A Model of Sound Localization in Multiple Loudspeaker Reproduction