emPolygonizer - metaballs for SOFTIMAGE

Discussions concerning plugins for SOFTIMAGE©
User avatar
origin
Posts: 619
Joined: 09 Jun 2009, 11:59
Location: warsaw

Re: upcoming: emPolygonizer - metaballs for SOFTIMAGE

Post by origin » 04 Jul 2009, 18:56

owei wrote:Hi there..!
@origin
...well, performance extremely depends on the level of detail and the post-iso-field operations. emPolygonizer deals quite fast with at least several thousands particles . The test-scene above runs between 15 fps (low detail) and 0,5 fps (hight detail shown above) with 15k cached particles ;)
thats outstanding..! the problem with pwrapper is that even with ultraultra low-level of detail and high particle count (lets say, 20k is high!) the meshing is uberslow. So what we got I not realtime but 'wait 30sec for SI to respond'

Well I don't know what post-iso-field operation is ;) but what I would love to see in your plugin is that some kind of 'push' operator of pwrapper, which makes blobs not so sphere-like :d I hope Im clear.
cheers!

User avatar
ardworx
Posts: 23
Joined: 04 Jun 2009, 12:01
Location: Stuttgart, Germany
Contact:

Re: upcoming: emPolygonizer - metaballs for SOFTIMAGE

Post by ardworx » 04 Jul 2009, 19:08

owei: ok, i'll ask the other way around. whats the maximum particle count one could mesh with lets say 8gb of ram i a q9300 for example? :D
...just nitpicking, will test it out myself when its out. :-

greetz,
ardworx

User avatar
owei
Administrator
Posts: 840
Joined: 03 Jun 2009, 17:25
Location: Siegen/Germany
Contact:

Re: upcoming: emPolygonizer - metaballs for SOFTIMAGE

Post by owei » 04 Jul 2009, 19:19

@origin
...well, post-iso-field operations can do exactly what you describe. Just compare the first post video with the one I posted today...but as I said, there will be some other algorithms included which will work different from the now included marching-cubes ;)

..here is a quick test with 1mio particles ;) ....about 1 fps

[flvprev]http://www.si-community.com/download/cl ... _particles[/flvprev]

cheers,
oliver

User avatar
owei
Administrator
Posts: 840
Joined: 03 Jun 2009, 17:25
Location: Siegen/Germany
Contact:

Re: upcoming: emPolygonizer - metaballs for SOFTIMAGE

Post by owei » 04 Jul 2009, 19:53

@ardworx
...you should be able to deal with 20mio particles, depending how much parameters your particles have. Remember, that ICE even needs much ram for its particles too. In those cases it is better, to emit "own" particles via "sample set" or something and giving them only those parameters you really need ;)

User avatar
Tekano
Posts: 488
Joined: 09 Jun 2009, 14:49
Location: London, UK

Re: upcoming: emPolygonizer - metaballs for SOFTIMAGE

Post by Tekano » 05 Jul 2009, 12:09

Excellent, that's a huge improvement with that 2nd SPH video. well done! :)
Gossip is what no one claims to like, but everybody enjoys.

kotsoft
Posts: 4
Joined: 07 Jul 2009, 14:45

Re: upcoming: emPolygonizer - metaballs for SOFTIMAGE

Post by kotsoft » 07 Jul 2009, 14:59

this is looking good! will it be supporting things other than spheres, like ellipsoids and stuff? also will it support things like "relax" and "smooth" in realflow?

also, i think this might be a good read:
http://multires.caltech.edu/pubs/meshextraction.pdf

also, the current implementation is based on http://www.iro.umontreal.ca/labs/infogr ... S/pvfs.pdf. so we actually use two kernels for the fluid. i've done some 2d tests with this and i think if you match the isosurface stuff with the kernels in the SPH, it should improve the results a lot and fit the simulations a lot better. maybe it could be an "auto" setting?

User avatar
Mootzoid
Administrator
Posts: 243
Joined: 03 Jun 2009, 18:48
Location: Germany/France
Contact:

Re: upcoming: emPolygonizer - metaballs for SOFTIMAGE

Post by Mootzoid » 07 Jul 2009, 15:38

@kotsoft

Hello, nice to have you here! I of course know who you are as I was thinking about implementing some SPH stuff in ICE last year, but then I saw that Thiago Costa was already working on that, so I prefered to do something else. Simply love your real time physics demos! :p

The two links are very interesting, especially the mesh-extraction stuff, thanks for sharing them.

