TTS-driven Synthetic Behavior Generation Model for Embodied Conversational Agents - Coverbal Synchrony in Human-Machine Interaction - page 325

CHAPTER 13
TTS-driven Synthetic
Behavior Generation Model
for Embodied Conversational
Agents
Izidor Mlakar, Zdravko Ka`´ i ´ and Matej Rojc
1. Introduction
Communication is a concept of information exchange. It incorporates
speech perception, visual perception, and understanding. These
concepts are correlated by sets of complex and interspersed conscious
and subconscious processes. The complementary nature of multimodal
information suggests that verbal information is accompanied by
spontaneous and planned body language, especially by gesturing
using lips, hands, and arms. In fact, the perception of a speech sound
is affected by a non-matching lip/mouth movement (Sargin et al.,
2006). Most researchers agree that speech-synchronized coverbal
behavior is essential in communication. Subsequently, artificial
interfaces should also integrate natural speech-gesture production to,
at least, some degree. The communication between different entities,
whether it be human or synthetic, commonly reflects personality and
emotions (Wallbott, 1998), idiosyncratic features (Gallaher, 1992),
and communicative functions (Allwood et al., 2007a). These non-
verbal features are regarded as coverbal behavior. Coverbal behavior
benefits the understanding process of the person being addressed