DynEmo is a publicly available database of significant size containing dynamic and authentic EFE of annoyance, astonishment, boredom, cheerfulness, disgust, fright, curiosity, moved, pride, and shame. All EFEs' affective states are identified both by the expresser and by observers, with all methodological, contextual, etc., elements at the disposal of the scientific community. This database was elaborated by a multidisciplinary team. This multimodal corpus meets psychological, technical and ethical criteria. 358 EFE videos (1 to 15 min. long) of ordinary people (aged from 25 to 65, half women and half men) recorded in natural (but experimental) conditions are associated with 2 types of data: first, the affective state of the expresser (self-reported once the emotional inducing task completed), and second, the timeline of observers' assessments regarding the emotions displayed all along the recording. This timeline allows easy emotion segmentations for any searcher interested in human non verbal behavior analysis