BerconMetaball 1.10

[MasterBercon]

Somebody else complained about the same thing but I haven't been able to reproduce that problem myself. All my DR tests have come out perfectly. Did you make sure you updated the the plugin to all the nodes too so they ain't running the old version?

[simmsimaging]

I went through and re-installed on all nodes to be sure. Same issue. It fails with any one node as well as with all.

I have zipped up a simplified version if you want to take a look:

Please download it by right / control clicking the link below. Choose "save as" / "download":

http://www.simmsimaging.com/upload/f..._Issue.zip.zip (66.9 KB)


It works fine as long as you are not using the particle size from Pflow.

Thanks.
b

[Da_elf]

why does this look like ISO

[dbuchhofer]

probably because by definition it is? :P

[Da_elf]

i just thought of that hehe. i was refering to dimo's iso but yours is particle based his is texture based

[MasterBercon]

Well metaballs are simply based on distance to point cloud. If I would implement distance to mesh it would be the same things as dimos iso plugin. However computing distance to triangle cloud (mesh) isn't that easy, when it has to be very very fast. If you guys run into some nice algorithm to do it fast I'd be happy to implement it.

simmsimaging, I'll check that on monday when I get back to work. Using DR through VPN is pretty slow.

[percydaman]

kinda offtopic, but thought this was cool.

creating metablobs in photoshop.

http://www.youtube.com/watch?v=SWad5QyjoaA

[tricky]

kinda offtopic, but thought this was cool.

creating metablobs in photoshop.

http://www.youtube.com/watch?v=SWad5QyjoaA

Now that is very cool!

[saudade]

Well metaballs are simply based on distance to point cloud. If I would implement distance to mesh it would be the same things as dimos iso plugin. However computing distance to triangle cloud (mesh) isn't that easy, when it has to be very very fast. If you guys run into some nice algorithm to do it fast I'd be happy to implement it.

simmsimaging, I'll check that on monday when I get back to work. Using DR through VPN is pretty slow.

Thanks for the plugin, you are a genius!
In these days I'm playing with pflow and and I can say that your plugin can change the way to render liquids.
The geometry generated is like a cream, is very hard to control the elasticity or relaxation (like realflow does). Do you think is possible to add this feature?

Regards

[saudade]

a simple test...

[MasterBercon]

Unfrotunately since the plugin is not working with a mesh, you can't relax it. The only way to get similar effect is to play with the Threshold value and Radius. Using high threshold (2.0 for example) and large radius should give you something relax-like but its not the same thing.

You can also try the Use average mode and play with cutoff/trim values.

I still haven't found any problems with the DR. I wonder if the problem might be pflow not computing Size for the particles...

[simmsimaging]

I have not had time to come back to it, but I was half waiting to see what you found anyway. Are you able to reproduce the problem at all?

b

[saudade]

another test

[MasterBercon]

I've now managed to complete the 1.11 version and I'll just have to compile it for all max versions.

New features currently:
- Support for simple particle systems (basically all particle systems that come with max)
- Supports >1.0 values in texture strength in Multiply mode
- Slightly faster (Sorting bounding boxes along the ray is no longer brute force, this is most noticable with >1 000 000 particle counts)

I can't think of anything else to add, fix or improve so it should also be the final version.

I still haven't been able to reproduce the DR bug. Rendering on multiple machines with variating blob size works just fine: http://www.niksula.cs.hut.fi/~jylila...taballs_31.jpg

So I asume this bug is not related to the metaballs but has something to do with pflow and different machine setups.

EDIT: With 10 000 000 particles the render time decreased from 34 seconds to 27 seconds. This is a lot considering the preperation time (which is resolution independent) is quite long with that many particles. The new BB sorting affects only the actual rendering process so it might actually cut the rendertimes by 25-50% in some rare situations. I didn't bother doing any extensive testing on it.

[percydaman]

excellent work bercon. I see this mentioned all over the place. Whats your next pet project?