I'd argue it depends on the methodology by which they're implementing the technology. They could be using the directionality to amplify the "target" signal and identify the signals that the wearer desires to mute and then run a fast algorithm to reduce that sound.