Slashdot Mirror
Open Content Music Database Launched
An anonymous reader writes "The open source music database MusicBrainz was launched officially today. The data is partly in the public domain, partly under an open content like license. It includes artist/album/track information, with more to come. There's support for CDDB-like CD identification (actually, a lot of the current data was imported from freedb), but also identification of single tracks via audio fingerprints (TRMs). Help both in adding new content by tagging your music collection and consolidating the existing data is welcome. Also check out some technical information on the XML database at IBM developerWorks."
Now if only they would allow you to upload play lists, and classify your Mp3's =)
---
A friend of mine wanted to open up a punk CD store on the web. Being the nice friend I am I helped him import a large amount of data from the CDDB into his OScommerce store (Os commerces is an open source store package, pretty cool)
After much alignment of tables, farting around with the data eventually we got it right but with one small detail left out....
We didn't have cover art images...
So frantically he tried copying the images from other sites, then he kept insisting there was a way I could easily parse the obsfucicated data from other stores (album art gif's are never the same name as the album)
So eventually he gave up on it, but it got me to thinking, would the cover art be something unlawful for a CDDB type of entity to host?
I find the idea of a program that can identify my MP3s by audio fingerprint, and will submit that information to somewhere on the Internet a little creepy...
I had brought up to my friends several times, how it would be great to start something similiar. The metaratings are a great idea, providing the database openly to the public is great, and i'm falling in love with their tagging utility.
And it's all non-profit! (and will likely get better each and every day now that it has all this slashdot traffic)
I am this close to posting the 28 meg mysql database on my school account, but I think the coe admins would kill me!!
hooray! it's a sex wiki
than open source music is a simple way of getting samples put onto vinyl. Ive seen so called "computer dj" programs and its still nothing close to cutting on real vinyl.
Though I do like the idea of freeing up music, beats, samples etc, with out a good medium to manipulate the audio this is less of an achievement than it seems.
The reason I say this is because Im a skrxtch dj so this problem of manipulating the audio affects me much more than someone who's mostly doing music on the computer [various forms of electronica]
Another interesting point is that that most skrxtch records encourage reusing the samples and beats. In fact I have a copy of Tales from the Crate next to me that on the cover reads "Unauthorized Duplication is prohibited. Unauthorized flipping, mixing and juggling is recommeneded" [Thats an approximation, I couldnt find the sleeve]
If you get an error, type "OVERRIDE" or "SECURITY OVERRIDE" and then try the optimize command again.
(Insert the usual IANAL disclaimer here).
This sounds like a reasonable assumption at first, but if you look at it more closely, on which legal basis could Amazon.com prevent you from using these images? They are not the copyright holders of the art. All they did was a simple scan/resize/save_for_web...
Amazon could overlay a "www.amazon.com" on the scans to make the reuse more annoying, but then they could face suits from the artists...
Now the artists could prevent you from using their art. But if you are not defaming them, I don't see why they would. I think they will rather appreciate the publicity.
I code, therefore I am.
I looked at this project about 8 months ago and planned to use this setup for an open source media utility.
:
I stopped short at that time because
1) the TRM (song fingerprint) technology was owned by a seperate entity and was closed/private.
*Paranoid pondering* what if the TRM tech owners decided to charge for future use after the database was largely used and accepted. Although the database would remain open, they could charge for new fingerprints (song IDs). Not neccessarily a bad thing but we've all seen things how f'd up these situations can get.
2) the TRM generation took place on the server. Doing a batch of fingerprints would tie up a connection for quite a while.
My brainz a bit fuzzy on this but I think a portion of the actual audio data is uploaded and then processed on the server. I figured that generating TRMs completely client side and then uploading/matching song data to the server would do better for scalability.
Just the same I haven't looked a the project recently and it may have since changed.
Regardless I think its a pretty cool idea.
I find the programs that interface with musicbrainz to be very useful. The organizational view used by Zinf is probably superior to any other I have used, including iTunes and MusicMatch jukebox. It is great that we have this large database of data that can be accessed from client programs using an open api.
A system is needed to be able to allow users to provide feedback (and hence publicity to new music) and most importantly somehow give the artists some money for the work they produce, afterall they need to eat too.
---- The Open Source Record Label : : LOCARECORDS.COM
...is have it store a list of emotions associated with the songs, eg romantic, depressing, happy and so on. You could then build a playlist for whatever mood you were in, or mood you wanted to be in. This is FAR more effective than simply categorising music by genre (in my experience anyway).
I think there is a windows program called Moodlogic that did this but it was closed source, expensive, and it used a proprietery database. It also used the accoustic properties of the song, so it could identify badly-named mp3s.
A system that doesn't allow anonymous (unsigned) files to be shared.
And why should you be obliged to sign your own creation, or public domain one ?
How do you sign ? Is it free ?
When you can sign pubdom, what prevent you to sign unfree ?
Someone mentioned that this technology helped the music industry in its lawsuit against Napster (or another service?), showing which files were being swapped. I haven't been able to quickly find that post again, so here's my thoughts in the void.
Would technology that allows fingerprinting down to the file level, in conjunction with a user-supported (i.e. richly populated) database, actually help music file swapping? Conceivably, someone could integrate this into their service to indicate that a file was what it was called before it was transferred.
Granted, there are other ways to fake a file than just giving the "right" name to a bogus file of the right size. But I imagine something like this (along with checks) could make it much more difficult to kill PNP by populating services with bogus files.
A half-baked idea, but my two cents' worth anyway.
I started my own intranet movie database. The problem is that populating it takes so much time, even with the helper apps I made. I really wish someone with the server resources would allow developers to work on a movie database similar to CDDB. The benefit of it would be that people would be able to import data (title, description, rating, length, actors, etc) via XML/etc to use to populate an offline database of movies they own. This personal database could be used to do custom searches on movies the person owns. Imagine having guests stop over for movie night, hand them a wireless PDA or wireless keyboard to media PC (widescreen) for movie selection; they search by a theme, "Camping" and get two results, one of which you end up pulling off your shelf to watch with them.