Perceptual evaluation of cartoon physics: accuracy, attention, appeal

  • Authors:
  • Marcos Garcia;John Dingliana;Carol O'Sullivan

  • Affiliations:
  • Graphics Vision and Visualisation Group, Trinity College Dublin;Graphics Vision and Visualisation Group, Trinity College Dublin;Graphics Vision and Visualisation Group, Trinity College Dublin

  • Venue:
  • Proceedings of the 5th symposium on Applied perception in graphics and visualization
  • Year:
  • 2008

Quantified Score

Hi-index 0.00

Visualization

Abstract

People have been using stylistic methods in classical animation for many years and such methods have also been recently applied in 3D Computer Graphics. We have developed a method to apply squash and stretch cartoon stylisations to physically based simulations in real-time. In this paper, we present a perceptual evaluation of this approach in a series of experiments. Our hypotheses were: that stylised motion would improve user Accuracy (trajectory prediction); that user Attention would be drawn more to objects with cartoon physics; and that animations with cartoon physics would have more Appeal. In a task that required users to accurately predict the trajectories of bouncing objects with a range of elasticities and varying degrees of information, we found that stylisation significantly improved user accuracy, especially for high elasticities and low information. To assess attention, many simulated objects were shown to participants on which words appeared at random, the task being to speak and remember them. Our results do not confirm that attention can be directed in such a scenario using cartoon physics. However, a game with cartoon physics was chosen to be more appealing almost twice as often as one with no stylisation applied. We conclude that stylised motion can be a valuable tool to improve physically based animations.