Most of our social interactions involve combining information from both the face and voice of other persons: speech information, but also crucial nonverbal information on the person’s identity and affective state. The cerebral bases of the multimodal integration of speech have been intensively investigated; by contrast only few studies have focused on nonverbal aspects of face-voice integration. This work highlights recent advances in investigations of the behavioral and cerebral bases of face-voice multimodal integration in the context of person perception, focusing on the integration of affective and identity information. Several research domains are brought together. Behavioral and neuroimaging work in normal adult humans included are presented alongside evidence from other domains to provide complementary perspectives: studies in human children for a developmental perspective, studies in non-human primates for an evolutionary perspective, and studies in human clinical populations for a clinical perspective.