@strypey @switchingsocial @musicman

Well, based on my preliminary research, to mirror the FMA we need:

1. An estimation of the size of the FMA (number of tracks, and total size of files)
2. A funkwhale instance were we can import the music
3. A script to crawl/download the whole archive
4. Permission from the FMA to run the script (since it may increase their bandwidth cost)
5. Funding to cover the hosting costs, if relevant

@strypey @switchingsocial @musicman

I think I can provide 2. and 3., so any help on the other aspects would be amazing. If you have the time to research 1., contact the FMA about 4., provide hosting resources (or funding), that would be awesome!

@strypey @switchingsocial @musicman thank you for the clarification. Since Archive.org seem to allow (and even encourage) scraping / downloading, and I found a way to get the tagged files (by downloading the originals), that's a few questions/problems solved!

@strypey @switchingsocial @musicman

Just to let you know, following our discussion from last week, I'v started writing a small utility to grab the FMA content from archive.org.

I plan to mirror at least a part of the FMA on open.audio using this :)


So, I've imported ~25Gb of music from the Free Music Archive on open.audio using this method.

It worked beautifully, and it's now available for the whole #Funkwhale federation. Follow this library from your Funkwhale instance and you'll have access to it:


You can also visit open.audio/library/ directly to listen to all this new content! (you don't need an account ;)

Since this Proof of Concept work, the next step will probably be to mirror the complete FMA. I expect the whole thing to weight many, many terrabytes though, so we'd have to figure out how and where to store that!

I've launched a download task for ~2000 more albums (roughly 5-10 times what was already downloaded). I expect it to take an additional 200Gb of disk space.

This import went well too, and I launched another mirrorring tasks for 4000 additionnal albums before going to sleep yesterday.

All of this is currently importing into open.audio, but I'm confident saying that in a ~60 minutes, almost 350Gb of additionnal CC music (6000 albums, 40K tracks) from the FMA will be available!

In total, I'd say we have roughly 40% of the FMA mirrored on open.audio, which takes less space than I initially expected.

@funkwhale ha oui fma, funkwhale.mochi.academy, l'instance de @shiro , je connais.


@gordon @shiro alors je gère déjà 3 instances, on va peut-être se calmer :D

@funkwhale @shiro nan mais je parlais à Shiro, c'est son domaine 😃

@funkwhale @shiro et du coup vu qu'il n'a pas réagi, il doit être en train d'installer l'instance :blobnom:

Sign in to participate in the conversation

cybrespace: the social hub of the information superhighway jack in to the mastodon fediverse today and surf the dataflow through our cybrepunk, slightly glitchy web portal support us on patreon or liberapay!