|Archiving status||Not saved yet|
|IRC channel|| (on hackint)|
(formerly #freeallthesounds (on EFnet))
Freesound is a website for free audio material exchange, founded through the International Computer Music Computer Conference in 2005.
A free account is required to download sounds on the website. Uploaders can choose licensing agreements for their sound, CC0, CC BY, and CC BY-NC - but emphasis is placed on collaborative potential of open sounds.
User-specific generated keys are used to access Freesound's API. The API tools search, retrieve metadata, retrieve comments and generate download links for sounds - this could possibly be used to scrape and archive the site.
According to Wikipedia ~230,000 sounds were present on the site as of February 2015. There were an estimated ~1TB of sounds in 2014.
A 2016 machine learning project available on Kaggle aggregated "230,000 sounds and the associated tags" from the Freesound database using the aforementioned API.
The Kaggle dataset alone does not contain any sound samples, but the provided sounds.json file contains 230,000 Freesound links to download ~192kbps sounds in both ogg and mp3 format with associated tags. Data such as title, description, uploader and comments are missing, but the file URLs may be a useful starting point for archival.
The website's own browser of sounds lists more than 420,000 items as of August 2019.
The site shows no immediate signs of going under.
It seems at one point the site relied on university-provided research hosting, and has had a troubled past with downtime. In June 2008 the site's domain switched from freesound.iua.upf.edu to freesound.org - since that point reliability has improved.
It seems there was some fairly in-depth discussion about the process and politics of archiving Freesound in #archiveteam IRC May 2016. However, users dispanded to IRC channel '#freeallthesounds' which isn't archived - it is unclear whether progress was made.
- • • •
- • • •