Dataset Open Access

MAESTRO Synthetic - Multi-Annotator Estimated Strong Labels

Irene Martin Morato; Manu Harju; Annamaria Mesaros

Citation Style Language JSON Export

  "publisher": "Zenodo", 
  "DOI": "10.5281/zenodo.5126478", 
  "author": [
      "family": "Irene Martin Morato"
      "family": "Manu Harju"
      "family": "Annamaria Mesaros"
  "issued": {
    "date-parts": [
  "abstract": "<p>The dataset was created for studying estimation of strong labels using crowdsourcing.</p>\n\n<p>It&nbsp;contains 20 synthetic audio files created using Scaper, the reference annotation created with Scaper, and the annotation outcome.&nbsp;Annotation was performed using Amazon Mechanical Turk.</p>\n\n<p>Audio files contain excerpts of recordings uploaded to Urban Sound 8k dataset).&nbsp;Please see FREESOUNDCREDITS.txt for an attribution list.&nbsp;</p>\n\n<p>The dataset contains:&nbsp;</p>\n\n<ul>\n\t<li>audio: the 20 synthetic soundscapes, each 3 min long</li>\n\t<li>ground truth:&nbsp; the &quot;true&quot; reference annotation created using Scaper</li>\n\t<li>estimated strong labels: the reference annotation created from the crowdsourced data</li>\n\t<li>audio tags: the weak labels corresponding to each 10 s segment of the soundscapes, as annotated</li>\n</ul>\n\n<p>For details on the annotation procedure and label processing methodology, see the following paper:</p>\n\n<p>Irene Martin&nbsp;Morato, Manu Harju, and Annamaria Mesaros.&nbsp;<em><a href=\"\">Crowdsourcing strong labels for sound event detection</a>.</em>&nbsp;In IEEE Workshop on Applications of Signal Processing to Audio and Acoustics (WASPAA 2021). New Paltz, NY, Oct 2021.</p>\n\n<p>&nbsp;</p>\n\n<p>&nbsp;</p>", 
  "title": "MAESTRO Synthetic - Multi-Annotator Estimated Strong Labels", 
  "type": "dataset", 
  "id": "5126478"
All versions This version
Views 1111
Downloads 2222
Data volume 1.2 GB1.2 GB
Unique views 1010
Unique downloads 44


Cite as