Plausible motion: conclusion

1
Plausible motion conclusion

• Ronen BarzelJohn Hughes

2
What youve heard today

• Freedom in models
• Allows you to achieve various goals
• measuring plausibility
• constructing desired motion
• efficient simulation
• all physically correct

3
Choice of model is just a choice

• CG has adopted models from engineering
  disciplines
• But properties of interest to engineering are
  different from those of interest to animation.
• We can choose models that meet our needs
  visually observable plausibility, ease of
  control, efficiency of computation

4
If observable phenomena are important

• Can use phenomenological models rather than
  physical
• Faking Dynamics (Slinky dog), CGA 1996
• Structural Modeling of Flames, Siggraph 2002
• Continuum between
• Physical model that captures features of interest
• Phenomenological model of features of interest

5
Not just animation

• Same principles can apply to rendering
• engineering problem of light transport, vs.
• creating a plausible image

6
Open Questions

• Making user interfaces for plausible-motion
  specification and exploration
• Find the right spectrum of models and search
  algorithms for wide classes of problems

7
Other things to see

• Today 1015 ?
• Efficient Synthesis of Physically Valid Human
  Motion, Fang, Pollard.
• Today 545pm
• Evaluating the Visual Fidelity of Physically
  Based AnimationsOSullivan, Dingliana, Giang
• Perceptual Metrics for Character Animation
  Reitsma and Pollard
• Thursday, 800am
• Keyframe Control of Smoke SimulationsTreuille,
  McNamara, Popovic

8
We hope weve

• Conveyed a new world view about simulation
• Inspired future research and development
• Made you think

9
Thanks

• to our speakers
• Stephen Chenney, University of Wisconsin
• Jovan Popovic, MIT
• Ronald Fedkiw, Stanford University