apulSoft say that splitS isn’t actually a de-esser, but rather a mixer than makes it possible to balance the level of sibilances with the rest of the vocal signal.