Freesound Forums

Hi. I've used FreeSound for years and have donated a little, but I've never posted here. What I'm going to say will probably be mildly controversial amongst the audio professionals here, but please hear me out.

1. Most users of FreeSound are, like me, needing foley for a project.

We wildly range in variety, from seasoned audio engineers who are trying to avoid a need for fieldwork to record to people trying to source sounds for their first attempt at a video game. To say that FreeSound's contribution to the world's content is huge is an understatement; this is one of the key content locations on the Internet.

2. FreeSound's bandwidth is obviously a huge part of the cost of operating it. I have no idea who provides most of the funding that keeps FreeSound operating- I presume it's Universitat Pompeu Fabra- but we should all be aware that it's making use of a limited public / private resource, and our donations probably do not cover more than a fraction of the real costs.

3. Based on 1 and 2, I have a suggested idea for technology and policy that I would like to propose.

A. FreeSound should have an explicit, two-tier system for bandwidth, like Nexus Mods does, where paying members get a different bandwidth allocation than free-tier users do.

The cost should be pretty reasonable; let's say $20 a year. The number of people wanting to sign up for this would be fairly small at first, but I suspect it would grow, if users were aware of how it works and why it helps sustain the project financially.

I think it would work better than the Donate button, in terms of real-world revenue; those of us who use assets from here often (in my case, I'm probably downloading foley a few times a month, in bursts) would be primary targets (in my case, eating $20/year for a FS subscription would just be a business expense).

B. FreeSound should provide only OGG-compressed sound for free-tier users, and should have OGG using the highest-quality settings available to all users, created upon upload.

This would save a terrific amount of bandwidth, speed up the servers dramatically, and in most cases, would still be at an entirely acceptable level of quality.

I understand that both of these suggestions will be met with reservations.

I am not claiming to be some famous person or an important academic or an expert on the provision of free resources on the Internet. I do think, however, that FreeSound's mission is very important, and that people like me would be largely unable to do our work without it, and I have noticed that users uploading here use wildly different levels of quality on their field recordings and synthesized sounds, and that causes dramatically different amounts of load on FreeSound's bandwidth and servers.

I don't want to take away high-end field recordings from users who need them by requiring everything to be compressed, but I do think that considering either walling it off from free users or at the very least, having two download options- the compressed version being the "recommended option"- would dramatically improve FreeSound's ability to operate over time, and having a paid subscription that gave users access to more bandwidth would strongly incentivize professional users to support the project financially without creating a paywall barrier for people who lack the means to support the project.

At any rate, those are my thoughts, after years of using this resource. I love that this exists, and I would like it to sustainably remain an important resource for countless creatives.

Hi there
I don't think paid tier for lossless is in the spirit of freesound.
First question to ask is freesound currently sustainable?
There's sustainability reports for the previous years if you're interested.
I beleive the admins would let the community know if otherwise. So it doesn't seem like a problem.
In-fact, I think there was ideas of bumping up the preview quality ?

I try to compress .ogg on longer sounds, and use flac or wav for short foley. Being mindful of storage data vs quality when it's needed.
First of all, there's probably a part to do from the users.

Similarly if freesound would offer a 2 tier system, they'd have more expenses, having 3 copies of each sound to store (original, .ogg low teir, .mp3 preview) and convert the audio with another pass.
Probably a subscription would outweigh the cost though.

But I shouldn't even fight on that front, real question is, is this a problem in the first place?

Cheers.

I don't want to cause any drama. People like you are why this is an amazing project.

Here's my take, as a user:

What I've experienced lately is that a lot of the field recordings in WAV, FLAC, etc. are now running to well over 100MB in size, and, at FreeSound's current per-user bandwidth allocation, are taking well over an hour to access. That's just one sound; typically, when I'm here, I'm looking at dozens, to find what I'll need to remix.

This is a problem that's been growing over the last few years, as uploaders get more and more access to high-bandwidth Internet access and conclude that they may as well upload their recordings in the original, uncompressed formats.

I get the impression that FreeSound's bandwidth is either being throttled per user or that with the growth of the platform, it's becoming more of an issue in terms of costs.

