VAM Database

The "Vera am Mittag" (VAM) database is a German audio-visual speech database recorded from a talk-show on TV. Its main corpus contains almost 1,000 labelled audio samples of spontaneous, unscripted emotional expressions.

The corpus was collected by the Communications Engineering Lab of Karlsruhe Institute of Technology, Karlsruhe, Germany, for research purposes, in 2005 (Dr. Michael Grimm and Prof. Dr.-Ing. Kristian Kroschel). The data collection was supported by grants of the Collaborative Research Project (SFB) 588 "Humanoid Robots - Learning and co-operating multimodal robots" by the Deutsche Forschungsgemeinschaft (DFG). The original TV show was produced by time 2 talk entertainment GmbH , Potsdam, Germany.

The labeling was carried out together with the Speech Interpretation and Analysis Laboratory (SAIL) of University of Southern California (USC), Los Angeles CA, USA, by Prof. Shrikanth S. Narayanan and his team.