DiRT Dust in RealTime

1
DiRT Dust in Real-Time
CS 426 Senior Projects
Team 3
Advisors

• Marcos Bagby
• Ryan Romero
• Brett Sulprizio
• Hiroko Uda

Sergiu Dascalu Brian Westphal Fredrick Harris,
Jr. Joseph Jaquish
2
Overview

• Project Description
• Practical Applications
• Objectives Development
• Design
• Demo
• Future Work
• Questions/Comments

3
DiRT (Dust in Real-Time) is

• Dust in Real-time (DiRT) is a 3D dust
  visualization program and interactive computer
  benchmarking tool designed to model real world
  dust dynamics in a virtual environment in
  real-time. The interactive benchmarking tool
  allows users to track and gauge the performance
  of their systems ability to render the dust by
  issuing real-time system and environment related
  reports. The system reports update in accordance
  with the users play. Similar to a video game,
  users will be able to play during the
  simulation by way of a simple vehicle simulator.
  This paper presents details of DiRTs UML
  requirements specification, software architecture
  specifics, high and low-level design details,
  user interface principles and snapshots.

4
Practical Applications

• Military
• The modeling of dust for use in simulations may
  help modern combat experts plan for or pre-empt
  the affects of dust caused by vehicles or the
  atmosphere.
• Video Games
• Low resource gaming engines.
• CAVE Project
• Immersive virtual environment.

5
Main Objectives

• Simulate/Model Dust
• Realistic
• Real-time

6
Development

• Three Methods
• Volumetric Fog
• Diffuse Reflection (future work)
• Particle Systems (future work)

7
Method Volumetric Fog

• Large affected area divided into small volumes.
• Different Densities
• Volumes will translate, morph, etc
• Translation, morphing specified by airflow
  parameters.
• Stencil Buffer and Alpha Blending with multipass
  rendering creates the illusion.

8
Design System Architecture
9
Design
10
Demo
11
Future Work

• Diffuse Reflection and Particle System methods
• Reflection probably most difficult to implement.
  Less resource intensive than a particle engine.
• Particle probably easiest to implement, but very
  resource intensive.
• Implement for CAVE use.
• Voxel Processing Method

12
Any Questions/Comments?