Chocolate on cookie ripples artifacts.

[jstrob]

Hi,
I did this test starting from the chocolate cookie sample scene. I used vray 3.0 new skin shader for the chocolate.

I would like to get rid of the ripples forming at the top. I guess they are due to the voxel grid. Should I increase resolution or just play with the render settings. Or maybe I could animated a clipping gizmo to place it at the top but it may be difficult to have a seamless cut and a perfectly smooth surface.



What are your thoughts?

[Ivaylo Katev]

could you share the scene?
we will try to find a solution, but will be better if you make few experiments changing the following parameters:
- non newtonian
- sharpness
- advection method
- surface tension

[jstrob]

Here it is

thank you!

[jstrob]

advection method: I tried the slow moving advection method, but it was giving a too much different sim. The liquid was getting spiky. I was using
sharpness: I had "strong surface mode" checked, so sharpness is not used.
non newtonian: I am simulating with a slight increase of the non newtonian parameter (1.8 to 2.2) I am waiting...
surface tension: I will try after with the surface tension (I try to change one parameter at a time at first).

In the meantime I have a question: would it be possible to mesh the liquid with frost (I got it) and then apply a relax on the mesh (with soft volume selection). If yes that would be the fastest fix I think.

[Ivaylo Katev]

well, i forgot to mention to exclude slow moving method, only classic and FT
frost is a particle mesher, the liquid is grid based, so it wouldn't work.
i remember that have observed this issue but can't remember in what situation.

[jstrob]

Ok. Iwill try FT then

Increasing non newtonian gives a liquid less visquous and the ripples happened a little later but still there. Increasing surface tension give a liquid with bigger drop and it result in bigger ripples too. Almost same ripples but with a bigger scale.

I also tried to fix the ripples by using the phoenix mesh output and using a gizmo cutter and it worked well on a single frame but now i have to try it by animating the gizmo to see if I can get a perfect animation that way. I also tried with a volume selection and a relax but for some reason the relax has a hard time making the ripple disapear completely.

But is Hybrido also grid based? Why can it use frost and not phoenix?

[Ivaylo Katev]

what about zero surface tension?

[Ivaylo Katev]

i just looked into your scene
i think your scene is based on the old chocolate sample, the viscosity is 0.97
so, my suggestion is to start from the new chocolate sample, that must be included in the current nightlies.
the viscosity in the new sample is 0.9 , because some a bug that decreases the viscosity effect was fixed.
the non newtonian is 0.1, if the scene is not installed for some reason

[jstrob]

I started from the new scene. But i increased non newtonian to 1.8 at some point to get the result in the video v001 posted at the first post here.

FT is very strange. The source keeps emiting in places were it was not anymore. And too many drops are created.

Non newtonian at 1.8 gives me the amount of spilling that I want at the end (around the cookie) If I increase it , the spill get larger. So I guess if I lower it the spill will be shorter.

I will try with non newtonian 1.8 and zero surface tension.

I am curious to try with same settings as my v001 (visc: 0.97; non N: 0.18; s.f.: 1.0) but with increased steps per frame let's say 25 instead of 8.

[jstrob]

And now if only I could get a simulating license or a rebate for a second one, it would go faster!

[jstrob]

With zero surface tension, the spills also gets too large. The dripping is larger too.

[Ivaylo Katev]

well, in the original scene there are no ripples. at which point you got them?
let start from the point which part of the sample scene result you would change?

[jstrob]

What I wanted to change in the sample scene was this:
1- the chocolate hit the floor instead of passing through it
2- liquid a bit thicker.
2.5- I also changed the gravity from 1 to 1.1
3- the source has a small back and forth motion to give the impression a human hand is holding the source.

4- I did not do yet in my v001 but I want the source to stop emiting a bit earlier. I tried animating discharge, but it did not work. It is working by animating the source out the box.


to change that I added the floor, animated the source and made it a little bigger. Increased viscosity from 0.9 to 0.97, increased non newtonian from 0.1 to 0.18 and then came the rilpples.

I will try to start back from the sample scene and see where the ripples got introduced. Maybe I should keep viscosity at 0.9 and lower non newtonian instead.

(p.s.: increasing steps to 25 from 8 gives the exact same result...)

[jstrob]

I took the sample scene and just animated the source so it got out around frame 230 and i simulated until 371. we can see some grid artifact but not really going into ripple. It's easy to get rid of them by just rendering with a mesh and applying a relax.



I have still to try with viscosity at 0.9 and lower non newtonian.

I will have to redo the test with more sample (25 instead of cause my first test in that regard was invalid (I forgot that I changed the input for rendering on the farm... so the simulation was calculating but it was still showing me the old one from the network!!!)

[jstrob]

I think i got it.

I did not remember , but I scale the cookie by 150% to get the scale I wanted. By scaling it , each polygon became bigger in regard to the cel size. I now added a subdivision level to the meshsmooth and it seems to create far less ripple. I will be able to get my same v001 sim but be able to remove the ripple with the relax. I'll post the result as soon as I get it!