Polarity, phase and geometry

Following the teaching of the early Bell Telephone Laboratories´ experiments in auditory perspective, wide speaker spacing is needed to realize accuracy of geometry. This results in substantially random phase, so that polarity is relatively unimportant. This is supported by Lissajous figures of two-channel stereo signals. It is still good practice, however, to observe polarities, if for no other reason than to permit monophonic reproduction over a stereophonic array. Where the stereo signals contain a strong monophonic component, correct system polarity is better than random polarity. In some stereo situations, bass is improved by correct polarities.