A speaker system and method, in which the intelligibility of reproduced speech or music sound, derived from an audio program signal, is enhanced by means of at least one of a first and second transfer function of a signal process applied to the audio program signal. The first transfer function incrementally varies the volume of the reproduced sound directly as a function of the volume of ambient noise. The second transfer function incrementally or continuously varies the high frequency response of the reproduced sound inversely as a function of the volume of the ambient noise. The ambient noise can be measured by a microphone, for example, located on or near the speaker system and can also be electronically averaged over a predetermined time period. The signal process provides an audio output signal that is applied to at least one amplifier and at least one speaker.