When using enhance speech the alst ~20% of the audio track the level just drops out. It looks so far to be one chanel but it renders an otherwise useful tool useless. Adjusting the mix amound either degrades the track sound (lovering) or makes the audio frop even worse (raising).