Baking Hypervoxels

Is there a feature for baking new hypervoxels in LW 2019? Can't find any references to it. Thanks for all the assistance over the years.
 
Old hypervoxel feature doesn't have new shading features and gives you options for sprites and slices. New hypervoxel feature uses new shading methods but I don't see a way to bake them out to make render times more manageable.
 
No, the new volumetric shading system listed in the Object Properties tab_Primitive Type_ Volumetric_Use Nodes_Edit Nodes and/ or adjust Emission/ Scale, Scattering/ Scale Etc. When adjustments are made to the Voxels I would like to bake them out so that calculations don't have to be made at render time... just use the baked file at render time instead.
 

if the option is gone i would presume they left it out in the 2019 version.
was never a big fan of baked HyperVoxels myself.
options today could be instanced cloud objects.
not the same, but an alternative.

 
Hey Erikals, not sure if the option is gone or placed in some other tab. I never had a need for any such thing until now. Thanks for the tip. Will look into that.
 

prometheus

REBORN
if the option is gone i would presume they left it out in the 2019 version.
was never a big fan of baked HyperVoxels myself.
options today could be instanced cloud objects.
not the same, but an alternative.

You can´t bake the new volumetrics from volume item, you can´t even convert a volume item to a vdb format.
You can bake the old hypervoxels out the same though in 2019 as previously.
Instancing is something completely different and has nothing to do with reducing a volume to image baked slices, and you can´t instance a volume item.
 

lardbros

Not so newbie member
You can´t bake the new volumetrics from volume item, you can´t even convert a volume item to a vdb format.
You can bake the old hypervoxels out the same though in 2019 as previously.
Instancing is something completely different and has nothing to do with reducing a volume to image baked slices, and you can´t instance a volume item.
You can convert a volume into a vdb, just for info
 

prometheus

REBORN
You can convert a volume into a vdb, just for info
No..that I do not think is correct.

As I am aware of ..there is no way you can for instance turn a primitive item volume mode in to a vdb file, that I have myself asked if it was possible to do..especially since I would have liked to have my volume item clouds with the fractals turned in to vdb and exported then to blender, but as far as I know..no one has verified it can be done, nor have I found any way to see how it would be possible.

You are probably refering to other volume types, shape item or meshes, but that defeats the purpose of trying to "bake out" the newer so called hypervoxels(yes they are called that as well sometimes)

For further conversions of "volumes" to vdb, you would have to work with mesh to volume, or fog to level sev, gas solvers and such, but there Is NO WAY ..you can get the volume item volume data with fractals in to that and convert to vdb.

If you by any chance have stumbled upon a way, I would bow in front of you and ask ..how? but I urge you to check that properly by makein a primitive item volume mode such as clouds with fractals and convert that to a VDB if you think that is possible :)
 

lardbros

Not so newbie member
Ohhh, you mean the volume in the 'Primitive Type' dropdown?
What's the benefit of using that method, over creating the volume entirely within vdb and with any mesh shape you want?
 

prometheus

REBORN
Ohhh, you mean the volume in the 'Primitive Type' dropdown?
What's the benefit of using that method, over creating the volume entirely within vdb and with any mesh shape you want?


Yes, I mean exactly that, since that is in essence what richard herbert would need when asking for baking hypervoxels (assuming volume and not surface)

1. Fractals: You can´t add fractals the same way on to a volume, there´s two ways mainly in the vdb tools, where one is only affecting the surface of a vdb volume, not the edges, the second is to use advection noise driven by a fractal, this has to be simulated through playing the timeline and is a huge time vaster and hard to get feedback until it is simulated and can be very unpredictable.
Or you could simulate with gas solver with fractals, but same there with simulation.

With a volume item you get instant feedback with the fractals and they are represented as applied properly cutting through a volume, and is faster in general, I would also say by experience that they seem to render faster as well, but can be hard to compare since if you work with real life scale clouds like several km´s in size and try to do that with a gas solver and such, that would be very slow to deal with.

2. VDB´s are VDBs and not a volume item, which means they are based on voxel size where a volume item isn´t, they have the same settings for voxel quality which is different and called upon as Step size, for VDB it is crucial to have good enough voxel size for volume resolution as well as step size, and the volume item only have the step size to worry about.

3. Mesh shape? well as mentioned above, you have to simulate the noise which you do not have to do with volume item, with volume item it´s by far easier to make animated skyscapes with cloud movement as opposed to do that with a mesh item, though that is doable to some degree with a fluid simulation.

Now have to mention blender again, since it has a mesh to volume, and a volume displacer, so it´s easy to model a cloud mesh, turning it in to a fog type cloud volume, and then add fractals by displacing the actual volume result, it relies on voxel size there similar to openVDB, but it´s not properly implemented so you can´t export that out to Lightwave, but otherwise that method is faster than Lightwaves advection for the fractal noise, since it doesn´t need simulation.
Downside with blender here is of course it´s lack of fractals compared to how much we have in the fractal library for Lightwave.

In blender I also do it the way of creating a mesh, and simulating the fluids to only fill the mesh by turning off temp diffusion, and make sure it´s filled inside and with proper fluid offset from the surface, this will yield a volume from the mesh based on fluids but static, I could then add forces, wind, texture force, turbulence, vortex and simulating to get undulations on top of that which in fact is more realistic behavior than just a fractal texture, so it is only those forces creating the cloud undulations and not the actual vortices or temp forces from a smoke simulation that does the trick.

You can approach that technique in lightwave as well..almost, by creating a mesh, fill it with gas solver and turn off the other forces in gas solver, then add fractal forces, but that is where you hit a limit since there are no True wind forces that can affect gas solver fluids, fractals and vectors yes..but no proper wind or turbulence force.

