Jump to content
hopbin9

Can Houdini do this fluid setup?

Recommended Posts

Here's the video:

Sorry, it's strangely dark. I don't know why vimeo messed with the gamma again.

post-4013-130096051564_thumb.jpg

Share this post


Link to post
Share on other sites

Yes it is, I am preparing a video of it now. I think I got a reasonably good result.

I actually think it is also possible to do convincing bits falling from a melting object. I'm not sure I can do that yet but it may be a matter of masking/scaling forces. There is also a useful gas elasticity node but I can't get it to work properly because it always asks for missing attributes and I'm not sure where they come from (hand made or supplied by another node, the docs are not clear on it).

It is also possible to fake a solution to the problem of "hanging" geometry by continuously emitting from the object you melt and advecting it by its own smoke sim as it were.

for the falling bits, may use the procedural shrink or eroding SDF and release the particles which can then collide with the rest SDF.

last year I did the melting project but failed at converting the particles into sph simulation at the mid simulation continuously. so I finally resort to realflow to solve the liquid look. :(

particles.mov

Share this post


Link to post
Share on other sites

Here's the video:

Sorry, it's strangely dark. I don't know why vimeo messed with the gamma again.

Macha, thanks for sharing :P I will try it too. for the vimeo I used to convert my video to flv first and then upload and hope vimeo would not double convert it. I have a HD flv converter which can protect the quality of the video very well, if you need I can send you :P

Share this post


Link to post
Share on other sites

There is some very interesting stuff going on in this thread. I don't exactly understand how the cloth solution would produce the right results?

I gave it a try today, and this is what I have so far. It seems to be in the right direction, except that if I increase the viscosity settings of the fluid container then all the particles explode!

test02.hip

Share this post


Link to post
Share on other sites

It seems like the size of the particles in the fluid are fixed and I can't seen to figure out how to tell Houdini to make the particle's radius smaller. Otherwise, if the spacing of the particles is overlapping then they seem to explode when viscosity is set to a high value.

Anyone know how to adjust this?

Share this post


Link to post
Share on other sites

Ok, so I'm a bit of a noob when it comes to fluids.

If I increase the substeps to a large about like 150 min-substep, then it starts to become more stable.

In the attached file, the substep of the fluid solver is 200 with a viscosity of 1,000. Seems to solve the first frame fine, but if you Space+G in the viewport to zoom all. There are these strange particles forming a large shape.

It doesn't render the fluid. So it seems to be failing.

test03.hip

Share this post


Link to post
Share on other sites

Ok, I found that if I increase the Rest Density the simulation becomes stable again and solves the exploding problem.

That is what I have now, and it seems better but looks more like poring paint then cream.

test04.hip

Share this post


Link to post
Share on other sites

Well, I can't say this has been easy. Viscosity in Houdini seems to me to be an unstable parameter. Anything to strong causes the simulation to be unstable. I would think this kind of fluid to be a common need in most projects.

Anyway, this is what I was able to get done tonight.

test05.hip

post-5539-130132612495_thumb.jpg

Share this post


Link to post
Share on other sites

I found this on YouTube. Looks like it was done with Houdini 9, and I'd like to know how they setup the viscosity to be so strong on the fluid.

Anyone know the history of how this video was created?

Also, here's a nice viscosity rendering done in Blender. Very nice.

Edited by hopbin9

Share this post


Link to post
Share on other sites

There is some very interesting stuff going on in this thread. I don't exactly understand how the cloth solution would produce the right results?

I gave it a try today, and this is what I have so far. It seems to be in the right direction, except that if I increase the viscosity settings of the fluid container then all the particles explode!

The reason for that is that SPH get's very in-stable due to stacking problems, 'the way' to get around this is increasing the sub-steps, how ever this results in countless hours of simulating. You do get a nice result, but it's at a price. That's why people try to modify FLIP for viscose simulations, which in it's core is a very good idea. It's just that FLIP needs a bit 'SESI' of love before it will be able to successfully accomplish that. Which leaves us holding our fingers crossed hoping for the best, as far as viscosity is concerned.

Share this post


Link to post
Share on other sites

I found this on YouTube. Looks like it was done with Houdini 9, and I'd like to know how they setup the viscosity to be so strong on the fluid.

Anyone know the history of how this video was created?

Also, here's a nice viscosity rendering done in Blender. Very nice.

It's using the 'elastic' option. You need to enable it on the particle fluid solver, under "internal forces" (where you would have enabled the viscocity force) just hit 'enable elastic force' then play with the settings.

Share this post


Link to post
Share on other sites

I found this on YouTube. Looks like it was done with Houdini 9, and I'd like to know how they setup the viscosity to be so strong on the fluid.

Anyone know the history of how this video was created?

just have a look at:

http://www.sidefx.com/index.php?option=com_content&task=view&id=1875&Itemid=262

it's video 13.

The scenefile is even included.

Share this post


Link to post
Share on other sites

Here's the video:

Sorry, it's strangely dark. I don't know why vimeo messed with the gamma again.

Couple questions for you Macha. Are you doing anything special here to maintain that nice curved bottom edge (right above the ground) as the teapot melts and points collide with the groundplane? It's an issue I've been trying to solve. As the advection occurs, the points sort of start stacking on top of eachother, and the geometry starts to lose it shape/ volume.

