FreeDB Is Shutting Down
freedb.org
freedb.org
I ran a similar service for years (daz.com) and eventually shut it down because musicbrainz was not worth competing with though I think we did a better job of tying the links between the various bands by tracking artists from one act to another, also studio artists which made it quite nice for music discovery.
I'm sorry you found it unsustainable, but from what I can find about Daz it seems to have been pretty much the same idea as Gracenote: take the data submitted by users out of their desire to share alike, lock it up, and then commercialize it by putting a bunch of semi-related functionality around it. Assuming this link is the correct Daz, anyway, it talks about mp3 downloads, used CD shopping, artist exclusives, and other things I was always very glad FreeDB didn't want to be:
https://www.j-oe.nl/2011/06/development-of-daz-com-share-you...
"Visitors and members can search for all kinds of music related information and listen to ten thousands various samples. DAZ.com offers members a free mp3 to download every day (see homepage) and two RSS feeds: the free mp3 of the day and the updates from last 24 hours. Feel free to point your reader to this feeds.
In the near future we will add more community features, create a better user-interface and offer a second hand cd-spot. At the moment I am contacting with signed and unsigned artists to discuss ways of working together."
Daz was essentially a wiki which allowed people to link artists, bands and media. Unfortunately we never got to the point where we got permission to stream media, but plenty of artists thought that was what we were doing (we just did the legal maximum sample that you could distribute under fair use). I can't count the number of angry 'cease-and-desist' letters that we received, invariably followed by sheepish apologies that they never actually downloaded anything from the site. But some artists really liked it and it was quite interesting to be 'pen pals' with some of the very well known names in the industry.
The whole reason I built it because I saw the writing on the wall with Gracenote, I contributed plenty to CDDB and suddenly saw my own work be used for profit.
Maybe instead of looking at a blog have a look at the archive.org snapshots of the website over time, they give a much better picture. I also built a media player / file sharing client that used daz.com for the meta data called mxchg.com.
But music on the web is very hard to make work as a long term thing, the bills have to be paid and even if the content is user generated you still need to supervise it and ensure that people don't pull tricks to vandalize it or destroy it.
I'm very much impressed with how Musicbrainz has been able to keep it clean, free and have such stamina, that's serious dedication.
Absolutely, and it's thankless work too, like how the #1 user-agent string on FreeDB's statistics page is "FreedbDemo" straight from their example code project. Moderation is more necessary now, though, with the Musicbrainz/Discogs paradigm where there's one canonical "correct" listing for every artist/album/whatever. One of my favorite parts of FreeDB/CDDB (and one of the reasons I use it to this day) is how many albums have multiple user-submitted entries that fit different tagging styles. I totally get how some people would consider that a downside, and it still doesn't preclude vandals crapflooding you with bogus garbage entries, but I don't consider a lot of the MB/Discogs tagging guidelines to be especially good. Any tag source I use—including the paid GD3—is guaranteed to need at least a little editing, and I find that FreeDB gets me closer to my ideal tagging format more often than any of the others.
I was toying with reading/writing my EAC database file a few months ago, and this news reinvigorated my interest in getting something nice and local running to serve all my new and (many) old clients.
my email jacques@modularcompany.com
If anyone is inconvenienced by the FreeDB shutdown and needs some short term service to get them by, while they work out what else to do, send us an email and we can see if we can open our FreeDB server.
http://www.onemusicapi.com/blog/2014/12/03/retrieving-freedb... (this talks about using our API, but we could open our FreeDB access point too).
Also, ISTM that it should be quite feasible for musicbrainz to track artists and "acts", in a pinch by cross-referencing wikidata.org which acts as a sort of shared focus for the whole Linked-Open-Data metaverse.
They do also store FreeDB disc IDs, for example: https://musicbrainz.org/cdtoc/lwHl8fGzJyLXQR33ug60E8jhf4k-
Artist credits can be tracked -- there are extremely fine-grained relationship models -- but most entries don't have the associated metadata because it's a hassle to input.
For my own system I used a music fingerprint written by Giancarlo Pascutto (I hope I spelled that right, it's ages ago), which used a really nice algorithm that did not see a single collision in a very large number of tracks.
Every time this happens I get more and more angry about the trash-heap-dumpster-fire services from exit-focused "enterpreneurs" that seek to replace them
How much could freedb have been costing MAGIX? Could they not have handed it off to someone else?
