Dataset Open Access

MAST rhythm re-annotated subset

Falcão, Felipe; Bozkurt, Baris


Citation Style Language JSON Export

{
  "publisher": "Zenodo", 
  "DOI": "10.5281/zenodo.2619499", 
  "author": [
    {
      "family": "Falc\u00e3o, Felipe"
    }, 
    {
      "family": "Bozkurt, Baris"
    }
  ], 
  "issued": {
    "date-parts": [
      [
        2019, 
        4, 
        2
      ]
    ]
  }, 
  "abstract": "<p>In the work described in the paper <em><a href=\"http://archives.ismir.net/ismir2019/paper/000052.pdf\">A dataset of rhythmic pattern reproductions and baseline automatic assessment system</a></em><a href=\"https://ismir2019.ewi.tudelft.nl/?q=accepted-papers\"><em>&nbsp;(ISMIR, 2019)</em></a>, Falc&atilde;o et al. applied efforts on annotating a subset of the full <a href=\"https://zenodo.org/record/2620357#.XKLB39szZuQ\">MAST rhythm&nbsp;dataset</a> with grades for a subset of performances.</p>\n\n<p>Out of a&nbsp;total of 1040 student submissions comprised by the original dataset, 80 performances (equally distributed between 20 references)&nbsp;were graded by annotators using a custom annotation tool. During this process annotators could play each&nbsp;<em>(reference, performance)</em>&nbsp;pair as many times as desired before deciding on&nbsp;one of the available assessments:&nbsp;<em>4-Perfect, 3-Minor errors, 2-Major errors, 1-Completely off.</em></p>\n\n<p><strong>Naming convention</strong></p>\n\n<p>This dataset is a collection of&nbsp;audio files, onset features and&nbsp;user-annotated data. Audio files are stored into the <em>Only Performances</em> and <em>Only References </em>folders, together with some extra files containing their original onsets, quantized onsets and scaling parameters&nbsp;(for more information regarding this auxiliary data, please check the paper). An indexed list of files (<em>listreferences </em>and<em> listperformances</em>) is also stored in&nbsp;each of these folders in order to allow for mapping between audio files and auxiliary data. Example:</p>\n\n<ul>\n\t<li>The 10th line&nbsp;in&nbsp;<em>Only Performances/MAST Onsets [Performances]&nbsp;</em>contains the pure onsets for the 10th audio file listed in the&nbsp;<em>Only Performances/listperformances&nbsp;</em>file (&#39;55_rhy2_per138459_pass.wav&#39;)</li>\n</ul>\n\n<p>Also, in order to check the reference which&nbsp;relates to a specific performance, one must also look for audio file names in the list of files: the <em>i-eth</em> audio file listed in&nbsp;<em>Only Performances/listperformances&nbsp;</em>is a specific performance whose reference is defined by&nbsp;the<em> i-eth</em>&nbsp;line of&nbsp;<em>Only References/listreferences.&nbsp;</em>Example:</p>\n\n<ul>\n\t<li>The 10th audio file listed in&nbsp;<em>Only Performances/listperformances&nbsp;</em>file (&#39;55_rhy2_per138459_pass.wav&#39;) is a performance whose reference is stored in the 10th line of&nbsp;<em>Only References/listreferences&nbsp;</em>(55_rhy2_ref162559.wav)</li>\n</ul>\n\n<p>Such convention is also applied to the folder that contains the annotations (<em>Performances Annotations</em>). Each one of the&nbsp;<em>annotator_<strong>X</strong>.txt </em>file lists the grades assigned by the <em><strong>X</strong>-ith</em> annotator, and the mapping between the assessments and their referred performances is indexed in&nbsp;<em>listfiles.txt</em></p>", 
  "title": "MAST rhythm re-annotated subset", 
  "type": "dataset", 
  "id": "2619499"
}
155
8
views
downloads
All versions This version
Views 155155
Downloads 88
Data volume 282.5 MB282.5 MB
Unique views 136136
Unique downloads 88

Share

Cite as