Page 1 of 1

A test using realflow, i'm getting problems with geometry.

Posted: Tue Jul 19, 2005 4:06 am
by ctl3d32
Hi all!

I'm getting bad geometry using realflow. Realflow gives me holes in the
fluid mesh. Tried to use 3dsmax CAP HOLES function, but it didn't work
100%.

Thanks for reading

Image

Image

[[[[[[Scene Setup TAB]]]]]]

Geometry:

Num Meshes: 7
Num Triangles: 269588
Num Vertexes: 135187
Num Normals: 131651

Materials:

Num Emitters: 1
Num Diffuse: 4
Num Metals: 1
Num Plastics: 0
Num Dielectrics: 1

Bitmaps:

Num Bitmaps: 0

[[[[[[== Render settings == ]]]]]]

resolution: 1024 x 614
maximum desired time: 1h 40m 0s
threads: 2
target sampling level: 25.00

>> Starting Render

>> Start Voxelization...
>> End Voxelization...


[0h 0m 27s] of [1h 40m 0s] , current sampling level : 0.00

[0h 1m 10s] of [1h 40m 0s] , current sampling level : 0.00

[0h 2m 12s] of [1h 40m 0s] , current sampling level : 0.85

[0h 3m 36s] of [1h 40m 0s] , current sampling level : 2.72

[0h 6m 7s] of [1h 40m 0s] , current sampling level : 4.51

[0h 10m 54s] of [1h 40m 0s] , current sampling level : 6.26

[0h 20m 5s] of [1h 40m 0s] , current sampling level : 7.99

[0h 30m 5s] of [1h 40m 0s] , current sampling level : 9.07

[0h 40m 3s] of [1h 40m 0s] , current sampling level : 9.81

[0h 50m 3s] of [1h 40m 0s] , current sampling level : 10.38

[1h 0m 3s] of [1h 40m 0s] , current sampling level : 10.85

[1h 10m 3s] of [1h 40m 0s] , current sampling level : 11.24

[1h 20m 8s] of [1h 40m 0s] , current sampling level : 11.57

[1h 30m 4s] of [1h 40m 0s] , current sampling level : 11.87

[1h 40m 22s] of [1h 40m 0s] , current sampling level : 12.13


Render finished successfully :)


Close this window or the frame buffer to End Session...

Posted: Tue Jul 19, 2005 4:10 am
by Maximus3D
Try increasing the total particlecount and generated meshdensity in Realflow, i think you should be able to do something like that there to get rid of the holes that i guess appear in your too lowpoly generated fuid mesh.

/ Max

hi

Posted: Tue Jul 19, 2005 4:16 am
by ctl3d32
Fluid density is set to 10.0 and Metaballs (Fluid Mesh) to 0.01. I thought it would give me a very high polycount (and it did). Let's do it again!!!!

Thanks for reading

Posted: Tue Jul 19, 2005 4:22 am
by Maximus3D
Alright.. well i'm not a Realflow user so i have no idea what does values mean. But 270,000 polygons for the entire scene including the fluid which take up the most isn't really that much. You could have atleast the double that. Also check your total particlecount, that should work in combination with your meshdensity to calculate the final fluid mesh detail look (i guess).
Is this a static or dynamic fluid mesh ? (animated)

/ Max

hi

Posted: Tue Jul 19, 2005 4:31 am
by ctl3d32
Hi!

It's a ten seconds animation. I choose 1 frame to render. I was looking
at my fluid configuration and i think i found the problem. I was giving the
fluid a fractal noise ( to avoid make it look too smooth ) of 0.8 in amplitude.
I have reduced it to 0.5 and doubled the polycount. Decreased frequecy of
the noise from 4.0 to 1.0. That seemed to solve the problem, but increased
simulation time. Let's see.

i will let

Posted: Tue Jul 19, 2005 4:50 am
by ctl3d32
Hi!

I will let it render overnight just to see the result!

Thanks for reading!!!

Posted: Tue Jul 19, 2005 8:20 am
by Becco_UK
Nice render. I sometimes get a few single polygon holes with meshes from RealFlow3 which are more noticable for static images.

Adjusting parameters in the filters and optimize boxes of RealFlows' mesh section may help.

Hi!

Posted: Tue Jul 19, 2005 1:58 pm
by ctl3d32
Hello All!

This is a little update.

Image

MAXWELL 1_2_2a_beta ( Engine core: rs1c )
[ Physically and Unbiased Based Rendering ]


## Scene Information ##

7 meshes ( 116176 triangles )

## Render Information ##

*settings:

resolution: 1024 x 614

*render time: 5 hours 0 minutes 0 seconds

target sampling level: 25
sampling level reached: 16

MaxwellBench 1.0: 24.160499

Becco

Posted: Tue Jul 19, 2005 2:02 pm
by ctl3d32
Hi Becco!

It was just the filter setting that was giving me holes in the mesh. I was
giving some it relaxation...


Thanks for reading!!!!

Hi

Posted: Tue Jul 19, 2005 2:17 pm
by ctl3d32
Hi all!

Here is a Wire of the camera view!!

See Ya!

Image

Posted: Tue Jul 19, 2005 2:44 pm
by Becco_UK
ctl3d32: Looking at the wireframe image you could get a big polygon count reduction by using the optimise function (using curvature setting) in RealFlow. Edge detail is preserved by retaining high poly's on the areas needed and the flatter/gentle curved areas are reduced quite good.

You probably already knew this already but may be useful for other people new to RealFlow.

Posted: Tue Jul 19, 2005 9:13 pm
by Becco_UK
Personally, if you revisit this, aim for at least 150,000 particles. Particle count has a direct relationship to final quality of the RealFlow mesh. No matter how much you adjust other settings nothing compensates for having too few particles.