A platform for the collaborative creation of open audio collections labeled by humans and based on Freesound content

Get started with our annotation tasks!

Explore, download and contribute to our datasets

The FSD is a large-scale, general-purpose dataset composed of Freesound content organised by the AudioSet Ontology

... more to come!

Our vision ...


The datasets creation process should be as transparent
as possible since data science results heavily depend on it


Data should be released under open licenses to
facilitate its distribution and help reproducible research


Datasets should be dynamic in the sense that
they can evolve over time in quality and quantity

... and our plan!

Freesound content

We leverage Freesound (>360k sounds) as a
source of open audio content for creating datasets


We rely on the community to sustainably crowdsource high-quality
sound annotations and to improve our platform workflows

The Freesound Datasets platform

Builds and maintains datasets following the outlined principles of
transparency, openness, dynamic character, and sustainability

What can I do in this platform?


Tools for exploring
the datasets


Datasets realesed under
Creative Commons licences


We collect annotations
through crowdsourcing


About the datasets
and the platform


Amend errors in
the datasets

If you use this platform or any of the hosted datasets in your work please cite our paper:

E. Fonseca, J. Pons, X. Favory, F. Font, D. Bogdanov, A. Ferraro, S. Oramas, A. Porter & X. Serra. Freesound Datasets: A Platform for the Creation of Open Audio Datasets.
In Proceedings of the 18th International Society for Music Information Retrieval Conference, Suzhou, China, 2017 [BiB]