I’ve released a new version of the Acoustid server today. This release changes the way fingerprints are matched, so it required the database to be rebuilt, which means all previous “track IDs” will be lost. I’m expecting it will take at least a week before the database is back at the full size. This is probably the last rebuild and the IDs will stay stable from now on, but in case I discover some serious problems it might be necessary to do it one more time.

I’ve just released a new version of the Acoustid fingerprinter. I’ve sneakily released 0.3 some time ago, but it was available only on Linux. The only change was support for reading of the PUID tags. This release brings that functionality also to Windows and OS X, plus a few bug fixes. Changes since the last version are: New icon: HTTP proxy support using the $http_proxy environment variable Support for reading PUIDs from tags Fixed submission of the file format You can download the new version on the Acoustid website .

I’ve released a new version of the Acoustid server that doesn’t require submissions to include MusicBrainz recording IDs. Applications can send textual metadata (track, artist, album, album artist, year, track number, disc number) instead and the server will try to match the tags to MusicBrainz by itself. The matching is actually not yet implemented, but the server is able to collect the data. The format of the metadata is described in the web service documentation .

A new version of Chromaprint has been released. The main change is addition of an example application that uses the public C API and FFmpeg to allow externals programs to use Chromaprint without linking to the library. Other than this, there were some build system changes and added support for the latest development version of FFmpeg. Download: Source code tarball (529K) Ubuntu packages Static binaries for the fpcalc tool: Windows, x86 (633K) Mac OS X, x86, 10.

Some time ago I realized that having the server code in Java keeps me away from working more on it. The original reason for writing it in Java was that I expected it would do harder work, but I really don’t need a “fast” language for simply running SQL queries. Additionally, since developing simple websites in Java turned out to be a pain, I used PHP to do the Acoustid website.

A new version of Chromaprint has been released. There aren’t many changes, mostly just compatibility improvements for Windows and Mac OS X, plus simple ctypes-based Python bindings. Download: Source code tarball (526K) Ubuntu packages Changes since version 0.2: Fixed compilation with MSVC 2010. Added support for calculating FFT using the Accelerate framework on Mac OS X and iOS. Added Python bindings.

Last week, with the help from Zvooq , the Acoustid service was moved one more time , this time to a much more powerful server. This move made it possible to import all submissions from the backlog, so the database now contains fingerprints for over 5M unique tracks. Unfortunately, only around 2.4M of them are matched to the MusicBrainz database, but I’m hoping that the ratio will improve soon. There are a couple of things on my roadmap:

I’m not good at releasing code, as this should have been done a long time ago, but I’ve finally released the version 0.2 of Chromaprint. The main changes are new functions in the public API for working with raw fingerprints and safe releasing of memory allocated internally in Chromaprint. Download: Source code tarball (520K) Ubuntu packages Changes since version 0.1: New public functions chromaprint_{encode,decode}_fingerprint for encoding/decoding raw fingerprints.

I’ve been meaning to write this post for a long time, but never really finished it. I hope it will help people understand how does the Chromaprint algorithm work, where do the individual ideas come from and what do the fingerprints really represent. It’s not meant to be a detailed description, just the basics to get the general idea. Being primarily based on the Computer Vision for Music Identification paper, images play an important role in the algorithm.

I’ve finally written a script to take a consistent dump of the Acoustid database in the PostgreSQL tab-separated format used by the COPY command. I do not have any tools for importing it into PostgreSQL, so it has to be done manually by running SQL commands, but if anybody is interested in playing with the database, you can download it here (2.7G after compression using bzip2, 6.4G uncompressed). The data is licensed under a Creative Commons BY-SA 3.