MusicBrainz - Automatic Music Identifier using its acoustic fingerprint algorithm
#1
Lightbulb 
MusicBrainz - Automatic Music Identifier (using acoustic fingerprint algorithms).

With MusicBrainz there would no longer a need for correctly tagged music files with MusicBrainz to properly identify your music files to finally bring your music collection in order and being able to use database a library mode functions!

MusicBrainz has very interesting acoustic fingerprinting code and API for music files, free and open source, (operated by The MetaBrainz Foundation, a non-profit corporation), very popular, and already has a cross-platform library (libmusicbrainz), would be great to have it built-into Boxee (and XBMC) so that it was used by the scanning process when building the database for the music library.

http://en.wikipedia.org/wiki/MusicBrainz
http://en.wikipedia.org/wiki/MusicBrainz#Fingerprinting
http://en.wikipedia.org/wiki/Acoustic_fingerprint

http://musicbrainz.org
Quote:What is MusicBrainz?

MusicBrainz is a community music metadatabase that attempts to create a comprehensive music information site. You can use the MusicBrainz data either by browsing this web site, or you can access the data from a client program — for example, a Audio-CD player program can use MusicBrainz to identify CDs and provide information about the CD, about the artist or about related information. You can also use the MusicBrainz Tagger to automatically identify and clean up the metadata tags in your digital music collections. Find out more in the introduction.

Quote:Acoustic Fingerprinting:
MusicBrainz initially used Relatable's patented TRM (a recursive acronym for TRM Recognizes Music) for acoustic fingerprint matching. This feature attracted a lot of users and allowed the database to grow at a fast rate. By 2005 it became obvious Relatable's fingerprinting solution didn't scale well to the millions of tracks in the database and the search for a viable replacement began. On May 12, 2006, Robert Kaye posted an announcement on the project's official blog about a partnership between MusicBrainz and MusicIP. Part of the agreement allows MusicBrainz to use MusicIP's MusicDNS service for acoustic fingerprinting (PUIDs). After a grace period of 6 months, TRMs will be phased out and MusicBrainz will rely solely on PUIDs. MusicBrainz uses RDF/XML for describing music metadata, which is available for automated processing via HTTP GET and POST methods according to REST architectural style for distributed hypermedia systems.

Magic MP3 Tagger is an example of a software which used MusicBrainz:
http://www.magic-tagger.com
Quote:What is Magic MP3 Tagger?

Your MP3 files are named incorrectly or in different ways? You don't want to listen to "Track01" files any more? Your music collection is messed? The correct release year of some files is missing? You've come to the right place.

Magic MP3 Tagger was specially designed to automatically identify and sort your music collection, which is a different approach than other tagging applications use. Magic MP3 Tagger does all the work for you by using a high sophisticated music identification algorithm!

In short, it's the tool you need to finally bring your music collection in order!


Identify your MP3 files automatically

How long would it take you to manually go through all of your music and type in the correct information? With Magic MP3 Tagger, your entire music collection can be fixed and organized automatically. This is possible by using a powerful dynamic identification algorithm which combines available song data with an internal music database under consideration of different naming patterns and similarities, various artist albums, compilation discs, known common misspellings and much more. This core identification process needs no internet connection and is therefore extremely fast.
Reply
#2
I believe Bobbin007 was going to look into this following devcon. Haven't heard from him since then though - perhaps he'll see this thread.

Certainly it's something we are interested in - we have support for reading the musicbrainz id's once they're generated, and storing them locally in the db, just not the generating side of it.

A patch would most certainly be welcomed.

Cheers,
Jonathan
Always read the XBMC online-manual, FAQ and search the forum before posting.
Do not e-mail XBMC-Team members directly asking for support. Read/follow the forum rules.
For troubleshooting and bug reporting please make sure you read this first.


Image
Reply
#3
I've used it a lot. It's the best. There is no other program that can id music files more accurate without user interference.
Image

Please add to my reputation if you find my posts usefull (+/- button below posts)
Ubuntu 12.10 minimal XBMC auto-install script :: XBMControl :: Xbmc XBOX Skins :: XBMControl for Android :: Owner of Sudo Systems
Reply
#4
I tried their Picard tagging program alright, didn't find it too helpful, got decent tagger from elsewhere though, do most of it manually following the guidelines.
Reply
#5
Which one of the apps did you use?

did u find it worked well...
Reply
#6
I'm using one called "Ultra Tag Editor" and just following the tagging guideline of people on last.fm and the musicbrainz/wikipedia directories for album info.

