Skip to Main content Skip to Navigation
New interface
Master thesis

Flow-noise en temps réel

Aymeric Augustin 1 
1 EVASION - Virtual environments for animation and image synthesis of natural objects
GRAVIR - IMAG - Laboratoire d'informatique GRAphique, VIsion et Robotique de Grenoble, Inria Grenoble - Rhône-Alpes, CNRS - Centre National de la Recherche Scientifique : FR71
Abstract : Procedural textures, such as Perlin noise, enable us to achieve high quality, resolution independant static effects. Our research aims at providing real-time animation for such noise. Thus, we can enhance the apperance of fluids -- especially geophysical fluids such as clouds, fire, water, lava, mud, etc. -- by adding animated noise at small scales. To achieve a realistic appearance, we must give the feeling of the flowing and swirling that are caracteristic of real flows. Flowing is obtained with advected textures, linked to a fluid solver. The main part of our work is simulating the swirling of the flow. We use flow-noise to add rotation to Perlin noise. Using a local measure of vorticity and Kolmogorov theory, we determine the amount of animation in the fluid at different scales. This ensures that our flow-noise has a physically correct spectrum. Computing flow-noise is quite intensive ; however, we show how real-time framerates can be achieved with current GPUs. We highlight a number of GPU-specific issues and optimisations for our algorithm, and we include complete source code and documentation for our pixel shader.
Document type :
Master thesis
Complete list of metadata

Cited literature [9 references]  Display  Hide  Download
Contributor : Team Evasion Connect in order to contact the contributor
Submitted on : Monday, June 6, 2011 - 2:08:51 PM
Last modification on : Friday, February 4, 2022 - 3:17:58 AM
Long-term archiving on: : Friday, November 9, 2012 - 2:35:54 PM


  • HAL Id : inria-00598384, version 1



Aymeric Augustin. Flow-noise en temps réel. Synthèse d'image et réalité virtuelle [cs.GR]. 2006. ⟨inria-00598384⟩



Record views


Files downloads