Binaural pre-processing for contralateral sound field attenuation and improved speech-in-noise recognition