More than zero. MAGIX acquired freedb in 2006, so I assume this decision was an easy one for some middle manager who has been there for a couple years and has no love for some crusty Old Internet service that's beyond obsolete (aka can't figure out how to put ads on it).
This corporate "has to be profitable" mentality needs to die. It's sad that we are in a place where even simple services require these supposedly ongoing maintenance costs, when they were originally designed as set-it-and-forget-it services
How can a "set-it-and-forget-it" web service exist from the service provider's POV? You always need to do small things like monitoring your servers and making sure they're all up-to-date and still working. This is ongoing maintenance cost. Additionally, you need electricity and some place to put your servers (be it virtual or physical).
Also, running something unprofitable out of your own pocket works as a hobby maybe, but not for a company at scale. You need to pay your employees. Even as a hobby it's problematic because you can only do it so many times until your regular income is exhausted.
Download: http://www.freedb.org/en/download__server_software.4.html
Found it a bit sad to see this notice followed by a recent news headline saying the project was guaranteed to continue:
Continuation of the free database guaranteed
With the acquisition of the www.freedb.org domain
MAGIX also takes on all duties regarding the
worldwide freedb communityI would encourage everybody to download the software:
http://www.freedb.org/en/download__server_software.4.html
And the db:
http://ftp.freedb.org/pub/freedb/
It would be good to have multiple copies floating around.
magnet:?xt=urn:btih:96208ba12ca7ffdac5b5adccf1a41e484de94f4a&dn=freedb&tr=udp%3A%2F%2Ftracker.openbittorrent.com%3A80&tr=udp%3A%2F%2Ftracker.opentrackr.org%3A1337%2Fannounce&tr=udp%3A%2F%2Ftracker.coppersurfer.tk%3A6969%2Fannounce&tr=udp%3A%2F%2Ftracker.tiny-vps.com%3A6969%2Fannounce&tr=udp%3A%2F%2Ftracker.torrent.eu.org%3A451%2Fannounce&tr=http%3A%2F%2Ftracker.torrentyorg.pl%2Fannounce&tr=http%3A%2F%2Fexplodie.org%3A6969%2Fannounce
It seems like so little for artists/tracknames of every piece of music ever published (to CD), because it would take 0.1% of my disk space, while knowing it's so much data.
.
freedb-complete-20191203.tar.bz2: bzip2 compressed data
curl -I http://ftp.freedb.org/pub/freedb/freedb-complete-20191203.tar.bz2
HTTP/1.1 200 OK
Date: Sun, 29 Dec 2019 02:42:04 GMT
Server: Apache/2.0.54 (Debian GNU/Linux)
Last-Modified: Tue, 03 Dec 2019 19:14:50 GMT
ETag: "542dff-3b460112-8295de80"
Accept-Ranges: bytes
Content-Length: 994443538
Content-Type: application/x-tar
SHA256 (freedb-complete-20191203.tar.bz2) = e005ec2ee3f3dabf2f5c45b49273358fcb796aae946cdfeeb31cd2e555b86978
MD5 (freedb-complete-20191203.tar.bz2) = 6f8be385155242e11bae5d5acb557b5d
.
cddbd-1.5.2.tar.gz: gzip compressed data, was "cddbd-1.5.2.tar", last modified: Sat Jul 1 17:47:47 2006, max compression, from Unix, original size 675840
curl -I http://ftp.freedb.org/pub/freedb/cddbd-1.5.2.tar.gz
HTTP/1.1 200 OK
Date: Sun, 29 Dec 2019 02:45:17 GMT
Server: Apache/2.0.54 (Debian GNU/Linux)
Last-Modified: Tue, 19 Sep 2006 10:06:55 GMT
ETag: "20bb2f-2bef1-a6738dc0"
Accept-Ranges: bytes
Content-Length: 179953
Content-Type: application/x-tar
Content-Encoding: x-gzip
SHA256 (cddbd-1.5.2.tar.gz) = bec016250b06fa7aed368ca163489b72cebc93c1ae804e5f342cea3c64f5e4e4
MD5 (cddbd-1.5.2.tar.gz) = 2f0e6a0ec6a4157a5c3874fca10b33ce
.
