MVSep MultiSpeaker (MDX23C) - this model tries to isolate the most loud voice from all other voices. It uses MDX23C architecture. Still under development.
