It only stands to reason that vocal modeling and manipulation is bound to happen once processing power is able to provide realtime response. Like all technology it will have warts and problems in its early phases but will improve over time.