Can this effect be replicated in LW (Gravity mist particle fx on solar panels)

Planeguy

Active member
I have phoenixfd in 3dsmax, but want to simplify the process as much as possible by not having to export an alembic or vdb into LW, but instead do it all in LW more naturally. Do you think HVs in 2019-2020 would be able to look anywhere close enough to this? I'd be happy if the particles look 90% as close.

If you had to do the mist/volumetric effect on the solar panels, how would you approach it using LW?

If it doesn't start at the time mark, it begins at the 1:11 minutes
(


(guys, keep it on topic please)
 
Well, decided against it. Particles just don't look right. Maybe doing a vdb might do the trick as it's more smooth for lack of a better word. So gonna simulate it in max and render the particles in LW.

Final_Render_026.png
 
Could you not take the LW object into Max to simulate, then render in Max as a pass (I don't know if you can import the camera settings....?)
Then comp it in Fusion/After Effects.
 
Could you not take the LW object into Max to simulate, then render in Max as a pass (I don't know if you can import the camera settings....?)
Then comp it in Fusion/After Effects.
The thing is that I want to render it all in LW to keep consistency.
 
There is always TurbulenceFD, but that's additional cost.
Turbulence would do the trick indeed, just can't get another tool. Have to use what I have at hand. Wish the vdb setup in LW had been more intuitive easier to use.
 
As I think about the problem, I'm considering what the particles in the movie are actually made from, and how this is definitely something that LightWave has dealt with before.

The particle fineness is very similar to the type produced by fog or volumetric lights and I think if I were to attempt this, I'd be thinking of a way to imitate the effect by using an animated volumetric light, since that's effectively what is happening in the movie.

The particles of dust surrrounding the space station's solar panels are effectively being disturbed by it's motion through it and is being illuminated by the planet and sun. This is almost exactly what volumetric lights simulate. They simulate the illumination of very fine particles in the atmosphere, so it kind of stands to reason that this option is a technique worth exploring.
 
As I think about the problem, I'm considering what the particles in the movie are actually made from, and how this is definitely something that LightWave has dealt with before.

The particle fineness is very similar to the type produced by fog or volumetric lights and I think if I were to attempt this, I'd be thinking of a way to imitate the effect by using an animated volumetric light, since that's effectively what is happening in the movie.

No..I wouldn´t go with volumetric lights, the effect is dynamic and particle based, you would struggle for sure with just vollumetric lights matching the objects.
True particles or true fluid dynamics is the way to go.
As for Not having TFD for fluids, it leaves you with two-three options...

1. use lightwave native fluid Gas solver
2. use blender (2.79. old fluid smoke recommended) and import in to lightwave
3. use Houdini to simulate the fluid and import to lightwave as vdb, though you can not use it commercially unless having indie version, for your own project it is allowed though.

Particles, it is doable with particles, and there are two ways foremost..

1. use particles with not so many particles, and let Old legacy hv sprite textures with a bit larger hv size and texture effects drive the secondary small undulation movement that breaks up the main direction motion.

2. use particles with a huge amount of particles, tiny tiny ..extremely small particles with very low density, render out, blur in compositing with fusion or after effects


Imagine similar stuff like this, but with a lot ..lot more particles, less density value, and blurring it, apart from the other aspects as the motion behaviour...


Or.. (only 363 000 particles, you should aim for millions, then blur comping, also possible to save to pxf to reload and offset particles) and some other tiny tiny particle effects.




And main dynamic principles but should be enhanced with millions of particles and using hv sprites in small size and density as shown above and you should get there..

 
Well, decided against it. Particles just don't look right. Maybe doing a vdb might do the trick as it's more smooth for lack of a better word. So gonna simulate it in max and render the particles in LW.

View attachment 152747

Actually .. I think fluids may be to over do it in a case like this, and why?..
Because the kind of motion needed doesn´t require fluid simulation to look good, and you would have to wrestle with the fluid simulation to behave correctly much more than working with particles, from your screenshot, you just have to large particles, to few of them, no initial motion undulation, and it doesn´t fade away per particle age or speed.
It could also use hv velocity stretch.

I think in many of these pfx cases in many movies of shuttles, probes or spacestations entering atmosphere etc, they have often used standard particles with post processing, it´s easier to tweak particles and blur in post.
 
To note about the Gravity movie, the particle life or density seen in the vapor from the panels are very very short lived, they fade off quite fast, and the direction of them all are a bit scattered and not so straight, so you may need to setup a couple of collision objects to do that, the rest is most about to get density and blur right and the fading off of it all.
 
