Ice cream melting - Mesh flick - Maya

[thhvfx]

Hi guy,

I'm working on a shot with ice cream melting. And I'm having the artifact or flicker on the mesh. Another version is my mesh doesn't seem to fall down on 1 side but the other side is working perfectly and that make the melting process doesn't look natural to me. I wonder if you guy know how to fix these issues. I have attached the Mov file and Maya file in the google drive link. Please have a look.

https://drive.google.com/open?id=0B6...F8tVmh5ZjlxdlE

Thanks,

[Gosho.Genchev]

Hello thhvfx, sorry for the late reply.
A couple of us have been trying to figure out a way to make your setup work. Here's a few notes from me:

1. In the second version of your scene, the Droplet Breakup parameter is animated but the Surface Tension is set to 0. The Droplet Breakup only works if Surface tension is > 0 (https://docs.chaosgroup.com/display/...uid-Parameters).
I tried setting the surface tension to 1 and the icecream starts to look like molten rubber but the flickering disappears. Animating it to go from 1 to 0 over the first 30 or so frames looks OK but there's still some random particles popping up here and there, partly because of the Droplet Breakup animation that you have going on there. If you set the droplet breakup to 0 at the start and increase it when the cone starts shaking, you'll get a similar result.

2. The thickness of the collision geometry (the icecream cone) is too low compared to the size of the grid cells. I tried increasing the resolution - it helps but there's still some penetration going on. I also went and increased the thickness by selecting the outer faces, converting my selection to vertices and moving them along the normal. That helped a bit but the look of the simulation changed too so I'm not sure you'd want to do that.

3. The source geometry is very low res which is probably the reason for the artifacts that you're seeing. I subdivided Mesh1 and those disappeared.

4. The bottom of the icecream cone is a separate geometry which seems to be causing problems. I filled the bottom of the cone instead of using the separate piece that you have there.

Hope any of that helps. If not, please post a screenshot with a red circle around the area of your simulation that you dislike so we can nail down the issue.

Cheers

[thhvfx]

Hi, Gosho

Can you please send me the scene that you already fixed those problems so I can try to pick it up from there.

Thanks, Gosho.

[Gosho.Genchev]

Hello thhvfx!

I've fixed an issue with the discharge modifier - in the uploaded scene the discharge was set up such that the liquid was appearing at the side of your source geometry. I presume you want it either at the top or bottom ?
I've closed the hole at the bottom of the icrecream cone, animated the surface tension, and increased the grid resolution to about 2M cells to help with the collision penetration problem.
You can try to enable Forward Simulation from the Simulation tab and Clear Inside on the PhoenixFD properties for the icecream cone - again, to resolve the collision problem.

With those things set up, I can't see any unusual behavior. It's still far from done but hopefully you can focus on the artistic side now that the technical problems are resolved.

If there's anything we've missed, please post a message again so we can check it out.

Cheers

edit: I've zipped and uploaded the wrong scene, apologies if you have already downloaded the previous link.

BEY206_025_010_IceCream_Melting_v006_15_02_fix.zip

[thhvfx]

Thanks Gosho. I will test out your fixed scene and let you know the result.

Cheers,

[thhvfx]

Hi Gosho,

I'm having another issue with the ice cream. So everything was fine except when I replace the low poly source with the hires one the fluid got some popping mesh. Please have a look at the play blast and you will know exactly what I'm talking about. I also attach 2 versions of the ice cream, they are the same scene just different emit source. Version 13 is good version with low poly, version 23 is version with the mesh issue.

https://drive.google.com/drive/folders/0B6D-0vS-tO7vbzQtOHNPWlJDVWc?usp=sharing

Thanks,

[Gosho.Genchev]

Hey thhvfx, sorry for the late reply. I had some trouble debugging your scene as the source geometry is quite heavy and there's a lot going on in there.

If I understand correctly, the problem is that using the highly detailed source causes issues with the liquid emission. May I suggest you use the highly detailed one as the initial fill geometry and the low detailed one as the source geometry for the Phoenix liquid source ? I tried that in my scene and it seems to work.

It's very important to try to keep the geometry for your simulation as clean as possible. Otherwise you're basically hacking the solver and the results are unpredictable.

One last thing - consider Mesh->Reduce to lower the resolution of your high res geo. The more polygons you have in your sources, the slower the simulation becomes. In the case of your scene, performance is taking a big hit because of that.

[Svetlin.Nikolov]

Most importantly - always make sure to use closed geometry - open mesh sheets are really not expected to work, so it may seems to work in one case, but will blow up in the next. If the simulator interacts with any open geometry, don't expect anything else to work properly.

[thhvfx]

Hi Gosho,

I did use both the low res and high res geo as initial fill geometry. Can you please send me the scene that works for you? Thanks.

Svetlin.Nikolov Thank you so much dude. My modeling skill is quite terrible I believe. I wish the modeler can give me a better mesh instead of open geometry. I will work on the model and resim again. Thanks

[Gosho.Genchev]

Hey thh, I'm uploading my test file for you but keep in mind I've deleted some nodes and changed things quite a bit to make sure everything is clean and working. It's unlikely that you'd be able to use it for anything other than a reference.