Differences between poker-faced and affective, 'real' and 'virtual' news readers

摘要: Much attention has been paid to the simulation and display of emotional states in applications which an Embodied Conversational Agent (ECA) takes role a real human who is expected have show emotions. However, what do if ECA replaces where protocol prescribes poker-face, likein case TV news reading? The question whether noticeably non-real scenario, user profits from benefits layer, even this contradicts ‘genre-practices’? To answer question, we are conducting experiment virtual readers evaluated, with different amounts both facial expression voice. In paper goals, design, preliminary results ongoing discussed]