I have lots of cool things planned for emPolygonizer. Tomorrow (hopefully!) I will release version 1.0. It already is a pretty nice version and the workflow and the performance are both quite good I think. The main work for this version 1.0 was to code the heart (or body) of this plug-in. Now that that work is done it is relatively easy to implement new features (or "algorithmes" as I call them).
This first version only has a basic marching cubes algorithm implemented. It works fine and produces nice results, but as the name already suggests, it is only basic and has a certain "blobbyness" to it, but I made some test with over 250.000 particles and my little laptop (which is running a crappy 32 bit OS) did just fine :-)
Next versions will support more sophisticated algorithms that will be designed especially for fluid/liquid things, a little like Pwrapper or the metaballs from motionblur.it ... but faster!!

One thing is sure: there will be many updates :D
...the current implementation is based on...
What current implementation do you mean? The one of Thiago Costa's SPH fluids in ICE?

cheers,
Eric

kotsoft
Posts: 4
Joined: 07 Jul 2009, 14:45

Re: upcoming: emPolygonizer - metaballs for SOFTIMAGE

Post by kotsoft » 07 Jul 2009, 16:04

yeah, the ICE sph right now uses the two density kernels stuff to give it the surface tension effects and try to prevent the usual particle clustering in sph. so i'm thinking that by using the same kernels you should be able to get the mesher to see the particles as the simulator sees it.

thiagocosta
Posts: 5
Joined: 07 Jul 2009, 16:01
Location: Montreal QC
Contact:

Re: upcoming: emPolygonizer - metaballs for SOFTIMAGE

Post by thiagocosta » 07 Jul 2009, 16:05

Hi Eric and Oliver, nice work guys!
I will get in touch with you guys offline.

User avatar
Mootzoid
Administrator
Posts: 243
Joined: 03 Jun 2009, 18:48
Location: Germany/France
Contact:

Re: upcoming: emPolygonizer - metaballs for SOFTIMAGE

Post by Mootzoid » 07 Jul 2009, 16:12

okay, I get it.
Is that data available as normal ICE data?

thiagocosta
Posts: 5
Joined: 07 Jul 2009, 16:01
Location: Montreal QC
Contact:

Re: upcoming: emPolygonizer - metaballs for SOFTIMAGE

Post by thiagocosta » 07 Jul 2009, 18:16

Yes, it's sph_q and sph_q2
Let me know if you can find it

User avatar
CiaranM
Posts: 87
Joined: 09 Jun 2009, 01:37
Location: London
Contact:

Re: upcoming: emPolygonizer - metaballs for SOFTIMAGE

Post by CiaranM » 08 Jul 2009, 02:13

Very impressive, Eric. :ymapplause:

bryitis
Posts: 35
Joined: 09 Jun 2009, 00:28

Re: upcoming: emPolygonizer - metaballs for SOFTIMAGE

Post by bryitis » 08 Jul 2009, 05:22

Things just keep getting better around here! Great work Eric, i look forward to playing with this.

User avatar
Mootzoid
Administrator
Posts: 243
Joined: 03 Jun 2009, 18:48
Location: Germany/France
Contact:

Re: upcoming: emPolygonizer - metaballs for SOFTIMAGE

Post by Mootzoid » 08 Jul 2009, 23:50

@everybody

thanks, guys :D ... glad you like the new tool !

I wanted to release version 1.0 today, but an unexpected wave of extremely unpleasant bugs spoiled the release X( ...coding can be such a fight... well, I got rid of most of the bugs, there's only one left in the Linux version and it's strong, angry and has fangs :((

cheers
Eric

User avatar
Mootzoid
Administrator
Posts: 243
Joined: 03 Jun 2009, 18:48
Location: Germany/France
Contact:

Re: upcoming: emPolygonizer - metaballs for SOFTIMAGE

Post by Mootzoid » 09 Jul 2009, 11:43

I finally got rid off all the bugs :-bd
So today should be "release day", finally.

Stay tuned!

User avatar
Hirazi Blue
Administrator
Posts: 5107
Joined: 04 Jun 2009, 12:15

Re: upcoming: emPolygonizer - metaballs for SOFTIMAGE

Post by Hirazi Blue » 09 Jul 2009, 12:08

Mootzoid wrote:I finally got rid off all the bugs
Famous last words??? :D
Stay safe, sane & healthy!

Post Reply

Who is online

Users browsing this forum: No registered users and 19 guests