Back Freesound’s coming of age

Freesound’s coming of age

Freesound’s coming of age
To further consolidate the Freesound platform and its unique potential for research on occassion of its 20th anniversary.

Freesound is one of the most important and highest-impact projects of the MTG and the DTIC/UPF. Initiated by the MTG in 2005, the 20th anniversary of Freesound is approaching. During all these years, Freesound has been consolidated as the world’s biggest platform for sharing audio clips under Creative Commons licences, currently hosting nearly 600k sounds and having more than 13 million registered users who download 70k sounds every day. These numbers are a clear indication of Freesound’s big impact on society. But beyond the general public, Freesound has also had a widespread impact both on the research community and the industry. We can count more than 600 papers (only a small fraction of them from the MTG) that cite one of the main Freesound reference papers, and which deal with different research topics. Also, a plethora of companies both large and small use the Freesound API to integrate Freesound into their products.

Over the years, Freesound has been sustained through a combination of funds from various related research projects, funds from UPF (mainly in the form of infrastructure), and occasional external grants and awards received from the industry. As part of the actions taken during the previous DTIC MdM program (focused on promoting open science), we were able to further develop a sustainability model for Freesound that would include a campaign for user donations. Since 2017 (when we started the donations campaign) we have received more than 337k unique donations from users all over the world. That has allowed us to spend more efforts on Freesound and provides a great sustainability basis for the future. The previous MdM action also allowed us to largely increase Freesound’s impact on the research community by partially supporting the development of the Freesound Annotator, a platform for the crowd-sourced annotation of Freesound content, and the creation of FSD50k, an open dataset of human-labelled sound events. That allowed us to use Freesound for carrying out state of the art research on topics such as sound clustering and classification.

Now that Freesound is coming of age and is approaching its 20th anniversary, we believe it is time to further consolidate the platform and its unique potential for research. To this end, we have set a number of objectives that maintain a tight relation with the current MdM program. These include to O1) carry out state of the art audio research taking advantage of the privileged context provided by Freesound; O2) carry out dissemination activities both within UPF and outside UPF to promote research collaborations and technology transfer; and O3) maximise Freesound’s impact in relation to UPF’s Planetary Wellbeing initiative.

Principal researchers

Xavier Serra

Researchers

Frederic Font
Panagiota Anastasopoulou
Sonia Espí

The project will be supported by the PhD Fellowship program at the Department of Information and Communication Technologies at UPF, and with Freesound funds to support coordination, development and dissemination.