Staging exhibitions: methods and tools for modelling narrative structure to produce interactive performances with virtual actors

CrossTalk is a self-explaining virtual character exhibition for public spaces. This paper presents the CrossTalk system, including its authoring tool SceneMaker and the CarSales exhibit. CrossTalk extends the commonplace human-to-screen interaction to an interaction triangle. The user faces two separated screens inhabited with virtual characters and interacts through a frontal touch screen. One screen features the exhibition’s hostess, an agent who explains exchangeable exhibits located in the opposing screen. The current exhibit is CarSales, a demonstration of automatically generated dialogue, performed by virtual actors. The physical presence of the characters is established through the separation of screens and intensified by inter-character conversations across screens, tying hostess and exhibit together. CrossTalk utilises a combination of both automatically generated and pre-scripted scenes, and a context memory to adapt to the user and the environment. CrossTalk’s authoring tool SceneMaker, in a strict separation of narrative structure and content, provides non-experts with a screenplay-like language to create installations for staging exhibitions.