Support automatic character set detection for tag data

[timoreimann]

Inspired by issue #54, is seems reasonable to support automatic character set detection to work around malformed or illegal tag text fields. Although the preferred solution should normally be to get down to the root of the trouble and fix the affected text encoding errors in the tag fields directly, it is a matter of usability/user experience to decode tag data to the best of knowledge available. Character set detection is the way to achieve this.

Web research and a quick test seem to indicate that juniversalchardet (based on Mozilla's Universal Charset Detection approach) is the Java library of choice. Its integration into Supersonic would be most easy if the employed Jaudiotagger library allowed decoding of textual tag fields by a specific character set. JAUDIOTAGGER-179 is supposed to provide this functionality but its status is yet open. Hence, additional coding effort may need to be invested into Supersonic.