Generating natural narrative speech for the Virtual Storyteller

If a storyteller recites a story he will try to liven up his speech by using certain rhetorical techniques, such as usage of a specific narrative speaking style and emotion display. Subject of this thesis is the generation of narrative speech. Some important aspects of narrative speech will be analysed and translated to rules which can be used to convert neutral speech into narrative speech. Based on this rules a module is implemented which automatically generates narrative speech based on a certain input text. The module can be used in the Virtual Storyteller Project, a multi-agent story generation system in which plots are automatically created, converted to natural language and presented by an embodied, virtual storyteller using spoken language with appropriate prosody and gestures. The implemented module is used to pronounce the plots that are created by the virtual storyteller in a narrative way.