Tornado simulations in the visual effects industry have always been an interesting problem. Developing tools to provide more control over such effects is an important and challenging task. Current methods to achieve these effects use either particle systems or fluid simulation. Particle systems give a lot of control over the simulation but do not take into account the fluid characteristics of tornadoes. The other method which involves fluid simulation models the fluid behavior accurately but does not give control over the simulation. In this thesis, a novel method to model tornado behavior is presented. A tool based on this method was also created. The method proposed in this thesis uses a hybrid approach that combines the flexibility of particle systems while producing interesting swirling motions inherent in the fluids. The main focus of the research is on providing easy-to-use controls for art directors to help them achieve the desired look of the simulation effectively. A variety of controls is provided which include the overall shape, path, rotation, debris, surface, swirling motion, and interaction with the environment. The implementation was done in Houdini, which is a 3D animation software whose node based system allows an algorithmic approach to the problem and integrates well with the current tools. The tool allows the user to create animations that reflect the visual characteristics of real tornadoes. The usefulness of the tool was evaluated among participants who had some experience in 3D animation software. The results from the simulation and evaluation feedback reveal that the tool successfully allowed the users to create tornadoes of their choice efficiently.
Identifer | oai:union.ndltd.org:tamu.edu/oai:repository.tamu.edu:1969.1/ETD-TAMU-2011-05-9417 |
Date | 2011 May 1900 |
Creators | Dwivedi, Ravindra |
Contributors | Srinivasan, Vinod |
Source Sets | Texas A and M University |
Language | en_US |
Detected Language | English |
Type | thesis, text |
Format | application/pdf |
Page generated in 0.1702 seconds