freedb-proto5-update: Perl script text executable
curl -I http://ftp.freedb.org/pub/freedb/freedb-proto5-update
HTTP/1.1 200 OK
Date: Sun, 29 Dec 2019 02:47:34 GMT
Server: Apache/2.0.54 (Debian GNU/Linux)
Last-Modified: Wed, 20 Sep 2006 15:05:29 GMT
ETag: "20bbac-1480-f00ce040"
Accept-Ranges: bytes
Content-Length: 5248
Content-Type: text/plain
SHA256 (freedb-proto5-update) = 4cc1f7b7909c845c55fe1e8ea4b4844e45d2645d7ddec74300517378c2558081
MD5 (freedb-proto5-update) = 0ec2fb029e9aeae76ddf25daea8ddd59
The files that were mentioned above should now be available at https://archive.org/details/freedb
After uploading the files I noticed that there were some results on IA already for FreeDB, and one of the results is even a user named freedb which I guess might be an official account of the FreeDB people.
He seemed annoyed when I showed him how CDDB/FreeDB worked.
The second version of the cluster looked like this:
https://www.clustercompute.com/
The first (of which I have no pictures) was just 10 whitebox PCs.
I'm sure it has something to do with the music genre these albums fall in (hardstyle), but as I said Discogs did have them all and I've heard the same from people with different tastes in music.
So I like to use TagScanner[1] to tag my albums. It searches both MusicBrainz and Discogs and apparently others as well.
P.S. I checked a single artist to see if Musicbrainz has caught up with Discogs yet a few months ago and the only things listed were a 5 year old album and some singles I believe. I just checked again and in the last 3 months some users added everything else.[2]
[0]: https://discogs.com [1]: https://xdlab.ru/en [2]: https://musicbrainz.org/artist/5347c7dc-a8bd-4a05-8ae1-a17fb...
In other words, from a "discid" (in Musicbrainz format or FreeDB format or otherwise), I want artist, album, and track titles, and optimally cover artwork.
Does anyone know how to get that, in JSON or XML format, from Musicbrainz?
https://musicbrainz.org/ws/2/discid/I5l9cCSFccLKFEKS.7wqSZAo...
will give you a list of "releases" for that discid. Use the release ID in this endpoint, with the included parameters:
https://musicbrainz.org/ws/2/release/8e061dc4-790e-4587-ba53...
and you'll get the track data. Not quite sure how to get the album art yet, but should follow.
(The Release lookup has a `cover-art-archive` JSON object which contains information about what can be fetched from CAA.)
I imagine as you get up and running, you'd end up with industry contacts, access to industry data sources and so on but I'd be interested to know how these sorts of things get started?
It has always been like that in the Web 1.0 days: just a bunch of people being anal about something.
And it worked beautifully.
Web 2.0: People anal about their egos and snipe and attack each other until everyone is the same level of sad and lonely.
Ah, memories.
Nostlagia.
I'm not strongly against flagging my comment, it's effectively void of substance (zero meaning). But if you flag zero meaning comments, you should even more flag negative meaning comments. By that I mean flagging most of logical fallacies, bullshits, cargo cults, echo chambers, and therefore half of comments on A(G)I topics.
By that I mean flagging most of logical fallacies I'm currently working on a tool to automate this.
[*] Assuming they care about putting the "CDDA" logo on their product, but luckily those non-standard outliers—Sony rootkit included—mostly came and went with the the 2000s.
It seems CTDB provides access to MusicBrainz, so that might be my alternative.
I'm sure a few clone sites will pop up and you could always run a local copy yourself.
If Exact Audio Copy doesn't allow you to set the API to connect to, you could always use some proxy software (I use Charles) to forward freedb links to your chosen replacement.
I do love the formatting and structure of MB's Style Guidelines page, but it goes somewhere I dislike from the very first Principles page: https://musicbrainz.org/doc/Style/Principle/Error_correction...
"Artists sometimes choose to present names and titles in ways that deliberately contradict the rules of the language they're in (e.g. unorthodox spellings) and/or the MusicBrainz Style Guidelines. To describe the way we handle such choices, we use the term 'artist intent.' The general idea is that if an artist intended something to be written in a special way, then MusicBrainz should follow that intent. Unfortunately, it can be difficult to find out what an artist intended. If you want to claim that some deviation from the Style Guidelines should be considered artist intent, the burden of proof lies on you. A seeming error may be considered evidence of artist intent if it is consistently found on all of an artist's official releases. The best evidence would be a statement of intent by the artist (e.g. edit 6892422: https://musicbrainz.org/edit/6892422)."
