View Full Version : 3 million of particles, velocity vectors test

07-21-2012, 09:31 AM
this is 3 millions of particles, coloring with distance to object, I wish we could have some sort of speed or velocity coloring here.

Sample very short clip here...

I could push this to over 8 millions of particles, Perhaps even more, god knows what the limit is since Lightwave 11 broke the 1 mil barrier.

That is at least on my desktop, but I cranked this out from my laptop, Lightwave will choke a little while cooking and letting particles settle when you set your high amount of particles, Perhaps Newtek can work a little on that to be more efficient.

Render times ..1280x720 per frame aproxim, 2 min and a few seconds.

I had some cool post processing with a little glow and extra turbulence distortions in AfterEffects wich gave an even more fluid look to it, but have to redoo that, this 3 million sample is a straight render.

I first tried a low res version on vimeo and youtube, uploade with the quicktime animation best settings, however both youtube and vimeo failed in
converting it decently.

Heres that version but only with 1 million particles and a 640 by 480 resolution.

Description as follows...

Lightwave particle field, flat field with 5mx5mx0 dimensions x,y,z axis.
motion driven by two fractal layers in particles velocity vector channels(not velocity force), one main wavy water from DPont nodes, one turbulence on top with multiply mode to give extra small undulation distortions.

Tiny hv sprites 5mm, and density lowered to 50 percent,( 15% on the 3million sample) thatīs it.

I could probably test with 6-8 millions particles too, not sure, this was done on my laptop and only 1 million
of particles, added some glow and crushing of levels and colorama to colorize a little.

Image attached are from the low res lower 1 million non colored version thou, only slight color from after effects and slight glow, shows original and some post process of that one of 1 million particles, but the newer 3 million anim sample on youtube is better, still some conversion artifacts.

Im generally loading up quicktime animation in best settings, thou that give huge files, I tried quicktime H264 but that gives wrong colors or bleached out movs.

the low res with 1 million version on vimeo here..came out with bad conversion thou, have to replace it...


07-21-2012, 10:17 AM
Nice result... hypnotic

07-21-2012, 11:03 AM
This is cool.

Mind to give a more detailed description how you did it? I tried to replicate from the given info but can't nail it. Doesn't have to be a Video or such, just a detailed step by step (1. do this, 2. do that etc). Did you use Nodes for this?


07-21-2012, 01:07 PM
This is cool.

Mind to give a more detailed description how you did it? I tried to replicate from the given info but can't nail it. Doesn't have to be a Video or such, just a detailed step by step (1. do this, 2. do that etc). Did you use Nodes for this?


Hmm..wasnīt the description on the youtube page clear, and also here in the thread?

Ill look over that, might post basic low res particle setup so you can se the motion etc, and simple voxel sprite added.

No nodes used.

I would like to localize the distortion fields by referencing a null with falloff aproximate 95 by 95 in y and x axis, that works ..but only if Im not moving the null, because the procedural texture distortion wont follow the moving null, Have to ask for some help there.

Got some more still samples up soon with postprocessing, will post animation samples of that a little later thou.

Anyway, that was a a particle emitter with 5x5 meter scale and 0 at z axis, you distort it in particles motion tab and use the velocity(m/s) texture buttons, and add procedural textures that is animated, like the wavy water procedural, but you have to download that from rman collection and dpontīs site.

you could try other procedurals such as ripple etc.

I think I can post the basic scene setup in low res format soon.


07-21-2012, 01:12 PM
here follows some post processing in AfterEffects, mainly turbulence distortion, added glow, and some color correction.

still images now, will post animation samples on vimeo or youtube, as soon as I sort out best compression and format etc, this effect is sensitive to saturation or gamma and lightness to give good impression, and I should really work on this with my ips monitor instead of my laptop.

Hereīs three different stages of postprocessing...





07-21-2012, 01:48 PM
trying to animate 1 million particles here, but feedback is too slow! render is fast!

07-21-2012, 02:48 PM
trying to animate 1 million particles here, but feedback is too slow! render is fast!

Yepp that is my conclusion too, the clogger is the feedback when using textures to drive particle velocity, or simply just adding the particles to the scenes when you have 1 million or more, 1 million works, 3 millions starts to be cumbersome, have to wait for some minute or so before particles settles.

And using wind dynamics...puh..that will be slow, textures driving motion directly in particles velocity channels are faster, unfortunatly they canīt follow a null reference moving it seems.

If I would go up to 8 millions it would choke the system almost, or at least you have to wait quite sometime for lightwave to respond, and it will probably make other work tasks in the background in windows slow.

So a better management of huge particle amount would be nice if they could work out.

I remember working with houdini wasnīt as nearly slow with much higher amounts of particles on the same machine.

Now then of course if we could have our own special custom particle renderer ala Krakatoa, that would be nice too, unfortunatly krakatoa for lightwave was out of the question when I asked those developers.
I think you can add particles in rendertime and using partitions for better handling of particles, so doing vortex distortions would be done in low res, but final render in higher res.
Now I really do not know to much about krakatoa, so I might shut my mouth up about that.

Anyway, still better today to be able to have the limit pushed so we can, with some pain produce higher amount of particles.

Also ...checkk your system specs again if it becomes to slow.

I really do not know what hardware part is responsible for handling particles, but I guess it is the Ram, or if it also has to do with the gpu ram, not sure the processor has any big impact on it.

My laptop isnīt as good as my desktop either, system specs on that one in my sign, I need to fix some things with it before I can work seriously.

07-21-2012, 05:56 PM
Very cool effect!

07-23-2012, 01:51 AM
trying to animate 1 million particles here, but feedback is too slow! render is fast!

Do not attempt working with textures and velocity vectors with 1 millions of particles when you set it up and tweak, You should use lower resolution of 50 000 to 100 000 perhaps so the system doesnīt become unresponsible.

You can get a good Idea anyway how it will look in the end, whey youre done tweking textures and motions, I would suggest save out one low res version with smaller amount of particles, and then increase particles to higher amounts when it feels ready for final, do a test render with 1 millions perhaps, and if that looks good, turn it up to 3 millions or more, and take a coffe or clean the house while the particles settles, then start render and go for jogging:)

It might be a good Idea to have specific hypervoxel presets ready for the different versions related to wich resolution you work with, one for 100 000 particles, one for 1 million and one for 3 million etc, and add them to preset shelf for faster workflow and testing of scene.

Might even cook that up myself and post, Itīs important to lower density in hv sprites in relation to when you raise the particle amount, and perhaps even lower hv particle size.

Glad It looked good Lino:)


07-23-2012, 01:57 AM
i often simulate with 10000 first, then 30000, and so on, and finally 1 mill....
saves time...

never jump directly to 1 mill...