Design of an Italian Text–to–Speech System