An audio device for vehicles includes two audio sources; three audio sinks formed of two headphone connections and a speaker system; and a touch-sensitive display. The audio device is configured to simultaneously play back the two audio sources; display one representation each for the audio sources and the audio sinks on the display; and detect a touch input, wherein the touch input includes a motion of a finger of the user on the display. The motion begins on the representation of an origin audio source and proceeds toward a destination audio sink. In response to the detection, the origin audio source is output on the destination audio sink.