Measuring Dramatic Believability

The concept of a synthetic character being “believable” in a digital performance has remained a wholly subjective and nebulous descriptor. Little work has been done to provide a useful metric for comparing or contrasting the relative “believability” of different characters. This paper proposes a general method for measuring dramatic believability based on the premise that the believability of a performance to an observer is dependent on the different expectations of that observer. This is meant as an exploratory process of identifying a new, general means of defining believability for use in academic research and character design.