Clouds - PowerPoint PPT Presentation

1 / 13
About This Presentation
Title:

Clouds

Description:

Altocumulus - puffy. 6500-20000ft. Water droplets. Stratus, stratocumulus ... Trace rays from eye, accumulating density. For each point along ray, trace ... – PowerPoint PPT presentation

Number of Views:42
Avg rating:3.0/5.0
Slides: 14
Provided by: mcsCsuh
Category:
Tags: clouds | puffy

less

Transcript and Presenter's Notes

Title: Clouds


1
Clouds
  • Based on Sect 5.3.3 in Parent

2
Cloud Types
  • Cirrus - wispy
  • gt20,000 ft
  • Ice crystals
  • Altocumulus - puffy
  • 6500-20000ft
  • Water droplets
  • Stratus, stratocumulus
  • lt6500 ft, water droplets, layered

3
Rendering issues
  • Amorphous, volumetric structure
  • Swirling, bubbling motion
  • Low-albedo (reflectance) techniques
  • Assume scattering effects are negligible
  • For cirrus
  • High-albedo techniques
  • Scattering is significant
  • For thick clouds

4
More rendering issues
  • Wavelength-dependent scattering
  • Self-shadowing, shadows on landscape
  • Volumetric shadowing expensive
  • Trace rays from eye, accumulating density
  • For each point along ray, trace another ray
    towards light source to determine illumination

5
Early approaches
  • Semitransparent surfaces
  • Fractal synthesis of plane textures
  • Ellipsoids with fourier-synthesized transparency
  • Randomly placed, overlapping spheres with solid
    texture, transparent near edges

6
Volumetric Clouds
  • Surface-defined clouds don't allow flythroughs
  • Volumetric, density-based models do
  • Particle systems can model motion well, but need
    large numbers of particles
  • Volume-rendered implicit functions
  • Can control implicit function with a particle
    system

7
Volumetric Cloud Examples
  • From Nishita, et al, Siggraph 96

8
Volumetric Cloud Example
  • Eberts method

9
Ebert's Volumetric Cloud Model
  • Two-level hierarchy
  • High-level control for animator
  • Implicit functions
  • Procedural low-level details
  • Turbulent volume densities
  • Benefits of approach
  • Abstraction of detail
  • Data amplification

10
Ebert Clouds (2)
  • Implicit density function as summed implicit
    primitives

11
Ebert Clouds (3)
  • Implicit primitives are non-solid
  • Before evaluating blending functions, the point's
    position is perturbed using noise and turbulence
  • Final density is a blend of perturbed implicit
    density and a turbulence function
  • Density(p) uD(perturb(p)) (1-u)turbulence (
    p )
  • Setting u1 gives cotton balls
  • Finally, density is modified by an exponential
  • - density pow ( density, power )

12
Noise and Turbulence
  • Noise(x,y,z) created by
  • Assign random values to grid vertices
  • Interpolate between vertices near (x,y,z) with
    splines
  • Turbulence(x,y,z) is bandlimited noise with a
    fractal spectrum (1/f)
  • For (fminfreq fltmaxfreq f 2)
  • Val fabs ( noise (xf, yf, zf) / f )

13
Animating Ebert Clouds
  • Animate the particle system of implicit
    primitives
  • Only 100-1000 particles needed
  • Can use turbulence, vortex, etc to control
    particle motion
Write a Comment
User Comments (0)
About PowerShow.com