Dear Freesounders,
Welcome to our 2020 Sustainability Report! Here we give an update on how 2020 went for Freesound in terms of sustainability, and present our plans for 2021. You’ll see that most of the information is very similar to previous years’ posts, but still, we think that this will be interesting for you. As usual, the report is split in a number of sections discussing specific aspects that contribute to the sustainability of Freesound, and a final section with a summary, conclusions and future perspectives.
Sound uploads
Sound uploads are an essential part of the sustainability of Freesound. Now that Freesound is reaching 500k sound uploads, it becomes even clearer its immense value, making Freesound an extremely useful resource, attracting many new users every year (both downloaders and uploaders) and certifying the world-wide impact of the Freesound website. Such a big impact is also a great motivation for the research we (and others) carry out around Freesound, and for helping us obtain funding to support it. In 2020, 46,441 new sounds were uploaded. That number is quite similar to previous years (check the evolution of total number of sounds uploaded from the 2020 in numbers blog post). Other related statistics of activity around sounds (ratings, comments, posts) also feature numbers similar to previous years. All in all, we can conclude that Freesound follows a stable (and healthy!) trend in terms of sound uploads.
User donations
In terms of user donations, 2020 has seen a significant increase with respect to 2019. We explained in our 2020 in numbers post that in 2020 the donations increased by around 30%, and that this number correlated very well with the increase of daily visitors that we had right after the COVID19 lock-downs that happened around the world. In 2020, we received 8149 donations (2302 more than in 2019), totaling 69,000? (around 21,000? more than in 2019). These are extremely good news for us. On the one hand, this consolidates the donations model that we introduced in 2017. On the other hand, this shows that Freesound has been a specially useful resource in those strange pandemic times. We thank everyone who donated to Freesound. This is a very important pillar for the sustainability of Freesound and is allowing us to invest more efforts into its maintenance and further development. Most of the development efforts carried out with resources coming from user donations were outlined in the recent December 2020 Community Update post. Due to the increase of donations in 2020, we will be able to boost our development efforts in 2021.
Contribution from UPF
Freesound is an initiative of the?Music Technology Group, a research group of the?Universitat Pompeu Fabra?(UPF) in Barcelona, Spain. In 2020, the UPF contributed to Freesound similarly to previous years. UPF provides the necessary IT infrastructure and basic maintenance (15 servers, 4 TB monthly data bandwidth, IT support staff). As we calculated for the past edition of the sustainability report, the expected cost of such infrastructure if Freesound was hosted in external services (such as Amazon Web services or similar), would be over 20,000?/year (only for hosting costs). Researchers from the university also dedicate time to Freesound related activities (either research, development or administration) and are paid by the university. In 2021, we plan to move all our servers to a new infrastructure also provided by UPF. This should allow us to deploy a much faster and stable Freesound, and solve some of the availability and speed issues we’ve been experiencing in the last year. As you can see, the contribution from UPF is huge, and it is only thanks to the combination of the different aspects discussed in this post that Freesound is sustainable.
Contribution from research grants
Research is at the very core of the Freesound philosophy and, in fact, it is where it all started. We have carried out lots of research activities around Freesound (see some details below), but in 2020 we did not get any new big research grant with a primary role for Freesound (like the AudioCommons project that we coordinated a few-years ago). However, we recently received a 36,000$ grant from the Grant for the Web call to experiment with the application of Web Monetization technologies in the Freesound Licensing project that we’re starting and discussed in a previous blog post. This is funding for development of a prototype, and so it’s not like the projects that we normally do, but we still consider it a research grant. If you’re interested in learning about the research that happens around Freesound (i.e. using Freesound data) not only at the MTG but also around the world, be sure to check the papers section of the Freesound Labs website. But, summarizing, the main research activities that we carried out in 2020 in relation to Freesound are:
- Further development of the FSD50k dataset and publication.
- Further development and maintenance of Essentia, the audio analysis library that powers Freesound sound analysis.
- Research on methods for automatically classifying audio events.
- Research on the analysis of urban soundscape sounds from Barcelona.
- Further research on clustering methods to be potentially applied to Freesound search results. This is getting much closer to release now.
Commercial usage of the Freesound API
Freesound has an API endpoint which allows third parties to develop applications that incorporate Freesound content. Usage of this API is free for non-commercial purposes, while commercial use of the API requires a commercial license. In this way we make sure that commercial applications using Freesound also contribute back to the community. Note that this is independent from the license of the sounds themselves, which need to be respected regardless of the API usage agreement. In 2020 we slightly increased the number of license agreements and the income they generate (~3,000?). We spent this money in the same development efforts described in the?User donations?section above. We’ve observed growing interest for the API so we expect this number to grow in 2021.
