Parameterizing Emotions For Procedurally Content Generated Game Levels : A Case Study of the Game To the Skies!

Abstract: Procedural Content Generation (PCG) in games often suffers from a lack of emotional connection with the player as well as a perceived sameness. This thesis aims to design and test a tool for PCG game levels for the game To the Skies! which incorporates an emotional intention based upon the color-emotion theory proposed by Fugate and Franco, 2019.  Russell’s (1980) Circumplex Model of Affect was used as a foundation for the parameterization of emotions, which were implemented in a PCG test level for the game, developed in Unreal Engine. By adopting research through design, this thesis goes through an iterative process where the tool is under constant improvement through quantitative testing using the Self Assessment Manikin method proposed by Lang & Bradley (1994). The tests indicate that, although colors can to some degree elicit emotions, the player’s perception of the test levels did not match the expected emotional response based on color alone for all emotional states tested. The thesis concludes that, while color-emotion association can aid in generating PCG with emotional intent, additional complementary elicitors are likely required, and further tests are needed that incorporate shapes, placement, and weather patterns.