3DS Max 2018 liquids foam not colliding

[Anonymous]

Hello!

As the headline states I'm having some issues with the foam in maxliquids (bifrost) not seeming to colide with my geometry.

I have allready taken some steps ro fix this since this was also the case to the liquid itself. So I am now at min 500 transport steps, max 5000 and min 12 time steps and max 20 with a master voxel size of 0,175 (or something).

The liquid is now coliding fine but some of the foam just does not seem to care and I'd really like to not turn up the timesteps anymore since the sim has allready gotten grulingly slow compared other engines out there.

Any advice on how to keep the foam in line or is my only hope to increase the time steps again?

Thanks!

Your post subject line has been edited by @hagen.deloss: Maxliquids foam not colliding

[hagen.deloss]

Hello @Anonymous,

thanks for posting on the forums, I look forward to looking into some fluid simulation details with you! To best help out, I will need some additional info. Would you mind taking a few screenshots of your simulation settings? Be sure to include your voxel size, display settings and your render settings.

Additionally I would love to see what MassFX Rigid Body settings you have on the object(s) that aren't interacting the way you intend.

I'm just now learning about how Bifrost works and about Fluid simulations in general! Happy to be learning alongside such a fun and supportive community 😄

I look forward to hearing from you,  

[Anonymous]

Heyo!

Here are the screengrabs. The machine I screenshot from is not where I usually work so there were no caches present hence why the viewport is empty and the sim is set to cache the mesh without having a foam and liquid cache made. Still, same settings giving me issues.

[hagen.deloss]

@Anonymous

Thanks for the additional info! I'm still learning about fluid simulations myself, so take whatever I say with a grain of salt, but I will try my best to troubleshoot this with you!

You say that 'some' of the foam is interacting with your collision object? What does that mean? Does the foam go right through your collision object? Maybe try cranking up the voxel size and amount of the foam itself? Have you tested different foam rendering settings (changing the opacity to a lower percent)?  Here is some general info on fluid simulations.

I also would love to see your collision object settings, even though it's probably a matter of tweaking the foam settings, couldn't hurt to see those. I will chat with my colleague about it further and see what we can come up with for you... a good fluid simulation requires a lot of finesse it seems!

Please select the Accept as Solution button if my post solves your issue or answers your question.

[Anonymous]

Hey again!

1. By "some" of the foam I mean that an amount of the foam interacts with the collision object like you'd expect but a lot of it just passes right through the collision objects, hence my predicament hinted at in the post headline.

   - I have tried to sim with the voxel size even lower than the settings in the image but with improvement, there are still parts of the foam passing through the collision object. I have also been cranking up the foam (turned down the foam creation thresholds) so that's not an issue, there is plenty of foam in the scene.

2. No, I haven't really tried different render settings for the foam as this is not really a rendering problem. I've gotten the render too look how I'd like now I just need the dynamics to stick, i.e not having the foam passing through my collision geometry. 

   - I see the foam passing through the collision geo just with the viewport preview.

As for the collision setting here they are:

[hagen.deloss]

@Anonymous

I ran my own simulation and followed along to a tutorial, poked around on some similar forum threads...here are a few things we can try to get your foam working right.

How thick is the collision object that the foam is passing through? If it's thin, make it thicker, apparently fluid simulations like thicker objects to interact with.

Whats the scale on your geometry in the scene in general? It could be that the fluid simulation is at a smaller/larger scale then the rest of your scene.

Please select the Accept as Solution button if my post solves your issue or answers your question.

[Anonymous]

Hey again! @hagen.deloss

The sim/collision geo has an additional outward extrude/shell on it to make it thicker which I can remove later but just to help the sim collide properly. I do believe its about 4 voxels/cells thick, taking the voxel size down it becomes even more ofc. 

As for the scene, its set up to be a can of liquid and the scene is set to CM as unit size and everything sized accordingly based on that. The can is 14 cm tall (maybe a cm or two too tall but nothing that should produce any scale errors like you mentioned).

Thanks for the time

[Anonymous]

@hagen.deloss

Just to further visualize the sim/issues I have here is a preview of the scene. You might need to go frame by frame to see it because it moves kind of quickly, but just keep watching the bottom of the can. Do note that I have added more geo around the can to act as "helper" collision geo which is why you see some strange behavior here and there. Also keep in mind that I am only displaying 1% of the particles or maybe less, displaying 100% really kills the viewport. 