Summary and perspectives for 2021
As you can see, in 2020 we improved, in terms of sustainability, compared with previous years. This was been mainly due to the increase in user donations. Thanks to that, we have already started spending more efforts in Freesound development and that will allow us to greatly improve the platform during this year. Freesound is more sustainable than ever, and we have plans to further consolidate our model with the addition of the Freesound Licensing sister-project (which should advance considerably during 2021).
We?d like to finish this post by saying thank you to everyone who contributed to Freesound during 2020, in particular to those who donated and those who uploaded and moderated sounds. We?ll let you know how things go next year in the 2021?s sustainability report.
frederic, on behalf of the Freesound Team
Maybe with all the money you got you can make the preview less squeaky 😀
I am talking about this:
https://freesound.org/help/faq/#i-hear-a-high-pitched-squeaking-in-the-audio-preview-what-is-that
Hi thanks for the comment. We’ve been considering the idea of using better quality previews but as Freesound grows in traffic, bandwidth is something we need to control a lot. Maybe in the future, who knows!
how would one make a donation? I don’t see how. Probab;y write in front of me and just don’t see it
Thank you so much for your efforts! Freesound is a wonderful resource!
Wishing you all the best!
Dovetailing on from my earlier comments on using a ‘bounties’ or ‘challenge’ model for finances, in using Freesound I feel there may be a few optimisations that might help improve performance (thus reducing costs/overheads).
I notice previewed sounds seem to ‘fully load’, rather than using chunk/buffered loading (kind of like what YouTube does). This means effectively a person ‘downloads’ a sound twice – once for preview (given the full load on play) and once again if they like it enough to save-as. This seems like it would consume (and thus cost) a fair bit of bandwidth.
I often find I don’t play the entire previewed sound, and I imagine it’s similar for most people, so only pre-buffering a small chunk initially would be an optimisation. It’d likely improve download speeds across the board as well by reducing preview overheads.
The other observations are in a different ballpark – currently sound format is not standardised, which means the same sound could be stored more efficiently (EG .wav, .raw to a lossless audio codec). This would also save harddrive space and reduce bandwidth overheads. Freesound doesn’t have to eliminate all bar one codec, just transform less efficient codecs to more efficient ones (so post upload, converting .wav to .flac or asking the user to pre-convert).
Freesound could likely retroactively convert the WAVs presently on site to a lossless audio codec as a batch job. In terms of a financial model, it could offer the original WAVs to donors (using the donation to offset the higher bandwidth costs).
Finally, there’s one more trick that appears to be missing – compression in general. Whilst older tools like .zip don’t offer much improvement, .7z offers improvements. Pre-compressing the files trades off on additional harddrive space (as you need the original file for previews and the compressed for downloads) to save on bandwidth, and it depends which one costs more.
I don’t know if you already use it, but you may also wish to make use of HTTP’s underlying gzip compression for streaming to help reduce overheads (https://en.wikipedia.org/wiki/Gzip).
One small issue: attribution requirements by sound uploaders is not standardised, and I’ve seen a few examples of people tagging sounds as CC0 but in the description requiring (not requesting) attribution. You could auto-assign a licence by having an optional attribution field at upload. If it’s left blank, it’s CC0/non-attribution, and if it’s filled in, then the licence is narrowed to a by attribution licence. The field should show an example format for the attribution to help standardise.
All I have for now.
hello, my name is Abby, I’m in school still (as in middle school) and I have no idea what I’m making a comment on, but I was just trying to spread a word of hope to who ever is reading this or ma need this. Remember to enjoy life while you still have it and maybe get in a little trouble, that’s ok (as long as it isn’t nothing to serious). just do it while you live so you don’t regret it later. Any way that’s it, have a good day, good bye.
Ola pendejos, with all due repstekp I want to congratulate you on your job and present thanks you for efforts. You are one of few truly useful open platforms for creators. I wish you all – all the best on the migration to new HW, and every challenge ahead.
Congratulations, and great job you’re doing. I’ve downloaded hundreds of sounds over the years.
Thank you abby for your message of hope and peace. I was feeling a down and bit hopeless today about all the greed in the world but your simple message gave a smile to this old man. Also thank you freesound for all the free sounds, I will be donating soon.
Thank you and congratulations on a job well done. Over the years, I’ve downloaded hundreds of sounds.
I appreciate your message of peace and hope, Abby. Because of all the selfishness in the world, I was feeling depressed and a little hopeless today, but your brief comment made this elderly man smile. I will be making a donation shortly, and I also want to thank Freesound for all the free sounds.