mp3hash overview

mp3hash is a suite of programs to generate databases of filenames and
associated hash values purely from the music data of mp3 files.

The mp3 files are identified by their extension: end_of(lower(fn))='.mp3'

The parser then extracts just the music data, ignoring padding or
meta-data.

The clever thing about this program is its performance: it is a
second generation implementation (the original implementation can
actually be found inside the archive and was written in C).  While
the C version uses memory mapping for copy avoidance, this java
implemenation still beats it out by half a percent or so: the java
version uses threading to perform scanning, reading, hashing, and
catalogueing in parallel, thereby optimising resource utilisation.