Arguing with other music nerds on the Internet about which one of us has the Factually Correct Tags is not my idea of fun, and FreeDB avoids this problem by supporting multiple entries for any given disc. I totally understand how some people would consider that a burden, but I smile any time I get to choose from a few entries on FreeDB because one of them will always be substantially closer to my ideal. For me this usually manifests in ways such as (but not limited to):
- Genre tags. According to Discogs half my CDs belong in a single bucked labeled "Folk, World, & Country" https://www.discogs.com/genre/folk%2C+world%2C+%26+country
- Use of Artist vs AlbumArtist
- "Featured artists" in the title tag vs in the Artist tag (vs not listed at all?)
- "Year" tags when an album's slightly-staggered multi-region release overlaps a year boundary, or when a different label has re-released a bit-identical disc with packaging labeled a different year. I have that problem with a lot of my CDs from Sire or Tommy Boy Records where the post-1985 pressings are distributed under Reprise with a new date: https://en.wikipedia.org/wiki/Reprise_Records#Revival_and_Re...
- Title capitalization, my most frequent hassle. I have many CDs with track or artist names in ALL CAPS ENGLISH SHOUTING style. Super super common on Japanese releases, and both MB and Discogs used to have rules specifically banning anything except Camel Cased English Literary Title style.
I say "used to" because they have both gotten better on that issue, but a good percentage of the older data was entered under the old Title Case Only guidelines and remains that way. For an example, compare the same album described on Musicbrainz https://musicbrainz.org/release/041f87d6-6583-42cd-8cf3-5f4a... versus on a fandom wiki https://remywiki.com/Beatmania_HIROSHI_WATANABE_beat_indicat...
This is the relevant Musicbrainz guideline that didn't used to exist: https://musicbrainz.org/doc/Style/Language/Japanese
"Although the Japanese script has no capitalization, it is very common for Japanese titles to contain words in other scripts. Japanese artists have a tendency to choose capitalization and punctuation for aesthetic reasons; and to be very consistent regarding case over all releases. For this reason, words in the Latin script on a Japanese release should be in the same case as on the album art if other available sources, such as official discography or record label pages, are consistent; not normalized according to English or other capitalization standards."
But then _that_ guideline rubs me the wrong way, implying "capitalization and punctuation for aesthetic reasons" is exclusive to one culture, and letting record labels and marketing materials override the artistic work itself. Here's an example from Scotland: bis' 2007 CD+DVD greatest hits release.
- Title Cased on Musicbrainz: https://musicbrainz.org/release/45a14f8d-687d-4309-aa83-1004...
- Title Cased on Discogs: https://www.discogs.com/Bis-We-Are-Bis-From-Glasgow-Scotland...
- obviously aesthetically lower-cased on the actual album I'm holding in real life: https://i.imgur.com/hlaIRHj.jpg
The FreeDB system we're losing allows me to share my tags alongside the Title Cased version, so everyone can be happy: https://i.imgur.com/xrFj66A.png
Obviously any metadata service at all is perfectly fine for the 99% of people who just want some idea of what song they're listening to, and I realize I'm making my life way more difficult than it needs to be by caring about this to the extent I do, but music is too important and personal of a thing for me to settle for 99%. In a world of otherwise constant chaos my music library is my one zen garden. The single thing I curate to something asymptotically approaching my concept of perfection. A more accurate log of my own feelings and experiences than any words. Replicating the FreeDB data on my local network wont be hard, and that database will undoubtedly serve me well for years as I discover different things new to me despite their chronological age, but we're losing something way more special here than just the data and a crusty old Perl CGI script.
FWIW, something somewhat similar to this will Soon™ be part of MusicBrainz too: https://tickets.metabrainz.org/browse/MBS-4501 – you might want to read through there and maybe even put your own thoughts about what it can be used for in a comment on that ticket.
> MB […] used to have rules specifically banning anything except Camel Cased English Literary Title style
Just out of curiosity: When was this? I’ve been involved with MusicBrainz since summer of 2006 (ie., for over 10 years) and this hasn’t been the case for as long as I’ve been around.