SEMEX - An efficient Music Retrieval Prototype

The applied techniques are based on the assumption that music is representable symbolically. This is the case for most music; consider, e.g., the score notation of western music or MIDI. These are not, however, ideal representations; the score notation requires a conversion into a computer-readable form, and MIDI contains a lot of irrelevant information. Instead, music is represented by sequences of symbols (or strings), where the symbols are taken from an alphabet corresponding to some particular attribute of music, such as the duration or the pitch of a note.