• Fingerprint (command line) client

    12 oct. 2007, 10h48m

    Hi folks!

    We have made another step towards the metadata cleaning project: the first beta of a usable fingerprint client is out!

    For the moment it's just a command line program, but we are working hard to include the library in the official client as soon as possible!

    Here is how it works: get your OS version (see below), and just type:

    lastfmfpclient yourMp3File.mp3
    If the mp3 is found in our databases, you should get something like:

    141878 Spank Rock Chilly Will 0.8913763 25 28
    141878 Spank Rock Track 09 0.04141196 1 28
    141878 Spank Rock/Rosalie Parker Chilly Will 0.034457896 1 28
    141878 Spank Rock 08 - Chilly Will 0.032753944 1 28

    The first column represent the fingerprint ID, the second the artist, the third the track name, and the fourth our score. Other values are used internally for debugging. (Note that we are still working to fix the album problem. This will likely involve MusicBrainz knowledge).

    If the entry is not found it will be automatically added, and after a few days you should find it listed.

    Here are the binaries:

    Linux - 32
    OSX - Intel

    UPDATE: if you are having problem running it on windows (especially vista), try to put those libs in the same directory of the program.

    UPDATE 2: I uploaded an new version (1.34) with several bugfixes. Please update if you can.

    UPDATE 3: The metadata server is currently down and will be returning only in january after we fixed some issues and update the dataset with the data we collected via the client (I can already tell you the response has been great! Wait for some really cool stats soon!). In the mean time you can still fingerprint and send the data if you wish.

    UPDATE 4: The metadata server is back with tons of new data! :)

    UPDATE 5: The links now point to the updated command line programs.

    UPDATE 6: Fixed issue with UTF-8.

    You can compile your own by checking it out from subversion at svn://

    Enjoy! :-)

    P.s. Yes, other file formats will follow soon!
  • New Tags!

    5 avr. 2007, 18h47m

    Maybe some of you have already noticed that we changed slightly the composition of the tag clouds. The reason is simple: we updated the algorithm which is "arranging" the tags you provide, so that it now reflects more what real fans think about their beloved artists/tracks/albums.

    This does not mean that some tags are being trashed: every tag is precious! What we have done is develop an algorithm that combines a wide set of elements to avoid - as much as possible - spammers and organized mobs (as the listeners of a famous female "artist" know very well.. ;) ).

    Enjoy! ;)
  • New artists similarity and stuff

    9 déc. 2006, 20h00m

    Hi everybody!

    Maybe some of you have already noticed: there is a new artists similarity algorithm running. We decided to be conservative and keep the old one with the (new) addition of a special filter that takes out the stuff that is clearly not related to the style of the artist. For instance, Beethoven had, among his top 100 similar artists, groups like Pink Floyd, The Beatles and the omnipresent Radiohead. Hopefully this is now fixed, yeah!

    We are very well aware that there are still tons of issues about these lists (which, by the way, are computed using your listening behavior), and we are working hard to address them (next step is probably audio analysis.. ;) ).

    About the neighbors, we have a new set of candidates that are being evaluated right now. Another problem is that the full dataset of possible pairs is very large (in the order of TeraBytes!), and it takes quite a while to compute (about two days) on our cluster. Therefore we want to be confident about what to run before starting the process and using the machines that could be running something else in the mean time.

    This is all for the moment. See ya on the next update!

  • New taste-o-meter new neighborhood

    8 nov. 2006, 19h08m

    Hi everybody!

    The new taste-o-meter is online. And the new neighborhood too!
    Now the two things are directly related, so it should not happen that you get "medium" or "low" on your neighbor anymore.

    How it works? Well, the whole neighbor computation is derived from a rather standard user-based collaborative filtering technique, with some additional "tricks". Bwahahahahah! ;)

    This technique returns a (real) value between 0 and 1, that represent the "compatibility" between two users. Since it is rather uncommon that two users share exactly the same set of artists, that number is generally between 0.4 and 0, with a lot of variations for each user.

    If we order these values for each user, we get the neighbors. To make an example, here is the list of my neighbors with the computed value:

    The value is used as the basis for the taste-o-meter.
    Since the top value varies a lot depending on how obscure your tastes are, we cannot use an absolute value to calibrate the taste-o-meter. Rather I coded a simple function that takes into consideration your first neighbor as base, and scale the other values accordingly.

    The drawback of this technique is that the resulting match is not symmetric, that is we could have the case in which:
    A->B : super
    B->A : medium

    This follows the principle of: "he is my best friend, but for him, I am only a good friend".

    From a logical point of view this sounds correct, but some good hearted people here at believes that we should make all friendships strong, and go for a symmetric version that return max(A->B, B->A) for both people and make everybody (?) happy. Tell me what you think... :)

    One last - but very important - note:
    Both neighborhood and the taste-o-meter are based on a 6 months dataset. If you (or your neighbor) did not scrobble nor listen to radio during that period.. well, no cookie! ;)

    We will switch to 12 months if we see this is not working.

    For now it's all, and sorry for my Swiss English! ;)