[hagen.deloss]

@Anonymous

I messaged one of the developers on this, I will relay any information I receive. 

In the meantime, I would suggest creating another collision object to test and see if the foam isn't interacting correctly across all colliders, maybe its the particular shape/modifiers of the first collision object? 

If you can take it a step further and recreate a rudimentary version of a similar scene and see if you can get the foam to interact the way you want it; you can replicate the settings and steps to the scene in question. 

Please select the Accept as Solution button if my post solves your issue or answers your question.

[klankster]

@Martin_Brattensborg I've just put a video up on Vimeo that should help in solving the problem: https://vimeo.com/272804598

Many simulations can have problems that come from (A) fast-moving colliders or (B) fast-moving fluids/foam. If a thin-walled collider is moving too quickly, it can move past the fluid during a frame.  Or if a fluid is splashing and moving quickly, it can shoot through a thin collider wall, especially if the collider is also moving.  It sounds like you took the steps of thickening the walls of your soda can and adjusting the Transport and Time Steps parameters, which will usually help.  I'm a belt-and-suspenders kind of guy, so I usually go for the "Nuclear Option" of adding a hidden, super-thick barrier that wraps the container that might leak.  For things like your soda can scene or pouring liquid into a thin-walled goblet, it's the best way to deal with leakage of this kind.

Take a look at the video and let us know if it solves your issue.

-Tom

[Anonymous]

@hagen.deloss Thanks for getting me in contact with one of the devs!

@klankster

Thanks for your reply, and thanks alot for putting in the effort to make a video! Funnily enough I allready have this setup applied to my scene (me thinking I was so clever and all to figure it out on my own and trying it) but maybe the extra objects are not thick enough?

I have the original can still present as a collision object but I have a tube around the can that slightly interpenetrates the can's walls and a cylinder doing the same on the bottom of the can in addition as colliders for the liquid.

See attached screenshot for more details.

This in addition to me having 12-16 timesteps and 1000-10000 position steps

[klankster · Accepted solution]

@Anonymous It could be the thickness isn't enough, depending on other factors in your simulation, including scale. Depending on the liquid attributes, the foam can really get some high velocity which can shoot it past the collider boundaries.  Your additional colliders are also overlapping pretty nicely which should help. Try making those additional colliders substantially thicker and see if that helps.

If not, I'd like to see your settings for the Emitter Objects, Collision Objects and Liquid Parameters.  Because you're working with a smaller scale (~1 unit = 1 CM), there are certain settings such as stickiness that can cause some odd artifacts.

Interestingly, with my rig I can't seem to duplicate your issue (or it's easily fixed), but that could just be one or two parameter differences...

[klankster]

Another thing to remember is that the Bifrost simulation engine assumes 30 frames-per-second animation speed, and tries to be physically correct.  If I'm counting correctly, from the point where your sim starts to the point where the can comes to rest is less than a second total.  That's a lot of movement for less than a second of time -- Maybe try stretching your time a bit so that action takes place over something like 3-4 seconds, and see if that helps.

[Anonymous]

@klankster

Making the additional geometry significantly thicker has indeed helped. It just was not thick enough in my previous setup and has allowed me to drop from 12 timesteps down to 4 and get a faster and better result.

My animation/render is going to be at 24 frmaes per-second, but I guess there is no way to adjust bifrost to use that instaed of the assumed 30 its using by default, would it even make a great difference?

Anyways, that has solved the issue I was having, thanks again @klankster for the response, and thanks @hagen.deloss for getting me in touch with him! 

[klankster]

Awesome!  So glad that helped.

I'm glad you posted this question -- I actually hadn't ever tried that kind of setup before (liquid inside a container with an escape hole) and it was fun to try out and problem-solve.

[klankster]

Oh, and by the way, the default is 30 FPS only when using NTSC in the Time Configuration dialog -- You can set any frame rate you like -- Film (24 FPS), PAL (25), or Custom (anything you want) and the simulation will use that.

[hagen.deloss]

Hey @Anonymous

What a journey! YaaaaY I'm super happy that @klankster was able to help us figure out a solution, and I definitely learned a few things about fluid simulation 😄

Oh and @Anonymous Feel free to post the final result on this thread if you want/are allowed to. Looking forward to chatting with you more around the forums!