A huge mistake to even release the vdb fluids without any such true force, they should have made such one implemented before releasing it, and to this day several iterations after the initial implementations, it´s still not there.
 
Last edited:

You can bake the old hypervoxels out the same though in 2019 as previously.
- cool.

Instancing is something completely different and has nothing to do with reducing a volume to image baked slices, and you can´t instance a volume item.
- yes, as indicated.

A huge mistake to even release the vdb fluids without any such true force
- i wouldn't say that, but certainly yes, it should be added.

another alternative could be to make clouds in Houdini, then import into LightWave.
or maybe, Blender, and import into LightWave.
haven't looked much into that tho'

and cloud sims are after all some of the most processor intensive task we could ask for.

exBang cpu.png

 

prometheus

REBORN
You can bake the old hypervoxels out the same though in 2019 as previously.
- cool.

Instancing is something completely different and has nothing to do with reducing a volume to image baked slices, and you can´t instance a volume item.
- yes, as indicated.

A huge mistake to even release the vdb fluids without any such true force
- i wouldn't say that, but certainly yes, it should be added.

another alternative could be to make clouds in Houdini, then import into LightWave.
or maybe, Blender, and import into LightWave.
haven't looked much into that tho'

and cloud sims are after all some of the most processor intensive task we could ask for.

View attachment 150742

My perspective is that I honestly believe they made a huge mistake on the vdb tools, and had they pushed it a bit further than leaving it so with too little samples and tutes as well as actual features, they would have had a chance to gain lot more upgrades as well as new customers, as it came to being, so many seem so confused on how to use it.
I think it was released so poorly that so many didn´t even know there was a fluid tool in there.
But we can beg to differ.(y)

Yes, you can save out vdbs in many forms from houdini, make a mesh, convert to cloudfx where the noise is applied, but you need to "cook" every vdb to a high resolution enough, and when doing that houdini can take quite a bit of time to accomplish that, the method is however a bit better when you can work with fractal noise staticly unlike lightwave requiring a time simulation.

With blender you can not work with the mesh to volume and volume displacement and send to Lightwave, it has no vdb export for that, unless the latest release have fixed that.
With blender you can do what I said about filling a mesh with fluids so you have a static volume, or and complement with a force to distort it..also requiring a time simulation.
And blenders volumetric material isn´t vdb so you can´t go that route either to send to lightwave.

For houdini you need to know on what nodes you should stand on and in which level, right click and just save out geometry, often in the end output within the cloudfx setup nodes, and just add the vdb extension by yourself when saving out the file, but there are several ways to do this.

It just struck me that the reason I can´t make a basic cup in 3D, is because all Ivé ever done is simulating clouds :D
 

prometheus

REBORN
thy wish is our command...




clouds are though to get right.
perhaps -overall- Houdini is the best ?

And now is where I smile like the Joker in Batman and say"why so serious" 🤡

And for real, nahh..that cup is taking too long to create and the handle isn´t fused as opposed to how most cup is made, it´s loose from the main cup.
I would probably skip drawing a spline as well since you can get none proper curve flow if not using it right.
So probably just start with a disc, extruding and taper/rescale bottom and end and thicken and bevel round bottom and top where I need, then use Artspherers bezier bridge tool which is excellent for the handles with nice tension and divisions to set.

As for best overall, perhaps houdini, depends on cloud types as we discussed earlier, lightwaves volume item isn´t that bad for larger skies apart from cpu render speed and the lack of multiple scattering, for lighting conditions where clouds show little such multiple scattering or none, and if you have a good CPU in your machine.

While converting a modeled mesh to volume is better handled in Houdini and being able to apply static noise which you can´t in lightwave (the volume noise on vdb surface is not the same here, that can be applied staticly, but only carves in to the surface of the volume.
To get good resolution though you need to lower voxel size quite a bit and that slows houdini down, it cooks the resolution so you actually can see the vdb volume in realtime, which you can´t do in lightwave.

But the fastest one to handle volume to mesh I would probably argue blender is, but it´s a close call between houdini and blender there.
But I can only speak for those, and I am also excluding terragen and vue for that comparison.
Maya can´t say, have tried it many years ago but not so much with clouds, same with max, cinema4D don´t know, Modo the last version I tried was a bit slow and similar to lightwave, but slower I think, and less fractals to use.
 

inakito

Member
Can you not mimic what "baking hypervoxels" is doing but with the new volumetric engine by hand? I meant, it should be a matter of setting up as many renders as slices you would need and play with your render camera clipping planes. Save each render as an image format that respects your transparency and apply back over the number of "cards" you may be needing, right? It may take longer and somehow be a bit more of a hassle, but if should do what you are after :)
 

prometheus

REBORN
Can you not mimic what "baking hypervoxels" is doing but with the new volumetric engine by hand? I meant, it should be a matter of setting up as many renders as slices you would need and play with your render camera clipping planes. Save each render as an image format that respects your transparency and apply back over the number of "cards" you may be needing, right? It may take longer and somehow be a bit more of a hassle, but if should do what you are after :)

Don´t see any point with it, why use the full realism volumetrics when you can´t use the glory of full volumetrics in it, considering it´s slices mapped to cards, it´s better to simply use old hv bake out, and use baked volumetrics which is built in to a null item.

Somewwat a bit more of a hassle and longer time to setup?..Not even close I would say, much more hassle and longer time to setup.

Even houdini seem to have skipped the image render to slices process I think with old volumetrics, the results are too unpredictable for what you see and what you end up with when baking.
 
Top Bottom