Some blurring samples, and I would also suggest setting up the pfx in seperate segments, so on one panel you work with a Lot of particles, when that looks good, clone it to other panel parts, and render them out seperately as well I would think would be the best, that way you can use more particles per part than trying to acheive it all at once which would slow down the system more.

1.jpg

2.jpg
3.jpg

4.jpg
5.jpg
 
Actually @Planeguy, the more I think of it, the more I seem to remember how in the old days, we used to use animated textures in volumetric lights to fake illuminated particle movement in jet exhaust engines and such like.

So if used subtly, and planned properly, I can see volumetric lights working quite well. The added bonus would be a speedy render too. Of course you'd have to be familiar with the little tricks and techniques that would deliver the effect, but it can definitely be done.

It's also worth noting that in many cases when working on certain types of shots, the rule of "Fake It Till You Make It" is really a lot more preferable to doing the shot the "realistic" way.

Of course, this all really depends on the project and the shot requirements, but yes, LightWave definitely has many workarounds to make it possible. You just need to decide which route you want to take. Real or fake.
 
Actually @Planeguy, the more I think of it, the more I seem to remember how in the old days, we used to use animated textures in volumetric lights to fake illuminated particle movement in jet exhaust engines and such like.

So if used subtly, and planned properly, I can see volumetric lights working quite well. The added bonus would be a speedy render too. Of course you'd have to be familiar with the little tricks and techniques that would deliver the effect, but it can definitely be done.

It's also worth noting that in many cases when working on certain types of shots, the rule of "Fake It Till You Make It" is really a lot more preferable to doing the shot the "realistic" way.

Of course, this all really depends on the project and the shot requirements, but yes, LightWave definitely has many workarounds to make it possible. You just need to decide which route you want to take. Real or fake.
Animated textures and volumetric lights can not properly depict the emission from solar panels and how it is spawning from the very same, you need true collision and particles for that.
If you aim for a midlevel effects, by all means go volumetric lights, jet exhaust engines are different though, if you wan´t to aim for a level higher then mid level effects, to go top notch, use particles or fluids, I am pretty convinced particles are what they used for the VFX of Gravity.

My advice is against volumetric lights and go with particles foremost, secondly fluids.
 
As I think about the problem, I'm considering what the particles in the movie are actually made from, and how this is definitely something that LightWave has dealt with before.

The particle fineness is very similar to the type produced by fog or volumetric lights and I think if I were to attempt this, I'd be thinking of a way to imitate the effect by using an animated volumetric light, since that's effectively what is happening in the movie.

The particles of dust surrrounding the space station's solar panels are effectively being disturbed by it's motion through it and is being illuminated by the planet and sun. This is almost exactly what volumetric lights simulate. They simulate the illumination of very fine particles in the atmosphere, so it kind of stands to reason that this option is a technique worth exploring.
That's a great idea! Gonna test some volumetric lights around and see how it looks. Cheers!
 
Some blurring samples, and I would also suggest setting up the pfx in seperate segments, so on one panel you work with a Lot of particles, when that looks good, clone it to other panel parts, and render them out seperately as well I would think would be the best, that way you can use more particles per part than trying to acheive it all at once which would slow down the system more.

View attachment 152748

View attachment 152749
View attachment 152750

View attachment 152751
View attachment 152752
Those look great! Thanks for the screenshots. Gonna give it a try. The effect is indeed very short lived, but the way you got it looking is 90% there. I do wonder how the last screenshot would have looked animated. The render before that one, looks very similar to the movie.

The problem I'm having is, once I got the particle size in the milimeters, I had to increase the particle limit to 200k with a birth rate of 100k and that destroyed my system lol. However, I think I may be able to go up to 300k, maybe 400k for one panel. The higher I go, the less responsive the viewport becomes however. It renders relatively fast though, but it becomes impossible to work with.

going back to layout to try these settings. Appreciated
 
Thanks. Will try this too. I'm increasing the particle amount and see how much my system can handle first.
Nothing wrong with testing volumetric lights, but from my experience, I don´t think that´s the way to go, nor textures on polyplates.

As for my last screenshot which was blurred with vector blur in after effects and natural mode which has a very interesting effect that can be further altered in many ways, I just don´t have the time now to render out animations, these things probably take some days, weeks to research and render out ..if you aim for top notch effect level.

If particle amount is an issue, then perhaps the other way with fluids, gas solver is there natively but it would require you to learn how to direct the smoke etc, and damped the bouyancy and density, temperature lift of the gas which is there by default, and how to push the gas with the right vector nodes, it´s by far more complicated to set up and control than using blender 2.79.

