Pancho – The Mobile Music Explorer

The ever growing music collections on mobile devices renders the habitual browsing schemes – such as browsing by artist or by often ambiguous and inflexible genre terms – obsolete, as the collection is not organized by means of user’s perception. During this thesis a visual browsing scheme for mobile devices has been introduced and developed on Android, a free mobile software platform. A high dimensional Euclidean representation of the “World of Music” was used to make song similarity computations cheap and thus accessible for mobile devices. For visualizing large music collections on small screens the three basic requirements local focus, global overview and orientation in space were defined, which make browsing in large music collections possible and intuitive. Furthermore two visual metaphors were introduced: The lens metaphor for housing the local and global view and the cake metaphor for providing information about song clusters such as genre information. Simulations in Android’s Emulator showed that the implemented methods are able to cope with large data sets containing more then 3’000 songs (∼ 20 GB).