Proceedings of the Annual Meeting of the Cognitive Science Society

Attractiveness research is typically conducted using static photographs of human faces, often cropped or otherwise edited. We present a database of videos of faces, while talking with synchronized audio that will facilitate more ecologically valid, multimodal research into face and voice attractiveness. Our database contains videos, synchronized voice recordings, and photographs of 20 male and 20 female German speakers under different emotional conditions (neutral, sad, happy, angry, flirty). Recordings were simultaneously made from three different angles (frontal, profile, 3/4 view) in front of a green screen. We report the results of a comprehensive validation of this database to make it usable for multisensory empirical aesthetics research, but also for face or emotion research. An online study was conducted to collect ratings on several dimensions including general attractiveness, allowing us to uncover the effects of predictors for facial attractiveness in a natural setting.