Any tagger app that supports ID-3 v.2 tags should do the job for you.
Reply
#7
Sweet Ill give it a try.
Reply
#8
The EchoNest releases Echoprints: The open source fingerprint era has begun
http://blog.musicbrainz.org/?p=945

Amplifind, the company that operates the MusicDNS/PUID service, recently sold its intellectual property and the PUID service will be going away eventually. It is exactly this reason why we’ve been uncomfortable relying on closed source fingerprinting software to make MusicBrainz tick.

Fortunately, I’m pleased to announce that the open source fingerprint era has begun!

Lukáš Lalinský has been working on acoustid for months now and today the EchoNest, in conjunction with 7Digital and MusicBrainz, has issued a press release that announces Echoprint, their fully open source fingerprinting solution.

The source code has already been released on github: echoprint-server and echoprint-codegen.

We’re pleased to announce preliminary support for Echoprint in MusicBrainz on our echoprint test server. The Echoprint system works similarly to how PUIDs work in MusicBrainz right now. You can use echoprints anywhere you can use PUIDs on the Echoprint test server. The version 2 of our XML web service (on the echoprint test server) now supports submitting and fetching Echoprints. To submit an Echoprint, refer to our web service documentation and example page and use echoprint wherever you’d use puid. For instance, to submit an Echoprint to the test server, POST an XML document like the one below to the /ws/2/recording resource:


[XML]<metadata xmlns="http://musicbrainz.org/ns/mmd-2.0#">
<recording-list>
<recording id="e97f805a-ab48-4c52-855e-07049142113d">
<echoprint-list>
<echoprint id="TRN5NGX1187AB4F786"/>
<echoprint-list>
<recording>
<recording-list>
<metadata>[/XML]

Although it remains to be seen when the Echoprint system will be mature enough for inclusion on the live MusicBrainz servers, going forward MusicBrainz will only support fully open source fingerprint solutions, starting with Echoprint and acoustid. We are saying no to more closed source solutions, which have never worked out well for us. MusicDNS/PUID is now officially end of life and should not be used anymore in new development.

We look forward to working with the EchoNest closely to finish up the development of the Echoprint system and to fully integrate it into MusicBrainz when it matures.
Reply
#9
I would love to see acoustic fingerprinting integrated into XBMC. I think that many aspects of the software (radio station streaming, music video identification and even library scraping) could benefit from this feature.

I did some research and found several programs and services that are capable of acoustic fingerprinting and song identification

Last.fm fingerprints
Uses an open source library to fingerprint and a web service to identify the song. Used by banshee-extension-lastfmfingerprint, for example.

MusicDNS/PUID
Open source fingerprinting with libofa, closed database queries. MusicDNS was sold to Sony-owned Gracenote and is expected to be shit-canned. Thus, MusicBrainz is ditching it for Echoprint & Acoustid (see above article from 6/23/11).

ENMFP (Echo Nest Musical Fingerprint) powered by The Echo Nest
Closed source fingerprinter, free for use subject to the terms of the Echo Nest Terms Of Service.

Echoprint powered by The Echo Nest
MIT-licensed fingerprinter named Echoprint Codegen, public API (works with Echoprints and proprietary ENMFP's), and a permissive license for their database.

Acoustid fingerprint
LGPL fingerprinter named Chromaprint, open web service.


From my research, here's where fingerprinting stands

Last.fm is a safe bet with 90 million songs indexed and a rock-solid API. ENMFP and Echoprints both use the same Echo Nest API; the ENMFP database spans 30 million songs, whereas the Echoprints database only covers 13 million songs. The Acoustid database is an unknown (small) size, not to mention inaccessible for the last week due to a database rebuild. Currently, the MusicBrainz database supports MusicDNS PUID fingerprints; Echoprint support is available on one of their test servers, and Acoustid fingerprint support will (probably) be added in the unspecified future.

There are 5 types of fingerprints, 5 databases (counting the two Echo Nest db's separate) and 7 ways to connect fingerprints to databases (given that MusicBrainz accepts PUID, Echoprint and acoustid).
BTC: 1JtXwJdGdE9YnYgThWBT2StFCU5sEYkbVD (personal), https://kodi.tv/contribute/donate-bitcoin (foundation). Donations in the form of controllers, especially ones that don't work in Kodi, are also appreciated.
Reply
#10
Obviously addons are the way to go, but how should this work? I think this would work, to have a new addon category, "Music identification services", that contains XML files, similar to scrapers, for each database above. The fingerprinting code would have to be included in core, unless binary addons are a possibility. Then, the XML file would request one or more types of fingerprints, and be in charge of prioritizing the results of the database lookup if more than 1 type of fingerprint returned results.