Sample below, just used a cylinder, added quicksmoke effect, and painted area where I want the emission to be, make sure to select the vertex group for the painted weight map in smoke advanced tab, turned off temp diffusion which raises the smoke otherwise, which then would yield the smoke not moving, then add a force wind to push the smoke where I want it.

change cache method to openvdb, and just play to the frame I want, the files are then just cached saved to desired folder.
Import to Lightwave, rotate pitch 90 degree, set up emission in this case only for the vdb, enter nodes, add openVDB info and select density channel, then a gradient curve with tweakings on the alpha, then plug that gradient node in to the emission channel ..which will soften damp the emission in a much better way than the direct texture channels in the vdb main panel.

Blender fluid smoke weight paint.jpg

Blender vdb to lightwave.jpg


just a Quick test so i didn´t match it with solar panel objects, or other space mesh items, and it´s also without collision..had I more time I would have worked with that as well.
Gas solver natively is as mentioned also an option, but I think it would be harder to learn for you than getting up and running with quick smoke in blender (2,79) painting in areas to burn or emitt, work with forces in blender..and just learn to save out the cached file and import to lightwave.
Density is exaggerated just to be more visible in this showcase, the end result would need a lot more damping to be more subtle in the end.

Now I have to make early dinner, woke up at 3 a clock early morning today since I went to bed at 19:00 early evening yesterday, so I need to drop the testing for some hours.
 
Last edited:
Nothing wrong with testing volumetric lights, but from my experience, I don´t think that´s the way to go, nor textures on polyplates.

As for my last screenshot which was blurred with vector blur in after effects and natural mode which has a very interesting effect that can be further altered in many ways, I just don´t have the time now to render out animations, these things probably take some days, weeks to research and render out ..if you aim for top notch effect level.

If particle amount is an issue, then perhaps the other way with fluids, gas solver is there natively but it would require you to learn how to direct the smoke etc, and damped the bouyancy and density, temperature lift of the gas which is there by default, and how to push the gas with the right vector nodes, it´s by far more complicated to set up and control than using blender 2.79.

Sample below, just used a cylinder, added quicksmoke effect, and painted area where I want the emission to be, make sure to select the vertex group for the painted weight map in smoke advanced tab, turned off temp diffusion which raises the smoke otherwise, which then would yield the smoke not moving, then add a force wind to push the smoke where I want it.

change cache method to openvdb, and just play to the frame I want, the files are then just cached saved to desired folder.
Import to Lightwave, rotate pitch 90 degree, set up emission in this case only for the vdb, enter nodes, add openVDB info and select density channel, then a gradient curve with tweakings on the alpha, then plug that gradient node in to the emission channel ..which will soften damp the emission in a much better way than the direct texture channels in the vdb main panel.

View attachment 152753

View attachment 152754


just a Quick test so i didn´t match it with solar panel objects, or other space mesh items, and it´s also without collision..had I more time I would have worked with that as well.
Gas solver natively is as mentioned also an option, but I think it would be harder to learn for you than getting up and running with quick smoke in blender (2,79) painting in areas to burn or emitt, work with forces in blender..and just learn to save out the cached file and import to lightwave.
Density is exaggerated just to be more visible in this showcase, the end result would need a lot more damping to be more subtle in the end.

Now I have to make early dinner, woke up at 3 a clock early morning today since I went to bed at 19:00 early evening yesterday, so I need to drop the testing for some hours.
openvdb in lightwave I couldn't get it always to do the thing I wanted.

The gas solver you made in LW above doesn't look bad at all. Some more fine tuning, and transparency, and it would look pretty similar. Appreciated
 
openvdb in lightwave I couldn't get it always to do the thing I wanted.

The gas solver you made in LW above doesn't look bad at all. Some more fine tuning, and transparency, and it would look pretty similar. Appreciated
1. don´t work with gas solver unless you have worked with it many months before 😁
2. That sample above wasn´t the gas solver, it was blender 2.79 smoke, emitted from a painted weight map, then saved out as vdb, then imported to lightwave and setup in the vdb, but no gas solver involved.

Results can later be colored with gradients in the vdb node in lightwave, can be great for nebula gas or trail color effects ala armaggedon asteroids...

Same blender vdb sample as previous post, but some changes in emission scale and color gradients.

Now all these samples are just made within some one hour or so with a hungry stomack, so imagine putting some real effort in it for some weeks.

vdb coloring.jpg
 
Back
Top