Blender particle motion to Lightwave

[clintonman]

View attachment 122034
A couple of days ago I discovered that Blender has particle based fluids as well as the domain based fluids. A video here http://vimeo.com/21617994 shows how to transfer particle systems from Blender to Modo via mdd files so I adopted the technique for Lightwave. The process requires a customized Blender Addon so the MDD files read properly into Layout. The tutorial only shows the basic process not how to get a good result. I've only taken the particle count as high as 25000. Much higher counts would be needed for a decent result.

http://clintons3d.com/tutorials/blenderFluidsForLightwave_2.html


The Blender domain based fluids to Lightwave tutorial can be found here:

http://clintons3d.com/tutorials/blenderFluidsForLightwave.html

 

[lertola2]

Wow. Thanks for this. That is very interesting.

 

[clintonman]

I did another test with 100,000 particles and the results were the same as the 25,000 particle test and it was dead slow to render.

View attachment 122139
Link to the first test video:
http://youtu.be/QTXK4AYSk-8

View attachment 122140
Did a third test again with 25,000 particles and changing the Hypervoxel settings made a big difference in the appearance. The trick was to reduce the hypervoxel Blend scale to a small value, 3%. It's not perfect but getting closer to something useful.
Link to the third test video:
http://youtu.be/G8jcPcuMLEk

 

[clintonman]

View attachment 122160
Getting closer. This test uses fewer particles than the other tests, 10000 spread out over more frames, and uses a new Blender particle to Lightwave pfx script so that particles can die and don't have all be born at the start like in the mdd method. This should make it easier to scale up the particle counts later. Another benefit is that Hypervoxel stretch option helps to merge the faster moving particles together reducing a lot of the blobbiness. The new scene also has improved materials and a texture environment to reflect. Simulation time was a couple of minutes and render took a little over 6 hours.
Video:
http://youtu.be/HJF6gfGWY6o

Blender Particle Converter Python Script
http://clintons3d.com/plugins/lightwave/index.html

 

[clintonman]

I got your message, whodini. The hypervoxel settings are attached here.

 

[Shiny_Mike]

Thanks Clintonman! Also works for particles sim'd with the "Molecular" addon: https://www.youtube.com/watch?v=_5QkPnPDcfc

to clarify, the mdd/single point mesh trick worked, haven't tried your script yet

 

[clintonman]

Thanks Clintonman! Also works for particles sim'd with the "Molecular" addon: https://www.youtube.com/watch?v=_5QkPnPDcfc

to clarify, the mdd/single point mesh trick worked, haven't tried your script yet

Neat! What is the mdd/single point mesh trick?

 

[Shiny_Mike]

Neat! What is the mdd/single point mesh trick?

Just meant using the particle instance modifier on a single point mesh and mdd export like in your vid. (then using MDD on lwo made of array of points in LW for instancing)

 

[stevecullum]

I gave your Blender particles to PFX script a try, but I keep getting errors - any idea what the issue could be?

I'm using Blender 2.75a and Lightwave 2015.3

Thanks

View attachment 129531

 

[clintonman]

I gave your Blender particles to PFX script a try, but I keep getting errors - any idea what the issue could be?

I'm using Blender 2.75a and Lightwave 2015.3

Thanks

View attachment 129531

Can you share the Blender file. It looks like it doesn't like the file coming from Blender. Nothing Lightwave specific at that point in the code where the error occurs.

 

[stevecullum]

Here is the first 20 frames. Thanks for having a look!

 

[clintonman]

I just found out that pressing the Bake button creates an extra bphys file. Try generating the particles by pressing the play animation button in the timeline instead.

 

[stevecullum]

Ah ok - that works now - great! Thanks for your help and the handy script

 

[clintonman]

Updated the script to fix the extra bphys file bug. Works with Bake button and compute from timeline methods.

 

[stevecullum]

Great! - thanks for the update