For example, Last.fm "music identification service" XML would invoke Last.fm fingerprinter and query Last.fm server. MusicBrainz XML would invoke three fingerprinters (MusicDNS PUID, Echoprint, Acoustid), query the MB database, and if the PUID and Echoprint returned matches it would prefer the Echoprint match, for example.

There's probably a better way to do this, any thoughts?
BTC: 1JtXwJdGdE9YnYgThWBT2StFCU5sEYkbVD (personal), https://kodi.tv/contribute/donate-bitcoin (foundation). Donations in the form of controllers, especially ones that don't work in Kodi, are also appreciated.
Reply
#11
Just noticed that VLC now features audio fingerprinting using AcoustId webservice parser with MusicBrainz

https://github.com/videolan/vlc/tree/mas...ebservices
Reply
#12
210237 (thread)

AcousticBrainz is a new API from the MetaBrainz people (responsible for MusicBrainz)

http://acousticbrainz.org/api

Quote:What data does it have?
AcousticBrainz contains information called audio features. This acoustic information describes the acoustic characteristics of music and includes low-level spectral information such as tempo, and additional high level descriptors for genres, moods, keys, scales and much more.
Such additional meta data would be great for Smart Playlists.


More info on http://acousticbrainz.org
Quote:Welcome to AcousticBrainz!

The AcousticBrainz project aims to crowd source acoustic information for all music in the world and to make it available to the public. This acoustic information describes the acoustic characteristics of music and includes low-level spectral information and information for genres, moods, keys, scales and much more. The goal of AcousticBrainz is to provide music technology researchers and open source hackers with a massive database of information about music. We hope that this database will spur the development of new music technology research and allow music hackers to create new and interesting recommendation engines.

AcousticBrainz is a joint effort between Music Technology Group at Universitat Pompeu Fabra in Barcelona and the MusicBrainz project. AcousticBrainz was originally envisioned by Xavier Serra, the founder and head of the MTG. At the heart of this project lies the Essentia toolkit from the MTG -- this open source toolkit enables the automatic analysis of music. The output from Essentia is collected by the AcousticBrainz project and made available to the public.

AcousticBrainz organizes the data on a track basis, indexed by the MusicBrainz ID for recordings. If you know the MBID for a recording, you can easily fetch from AcousticBrainz. For details on how to do this, visit our API documentation.

All of the data contained in AcousticBrainz is licensed under the CC0 license (public domain).
Examples

If you're wondering what this collected data actually looks like, have a look at the last 5 tracks that have been submitted.
Reply
#13

(2014-12-03, 13:23)RockerC Wrote: 210237 (thread)

AcousticBrainz is a new API from the MetaBrainz people (responsible for MusicBrainz)

http://acousticbrainz.org/api

Quote:What data does it have?
AcousticBrainz contains information called audio features. This acoustic information describes the acoustic characteristics of music and includes low-level spectral information such as tempo, and additional high level descriptors for genres, moods, keys, scales and much more.
Such additional meta data would be great for Smart Playlists.


More info on http://acousticbrainz.org
Quote:Welcome to AcousticBrainz!

The AcousticBrainz project aims to crowd source acoustic information for all music in the world and to make it available to the public. This acoustic information describes the acoustic characteristics of music and includes low-level spectral information and information for genres, moods, keys, scales and much more. The goal of AcousticBrainz is to provide music technology researchers and open source hackers with a massive database of information about music. We hope that this database will spur the development of new music technology research and allow music hackers to create new and interesting recommendation engines.

AcousticBrainz is a joint effort between Music Technology Group at Universitat Pompeu Fabra in Barcelona and the MusicBrainz project. AcousticBrainz was originally envisioned by Xavier Serra, the founder and head of the MTG. At the heart of this project lies the Essentia toolkit from the MTG -- this open source toolkit enables the automatic analysis of music. The output from Essentia is collected by the AcousticBrainz project and made available to the public.

AcousticBrainz organizes the data on a track basis, indexed by the MusicBrainz ID for recordings. If you know the MBID for a recording, you can easily fetch from AcousticBrainz. For details on how to do this, visit our API documentation.

All of the data contained in AcousticBrainz is licensed under the CC0 license (public domain).
Examples

If you're wondering what this collected data actually looks like, have a look at the last 5 tracks that have been submitted.


I would love to see this as a scrapper!
Reply
 
Thread Rating:
  • 0 Vote(s) - 0 Average



Logout Mark Read Team Forum Stats Members Help
MusicBrainz - Automatic Music Identifier using its acoustic fingerprint algorithm00