ssh Posted January 11, 2013 Share Posted January 11, 2013 This is proof of concept for simulating water inside of air field. The only difference in simulations is the presence of that air field. Done with standard Houdini flip tools. Inspired by ILM's Siggrpah talk on "Battleship". See the huge difference and much more realistic behavior. Watch it here: https://vimeo.com/57182252 3 Quote Link to comment Share on other sites More sharing options...
anim Posted January 11, 2013 Share Posted January 11, 2013 that looks amazing love the behaviour of particles in the air Quote Link to comment Share on other sites More sharing options...
lukeiamyourfather Posted January 11, 2013 Share Posted January 11, 2013 That's very nice! Did the water and air interaction add much time to the overall simulation? Quote Link to comment Share on other sites More sharing options...
ssh Posted January 11, 2013 Author Share Posted January 11, 2013 Thanks, water with air is 2-3 times slower then just water. Worth it. 1 Quote Link to comment Share on other sites More sharing options...
Erik_JE Posted January 11, 2013 Share Posted January 11, 2013 Very nice! Is there any pdf available on the talk? I tried googling but it seems my google skills are failing me. Quote Link to comment Share on other sites More sharing options...
chaindriver Posted January 12, 2013 Share Posted January 12, 2013 This is excellent Sergey! Great comparison. May I know how is this air field created? Is it some standard node/OTL that can be plugged into my DOP network? Quote Link to comment Share on other sites More sharing options...
ssh Posted January 12, 2013 Author Share Posted January 12, 2013 Very nice! Is there any pdf available on the talk? I tried googling but it seems my google skills are failing me. Thank you Erik, don't think pdfs exist. I know it from the guys who attended it on Siggraph. This is excellent Sergey! Great comparison. May I know how is this air field created? Is it some standard node/OTL that can be plugged into my DOP network? Thanks, this is very usual Flip setup with density being overrided. Some particles(air) have very low density and some particles (water) have high density. All is simulated in one step. E.g. no need to simulate air first then apply cache to water or anything like that. This is very simple setup. 1 Quote Link to comment Share on other sites More sharing options...
eetu Posted January 12, 2013 Share Posted January 12, 2013 First let me say: Awesome! Is there any pdf available on the talk? I tried googling but it seems my google skills are failing me. Probably it was more thoroughly described in the Siggraph talk, but they also talk about it in this episode of FXpodcast 1 Quote Link to comment Share on other sites More sharing options...
ikarus Posted January 12, 2013 Share Posted January 12, 2013 Thank you Erik, don't think pdfs exist. I know it from the guys who attended it on Siggraph. Thanks, this is very usual Flip setup with density being overrided. Some particles(air) have very low density and some particles (water) have high density. All is simulated in one step. E.g. no need to simulate air first then apply cache to water or anything like that. This is very simple setup. are you simulating the air field via a complete smoke solver or a stripped down version? Quote Link to comment Share on other sites More sharing options...
ssh Posted January 12, 2013 Author Share Posted January 12, 2013 are you simulating the air field via a complete smoke solver or a stripped down version? There is no smoke solver involved. Water and air are simulated with one flip solver. Quote Link to comment Share on other sites More sharing options...
ikarus Posted January 12, 2013 Share Posted January 12, 2013 There is no smoke solver involved. Water and air are simulated with one flip solver. ahh okay I just got what you meant heh. I'm suprised that density variaton would cause such a huge difference in time. Quote Link to comment Share on other sites More sharing options...
ssh Posted January 12, 2013 Author Share Posted January 12, 2013 Time difference is probably caused by the fact that you have to simulate much more particles. Because all the rest of container is filled with air. There might be way to optimize this and have air only where it's needed. I am checking it now. This is just a proof of concept to see how much can we get from this approach. Quote Link to comment Share on other sites More sharing options...
ssh Posted January 12, 2013 Author Share Posted January 12, 2013 As a nice side effect one do not need to do any additional tricks to get better behavior. Which saves fx artist's time a lot. Container, emitter, initial velocity - done. Quote Link to comment Share on other sites More sharing options...
lukeiamyourfather Posted January 14, 2013 Share Posted January 14, 2013 Thanks, water with air is 2-3 times slower then just water. Worth it. I completely agree. Quote Link to comment Share on other sites More sharing options...
Pazuzu Posted January 14, 2013 Share Posted January 14, 2013 (edited) Check this nice paper: www.cs.ubc.ca/~rbridson/docs/boyd-tog2011-multiflip.pdf Edited January 14, 2013 by Pazuzu Quote Link to comment Share on other sites More sharing options...
kris Posted January 22, 2013 Share Posted January 22, 2013 Looks fabulous!! Could you please explain more about the setup? Quote Link to comment Share on other sites More sharing options...
dvfedele Posted January 25, 2013 Share Posted January 25, 2013 Are you constantly filling up the areas that aren't water with air particles? Quote Link to comment Share on other sites More sharing options...
Akabane Posted February 20, 2013 Share Posted February 20, 2013 (edited) Still haven't done proper tests, but maybe if "all we need" is the air around the sim, we can use the particle selection for the main body of water, transform to a volume, expand the volume by the amount of air we want around our sim, and then subtract the water volume with the new one, we'd get the "outer shell" of the sim at every frame, then we can just scatter into that, bring that into our FLIP sim with density 0.01 or whatever and we'd have saver all the remaining space in the box. But maybe I didn't understand the system fully since I haven't tried it yet XD Edited February 20, 2013 by Akabane Quote Link to comment Share on other sites More sharing options...
eetu Posted February 20, 2013 Share Posted February 20, 2013 I don't think that will cut it, as we would like to have the larger scale motion of the air affecting our fluid. Quote Link to comment Share on other sites More sharing options...
ssh Posted February 21, 2013 Author Share Posted February 21, 2013 (edited) I was working on it for some time. Check it out: https://vimeo.com/60134752 Air is added only where it's needed basing on sdf of main fluid. No bounding box needed. One can control thickness of air shell, amount of air particles, limit air to specific hero regions etc. Also done some extra steps to proper detect droplets with this approach. Edited February 21, 2013 by ssh 5 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.