Most sound localization algorithms rely on correlation methods. They are inspired by a neuronal coincidence detection model with delay lines as postulated by Jeffress, which was actually found in barn owls later. However, recent research has shown that in mammals the decoding of interaural time differences (ITDs) in the medial superior olive (MSO) is based on excitation and contralateral inhibition. As shown in the PhD thesis of Encke, this neuronal circuit decodes the phase difference between left and right ear signals along the tonotopic axis provided by the cochlear filters. In this paper we have developed and implemented a physical bio-inspired model of the detection of interaural level differences (ILDs) and ITDs. The system consists of a three-dimensional printed human head with pinnae. We fixed electret microphones in both ear canals from the inside of the head and recorded signals in a room and outdoors. We used a complex gamma tone filterbank to extract both magnitude and phase of the ear signals. We calculated the magnitude difference (ILD in dB) between left and right ear to mimic processing in the left and right lateral superior olive (LSO) and the phase differences (MSO) to extract ITDs and evaluated the system.