Angelika Hönemann

Linguistics and Literary Studies
Project title: 
Expressing attitudes in adaptive audiovisual speech synthesis
Depending on the social status or a current situation, humans express different attitudes in communicative interactions, which are still difficult to be realized in multimodal speech synthesis. For their synthetic modeling, an extensive analysis of their audiovisual production and perception cues (acoustic analysis of speech, facial and head movements) is carried out. Based on these analyses, the expression of the attutudes uncertainty and surprise will be implemented in an audiovisual adaptive speech synthesis system as part of the overall Cognitive Service Robotic Apartment (CSRA) architecture. Its final evaluation will be carried out in this scenario, where the effect of attitudinal expression of a robot or artificial agent on humans in human-machine interaction is examined.