On the vimeo description you say that smoke is being emitted from the evolving particle sim. Im a little confused :), but are you creating an advected particle sim, using that as a smoke emitter, then using that smoke vel field to advect your final geometry? Is this a way to help the geometry maintain its volume as it deforms? Can you describe a bit more your methods for doing the experiment in he video? Thanks

Share this post


Link to post
Share on other sites
Couple questions for you Macha. Are you doing anything special here to maintain that nice curved bottom edge (right above the ground) as the teapot melts and points collide with the groundplane?

Yes, I finished the animation before it went all tits-up.

Seriously though, since it is advected by smoke velocity it takes rather a long time to completely flatten

The way this worked was to take the current geometry and emit smoke from it (via a particle system emitting from it). That way at each step you get a new smoke source position, always coinciding with the current state of the teapot and it holds together better.

If I was to do this again I'd try to incorporate some of the ideas that Jeff posted here:

http://www.sidefx.com/index.php?option=com_forum&Itemid=172&page=viewtopic&t=21927

There are also some promising looking gas microsolvers (elasticity!) but I have a hard time figuring out how they work. I created some attributes that I thought were necessary (mass, density, etc) but something is still missing. I'm mostly confused what to do with the popnet. In the flipsolver it is set to Velocity Update mode but I've so far been unable to build a minimal setup that works satisfactorily.

Edited by Macha

Share this post


Link to post
Share on other sites

There are also some promising looking gas microsolvers (elasticity!) but I have a hard time figuring out how they work. I created some attributes that I thought were necessary (mass, density, etc) but something is still missing. I'm mostly confused what to do with the popnet. In the flipsolver it is set to Velocity Update mode but I've so far been unable to build a minimal setup that works satisfactorily.

As far as I know, Elasticity belongs to the SPH method/solver, on the particle fluid object you have to set the Particle Type to SPH, this will ensure that all the attributes are created. It sounds cool but I'm afraid you'll run in to the same issues as with the SPH solver. From what I've seen from my own tests, is that you do get a result from this SPH method combined with FLIP. I ended up having the same type of behaviour like the Dough cutting demo, which is nice but not very realistic.

Share this post


Link to post
Share on other sites

Ah, no, sorry, I didn't explain well. I'm not really talking about recreating flip or sph. I am interested in finding a setup that includes elastic forces (those elastic microsolvers) and that is minimal in size. I know how the attributes are created but there's so much stuff in the particle fluid object that I want to strip it down to the bare minimum (needed for elastic forces). The docs are quiet about about many of the detailed microsolver information so I'll often end up experimenting my way through it.

Basically that crumblesolver I made a while ago was a kind of elastic force. I'm hoping to do it a different, easier way.

Ideally I want an empty object, a couple of solvers, a bit of data and thats it.

Edited by Macha

Share this post


Link to post
Share on other sites

I'm sorry, I figured that you tried to create elastic forces for the FLIP solver.

Well in that case I guess you'll end up creating your own spring force. I'm concerned that you'll run in to overshooting problems, not to mention issues with collisions and sort stuff. In terms of that, doesn't it easily get's very complicated again?

You might be interested in using a FLIP-a-like approach where you use a field to calculate the distortion of the velocity on the particles and then advect the velocity by this distortion amount which will result in a viscose type of fluid.

You can to look into the gas strain integrate and a gas strain forces (this comes from SPH), if you give a zero vector field as the 'vel' field to the gas strain forces it will return a vector field needed to 'undistort' the current velocity field. Next you can use a Gas Calculate (or a Gas Field Vop for more control) to add this field to the velocity field, as a result the fluid will try to stay together or be more viscose.

I have to note that I've tried to implement this for the FLIP solver and it worked for me until a reasonable level, but I ran in to general collision and volumeloss problems. Also the fact that the behavior of FLIP totally changes if you change the resolution makes things very difficult to be usable in a production environment.

Share this post


Link to post
Share on other sites

I'm going to put together an experiment with Fluids viscosity settings.

A blob of fluid will drop onto an angled platform, and the settings will vary depending on the take. So I might simulate the same setup 20 times, and then render the results.

I'm thinking of taking each simulation into a scene where each one is laid out across 20 platforms. Then the camera pans across each platform as the simulation plays (offset by the position of the camera). I'll place some text above the platform showing the setting.

This way there will be a video showing a progression of viscosity, and a reference tool for people looking for the right look.

I know how to setup a simulation to run on a machine, but I'm not sure how to handle the vary the parameters. If I use takes, then I have to create 20 different takes. Is there a way to vary settings in a procedural way?

Share this post


Link to post
Share on other sites

I know how to setup a simulation to run on a machine, but I'm not sure how to handle the vary the parameters. If I use takes, then I have to create 20 different takes. Is there a way to vary settings in a procedural way?

The wedge ROP is probably the tool you're looking for. Once you set it up (tell it which params to change each time, how many, etc), you only have to hit render once and you can get as many different sims written out as you want. This sideFX video is a good intro on how to use it.

http://www.sidefx.com/index.php?option=com_content&task=view&id=1396&Itemid=259

Share this post


Link to post
Share on other sites

Thanks, that is exactly what I needed.

While playing around with some settings. I created this Jello like fluid using strong Elasticity and Plasticity settings. I'm thinking these settings effect the shaping of the fluid a lot more then viscosity. I might be able to recreate the effect I'm after with a combination of the two.

Share this post


Link to post
Share on other sites

Create an account or sign in to comment

You need to be a member in order to leave a comment

Create an account

Sign up for a new account in our community. It's easy!

Register a new account

Sign in

Already have an account? Sign in here.

Sign In Now

×