SC Conference - Activity Details

Organizing, Visualizing and Navigating a Digital Music Library using PFASC and Network Workbench

D. Scott McCaulay  (Indiana University)
Eric Wernert  (Indiana University)
Posters Session
Tuesday,  05:15PM - 07:00PM
Room Rotunda Lobby
A digital music library is typically organized based on manually maintained metadata such as artist and genre. A more flexible and less labor intensive alternative is to automate the process using similarity analysis software capable of determining key signature, tempo, instrumentation and other features extracted from audio data. Network analysis and visualization tools provide a convenient and useful method to navigate and explore the library by clustering files with similar characteristics. Users can browse through nodes in proximity to any starting point and listen to snippets of files considered similar based on selected parameters. We are using PFASC, a parallel toolkit for audio similarity analysis developed at Indiana University, for extraction and measurement of audio features. Using MPI, tens of thousands of audio files can be analyzed in a matter of hours. Network analysis and visualization is performed with the Network Workbench Tool developed at Indiana University and Northeastern University.
   IEEE Computer Society  /  ACM     2 0   Y E A R S   -   U N L E A S H I N G   T H E   P O W E R   O F   H P C