-nd/4.0/legalcode">Creative Commons Attribution Non Commercial No Derivatives 4.0 International</rights>
    <rights rightsURI="info:eu-repo/semantics/openAccess">Open Access</rights>
  </rightsList>
  <descriptions>
    <description descriptionType="Abstract">&lt;p&gt;&lt;strong&gt;Abstract&lt;/strong&gt;&lt;/p&gt;

&lt;p&gt;The identification of the melody from a music recording is a relatively easy task for humans, but very challenging for computational systems. This task is known as &amp;quot;audio melody extraction&amp;quot;, more formally defined as the automatic estimation of the pitch sequence of the melody directly from the audio signal of a polyphonic music recording. This thesis investigates the benefits of exploiting knowledge automatically derived from data for audio melody extraction, by combining digital signal&amp;nbsp;processing and machine learning methods. We extend the scope of melody extraction research by working with a varied dataset and multiple definitions of melody. We first present an overview of the state of the art, and perform an evaluation focused on a novel symphonic music dataset. We then propose melody extraction methods based on a source-filter model and pitch contour characterisation and evaluate them on a wide range of music genres. Finally, we explore novel timbre, tonal and spatial features for contour characterisation, and propose a method for estimating multiple melodic lines. The combination of supervised and unsupervised approaches leads to advancements on melody extraction and shows a promising path for future research and applications.&lt;/p&gt;

&lt;p&gt;&amp;nbsp;&lt;/p&gt;

&lt;p&gt;&lt;strong&gt;Datasets:&amp;nbsp;&lt;/strong&gt;&lt;br&gt;
&lt;br&gt;
The symphonic music dataset proposed in this thesis (Orchset) is available at:&lt;/p&gt;

&lt;p&gt;&lt;a href="https://zenodo.org/record/1289786#.XnNV15P0mL8"&gt;https://zenodo.org/record/1289786#.XnNV15P0mL8&lt;/a&gt;&lt;/p&gt;

&lt;p&gt;Orchset is intended to be used as a dataset for the development and evaluation of melody extraction algorithms. This collection contains 64 audio excerpts focused on symphonic music. with their corresponding annotation of the melody.&lt;/p&gt;

&lt;p&gt;&lt;strong&gt;Code:&lt;/strong&gt;&lt;/p&gt;

&lt;p&gt;The source code of the melody extraction algorithms proposed in this thesis is available at:&lt;/p&gt;

&lt;p&gt;&lt;a href="https://github.com/juanjobosch/SourceFilterContoursMelody"&gt;https://github.com/juanjobosch/SourceFilterContoursMelody&lt;/a&gt;&lt;/p&gt;</description>
  </descriptions>
</resource>
128
124
views
downloads
All versions This version
Views 128128
Downloads 124124
Data volume 1.3 GB1.3 GB
Unique views 107107
Unique downloads 111111

Share

Cite as