Audio source selection circuit (QW) for an audio signal processor with inputs for source signals (F1, F2, NA, NB) provided by at least one source (Q1, Q2), with a processing device (V) which forms pairs of signals from the source signals, with a settable source selection logic (Qs) to which the channels of the processing device (V) are applied, and with outputs coupled to signal outputs of the settable source selection logic (Qs) and each forming an output channel (Co1, Co2, Co3), the processing device (V) including an automatic audio mode setting device which generates suitable intermediate channels (Cz1, Cz2, Cz3, Cz4) from the source signals (F1, F2, NA, NB) according to a source- and application-related mode.