Voice background data structure