I'd be happy to pay a reasonable fee to not have that problem and support FreeSound's continued existence, and I doubt if I'm the only one.

*************

FreeSound's financial report, pinned above, indicates that they're facing a $100K euro deficit. That's surely causing some problems.

Donations probably won't bridge that gap very well; very few users can afford to donate much, and those of us who can have all of the same problems seen at similar resources on the Web, like Wikipedia, where they have to run donation drives on a regular basis simply to get our attention.

Thus a suggestion to use a funding model that I know works: Nexus Mods is, like FreeSound, a huge operation, in terms of bandwidth and file-serving. They probably require 100X FreeSound's bandwidth and storage is probably 1000X. Their two-tier system means that non-paying users aren't frozen out of the service, but paid users get much better bandwidth and can download more files at once.

But they're profitable and healthy and can afford full-time maintenance staff, not merely a few grad students or professors on their spare time. I don't care if FreeSound is profitable, but I do care if it goes broke or is suddenly unmaintained.

The owner of Nexus is fairly easy to contact and is a nice guy, and I've talked to him a couple of times over the years as it went from being a fairly-modest website to the behemoth it is today.

FreeSound's board could get in touch and discuss this with them to see if the tiered model might be a good idea for them. I understand that, as an operation tied to a university, they may face some problems justifying changes without strong data to support their proposals.

*************

Having all sounds converted to OGG upon upload, and being offered as the first choice for users would significantly impact bandwidth requirements.

The change in storage requirements would be fairly modest. Somebody at FreeSound would have to write a routine to do OGG encoding upon upload, which isn't that big of a computational expense vs. providing the file access on a daily basis.

For pretty much all users, the experience would be, essentially, the same.

Unless you're a professional audio engineer needing a raw sound recording for a pretty specific purpose that's being used on a lossless medium (i.e., you're using a field recording done in WAV for something on a CD or vinyl, etc.), having the sounds in uncompressed formats doesn't make much sense.

99% of the use cases for FreeSound assets will be compressed before end-users ever hear it.

If a user's working on a presentation or needs a sound effect to throw into a YouTube video, or are remixing the sounds for a video game mod, etc., a tiny difference in quality won't be apparent. Unity, GameMaker, Unreal Engine, etc.- all of them compress sound files by default.

YouTube and every other streaming service, including professional ones, are all compressing audio. So, while there are arguments for having the original, uncompressed sound available... I don't think it should be the typical user's first choice, to be honest. OGG at a high bitrate will be fine.

I don't see this turning to drama, I see your well intended to freesound and the community and respect that!

I'm someone who can't distinguish mp3 320kbps to wav, and I know ogg is the superior compression format.
I see where you're comming from.

I think, lossless is still good for the average user, some people remixes sounds and gives back a alt version to the community sometimes better. Sometimes just differemt (slowed down whoosh)
It would be a pitty to lose on that. Including for professionals.

If freesound is at a deficit, and going with your idea, I think a middle ground should be struck where either lossless is available to everyone, just not the default option, or it'd available for samples say under 3 min, or the "lower tier" has a limited lossless download per day, like 5.

But if freesound needs money, I think alternatives should be explored first, like making the donation button more at the forefront, having better reports of costs and donations (Lichess does this well, staff, server maintenance costs, and donations income is all listed publicly).

Part of the beauty here, is its free for all, its a rare gem of the internet and I'd like the service to remain this way.
It's like if Wikipedia started having a summary tier and a full page tier it just wouldn't feel the same.

Sincerely, equal service for all, and donation model is my utopia.
Lichess, Freesound, Wikipedia are the very few I can think of like this and im so freaking proud we have that!

Kind regards.

