Sex differences in vocalization have been commonly found in vocal animals. It remains unclear, however, how animals perceive and discriminate these differences. The amplitudes and latencies of event-related potentials (ERP) components can reflect the auditory processing efficiency and time course. We investigated the neural mechanisms of auditory processing in the Emei music frog (Nidirana daunchina) using an Oddball paradigm with ERP. We recorded and analyzed eletroencephalogram (EEG) signals from the forebrain and midbrain when the subjects listened to white noise (WN) and conspecific sex-specific vocalizations. We found that (1) both amplitudes and latencies of some ERP components evoked by conspecific calls were significantly higher than those by WN, suggesting the music frogs can discriminate conspecific vocalizations from background noise; (2) both amplitudes and latencies of most ERP components evoked by female calls were significantly higher or longer than those by male calls, implying that the ERP components can reflect sex differences in vocalization; and (3) there were significant differences in ERP amplitudes between male and female subjects, suggesting a sexual dimorphism in auditory perception. Together, the present results indicate that the music frog could discriminate conspecific calls from noise, male's calls from female's ones, and sexual dimorphism of auditory perception existed in this species.