Qui parle ?

Abstract

This paper describes a study comparing three types of cues to signal the speaker inside a virtual environment: a speech balloon above the speaker's avatar; the speaker's name on-screen; and a representation of the speaker inside a 2D map. People are able to quickly and correctly identify the speaker, no matter which cue is used. However, people recall the dialogue least when the cue is the speaker's name. This low recall might be explained by the interference of a semantic cue on a semantic task