Hey !
Interesting discussion !
Personally, I rather agree with Sadiquecat.
Tell me if I’m wrong, but I think FS has been thought as a free resource from the beginning.
As so, it can be used by everyone, including hobbyist users, sound professionals, as well as researchers, and I think it’s fine like that.
Of course, I deeply agree that it’s very important to keep the website running, in terms of money, and in terms of usability, but choosing to discriminate users according to their ability to pay doesn’t seem like to be the right solution for me.
Furthermore, what about these users who strongly contribute to the community, uploading thousands of sounds, or sharing their valuable experience on the threads ? Should they pay to get access to uncompressed file, even if they already contribute a lot by their work or knowledge ?
I would be quite interested to hear the opinion of the FS team and community about this matter...

Wishing you all the best !

Kevin

As @Kevip888 argues, this site should not have borders.
Freesound is basically a free sound database for scientists, proffesionals and amateurs.
For me, it is therefore the reason I share so much high-resolution sound. I also add a lot of information to my sounds for that reason.

Creating a financial barrier for less wealthy people and organizations would be very disappointing to me.

I make everything available to the community for free. Let everyone benefit from that for free. In these confusing times, there are enough difficult boundaries to overcome.

kind regards,

xenoargh wrote:
I don't want to cause any drama. People like you are why this is an amazing project.

Here's my take, as a user:

What I've experienced lately is that a lot of the field recordings in WAV, FLAC, etc. are now running to well over 100MB in size, and, at FreeSound's current per-user bandwidth allocation, are taking well over an hour to access. That's just one sound; typically, when I'm here, I'm looking at dozens, to find what I'll need to remix.

This is a problem that's been growing over the last few years, as uploaders get more and more access to high-bandwidth Internet access and conclude that they may as well upload their recordings in the original, uncompressed formats.

I get the impression that FreeSound's bandwidth is either being throttled per user or that with the growth of the platform, it's becoming more of an issue in terms of costs.

I'd be happy to pay a reasonable fee to not have that problem and support FreeSound's continued existence, and I doubt if I'm the only one.

*************

FreeSound's financial report, pinned above, indicates that they're facing a $100K euro deficit. That's surely causing some problems.

Donations probably won't bridge that gap very well; very few users can afford to donate much, and those of us who can have all of the same problems seen at similar resources on the Web, like Wikipedia, where they have to run donation drives on a regular basis simply to get our attention.

Thus a suggestion to use a funding model that I know works: Nexus Mods is, like FreeSound, a huge operation, in terms of bandwidth and file-serving. They probably require 100X FreeSound's bandwidth and storage is probably 1000X. Their two-tier system means that non-paying users aren't frozen out of the service, but paid users get much better bandwidth and can download more files at once.

But they're profitable and healthy and can afford full-time maintenance staff, not merely a few grad students or professors on their spare time. I don't care if FreeSound is profitable, but I do care if it goes broke or is suddenly unmaintained.

The owner of Nexus is fairly easy to contact and is a nice guy, and I've talked to him a couple of times over the years as it went from being a fairly-modest website to the behemoth it is today.

FreeSound's board could get in touch and discuss this with them to see if the tiered model might be a good idea for them. I understand that, as an operation tied to a university, they may face some problems justifying changes without strong data to support their proposals.

*************

Having all sounds converted to OGG upon upload, and being offered as the first choice for users would significantly impact bandwidth requirements.

The change in storage requirements would be fairly modest. Somebody at FreeSound would have to write a routine to do OGG encoding upon upload, which isn't that big of a computational expense vs. providing the file access on a daily basis.

For pretty much all users, the experience would be, essentially, the same.

Unless you're a professional audio engineer needing a raw sound recording for a pretty specific purpose that's being used on a lossless medium (i.e., you're using a field recording done in WAV for something on a CD or vinyl, etc.), having the sounds in uncompressed formats doesn't make much sense.

99% of the use cases for FreeSound assets will be compressed before end-users ever hear it.

If a user's working on a presentation or needs a sound effect to throw into a YouTube video, or are remixing the sounds for a video game mod, etc., a tiny difference in quality won't be apparent. Unity, GameMaker, Unreal Engine, etc.- all of them compress sound files by default.

YouTube and every other streaming service, including professional ones, are all compressing audio. So, while there are arguments for having the original, uncompressed sound available... I don't think it should be the typical user's first choice, to be honest. OGG at a high bitrate will be fine.