Blackstone Posted February 3, 2010 Share Posted February 3, 2010 (edited) Hello guys, I want to make an tiny PRMAN DSO to interpolate an massive particle shape from an coarse one, and I thought several way to do this, Lookup neighbors from a seed particle, according the distance between sphere centric and the neighbors, then fill the space. Like this "Simulation of High-Resolution Granular Media". I prefer this, but how to control the high-resolution particle shape is very difficult. The particles were cached on each frame. We can use particle ID to track the movement of single particle, then we sample the track line as a spline, generate more particle at current frame. Maybe can't keep continuity. From seed particles, subdivide the origin one into multi-particles. Not good, I think. Inspired from here "Interpolatory Refinement for Real-Time Processing of Point-Based Geometry". Assumed the particles is an PDF which could be represent as an implicit function in an grid. Then we sampler this grid to make more and more samplers around the seed particle. Inspried from dozens of Monte Carlo papers. According an old post about how to make Voxel-Bitch, the author said use spline to do interpolation, but how did this could be implemented, I never know about that. The interpolation method should be fast, parrallel. Here is my implementation from the 1st method, not good, I think. So can you help me ? Thanks very much. Edited February 3, 2010 by Blackstone Quote Link to comment Share on other sites More sharing options...
Blackstone Posted February 3, 2010 Author Share Posted February 3, 2010 or the APR(Adaptive Particle Refinement) in physics or something like PIC(Particle in Cell) in CFD ? We build an screen-oriented grid and fill the grid with more and more particles ? Quote Link to comment Share on other sites More sharing options...
ykcosmo Posted February 3, 2010 Share Posted February 3, 2010 "build an screen-oriented grid and fill the grid with more and more particles ? " I did the dso in this way too Quote Link to comment Share on other sites More sharing options...
Blackstone Posted February 4, 2010 Author Share Posted February 4, 2010 "build an screen-oriented grid and fill the grid with more and more particles ? " I did the dso in this way too Hey, thanks for your reply Can you teach me how to interpolate the coarse particle shape ? Which method are you using ? Quote Link to comment Share on other sites More sharing options...
ykcosmo Posted February 7, 2010 Share Posted February 7, 2010 If you come to my house, I would show you the source code.... Quote Link to comment Share on other sites More sharing options...
Blackstone Posted February 10, 2010 Author Share Posted February 10, 2010 If you come to my house, I would show you the source code.... In fact, the all mathmatics knowledge is enough to me. I just did not know which way is the "best". Quote Link to comment Share on other sites More sharing options...
phong Posted March 4, 2010 Share Posted March 4, 2010 That sounds quite interesting! Can you explain to me please why a screen-oriented grid instead of a world-oriented grid is used? Because if it's screen-oriented, procedural particle creation will depend on camera transformation which might cause weird artifacts or not? Quote Link to comment Share on other sites More sharing options...
eetu Posted March 4, 2010 Share Posted March 4, 2010 This is all mightily interesting! How well do these methods maintain the coherence between frames? E.g. it sounds "filling a screen oriented grid" would not easily be coherent from frame to frame? Maybe I just haven't thought it thru well enough I've done some stuff along these lines (check here or here), but I never really checked what had been done before - thanks for great googling pointers.. eetu. Quote Link to comment Share on other sites More sharing options...
Recommended Posts
Join the conversation
You can post now and register later. If you have an account, sign in now to post with your account.
Note: Your post will require moderator approval before it will be visible.