Lightwave volumetric Items, clouds and sky moods, planet clouds, fog, godrays Part 2.


i think it looks very good.
just scale down if it gets too stacked, or add number of slices.

stack-50.png


 
...I think the problem is stacking 2d elements and with an emulation of getting depth, it gets step artifacts sort of, and lacks any full 3D volume on the sides, as we can see in your sample there is repetitve stacking....

The cyclone is a volumetric object, not a 2D element. Everything is fully 3D there. The texture map only controls the volumetric object. There is also no repetitve stacking, there is no stacking at all :).

You can see the exact setup here.

ciao
Thomas
 

prometheus

REBORN
t
The cyclone is a volumetric object, not a 2D element. Everything is fully 3D there. The texture map only controls the volumetric object. There is also no repetitve stacking, there is no stacking at all :).

You can see the exact setup here.

ciao
Thomas
What I ment by fully 3D is the same as 3Dtextures vs 2D textures, they can all be used in 3D space but 2D textures doesn´t allow for any other directional expansion than from one axis, unlike a 3D texture, you use a 2d image as density map, and it´s also it that is extended in to only one axis upwards, sure..no stacking, then it´s just the stretching/scaling of the depth height that is off, had you used purely 3D textures, that straight stretch wouldn´t have been a problem.

Now..you did use gardner clouds to soften it up a bit, which helps, but it still isn´t any volume outside of the directional stretch, think of it as displacement straight up, without any normal displacement direction.

Ergo, no fully 3D fractal textures for the height, thus a look that inherits a 2D image phase...extended to almost 3D.
If you map clouds from pure 3D fractal layers in Terragen for instance, you will not have this straight (flat look) stretching of the cloud height.

So ok, I was wrong about the setup, but the look is still not Fully 3D as in true cloud volumes have , and not a stretch from a 2d image.
The gardner clouds helped a bit, but it´s not taking any account to the full 3D volume in such a way that it can correct a 2D stretched volume based on 2D images.

It´s quite nice though, so don´t get me wrong, but still...if that cyclone would have been extended by ..or as a true 3D texture, the general volume propagating up in heigh would be rounder and with fractal variation.
though depending on, such low raise of the cloud height that is needed, may not require the full works, depends on how heigh you want the cloud layer peaks.

As for control of cloud height, try curves if you can..I think that is easier to control than gradients...once you get the hang of it.

As for coriolis and cyclone, being 2D textures only available in the scalar, color and bump layer, the outcome would most likely be the same as if you stretch an image bitmap initially.
The vortex cloud content however, is a deformation on a full 3D fractal Texture, it gives a much fuller 3D volume look than the method you adapted, though yours of course looks better in some aspects since it is originating from a true cyclone map, so the vortex cloud content would need a lot of tweaking still.

By the way, what scale was the volumetric cyclone here at?
approximated real life scale?
 
Last edited:
t

What I ment by fully 3D is the same as 3Dtextures vs 2D textures, they can all be used in 3D space but 2D textures doesn´t allow for any other directional expansion than from one axis, unlike a 3D texture, you use a 2d image as density map, and it´s also it that is extended in to only one axis upwards, sure..no stacking, then it´s just the stretching/scaling of the depth height that is off, had you used purely 3D textures, that straight stretch wouldn´t have been a problem.

Now..you did use gardner clouds to soften it up a bit, which helps, but it still isn´t any volume outside of the directional stretch, think of it as displacement straight up, without any normal displacement direction.

Ergo, no fully 3D fractal textures for the height, thus a look that inherits a 2D image phase...extended to almost 3D.
If you map clouds from pure 3D fractal layers in Terragen for instance, you will not have this straight (flat look) stretching of the cloud height.

So ok, I was wrong about the setup, but the look is still not Fully 3D as in true cloud volumes have , and not a stretch from a 2d image.
The gardner clouds helped a bit, but it´s not taking any account to the full 3D volume in such a way that it can correct a 2D stretched volume based on 2D images.

It´s quite nice though, so don´t get me wrong, but still...if that cyclone would have been extended by ..or as a true 3D texture, the general volume propagating up in heigh would be rounder and with fractal variation.
though depending on, such low raise of the cloud height that is needed, may not require the full works, depends on how heigh you want the cloud layer peaks.

As for control of cloud height, try curves if you can..I think that is easier to control than gradients...once you get the hang of it.

As for coriolis and cyclone, being 2D textures only available in the scalar, color and bump layer, the outcome would most likely be the same as if you stretch an image bitmap initially.
The vortex cloud content however, is a deformation on a full 3D fractal Texture, it gives a much fuller 3D volume look than the method you adapted, though yours of course looks better in some aspects since it is originating from a true cyclone map, so the vortex cloud content would need a lot of tweaking still.

By the way, what scale was the volumetric cyclone here at?
approximated real life scale?

In this case, I disagree with you. :)
The only difference to a real cloud is that the density of this cloud does not change in the Y axis. But I don't think the density in such a dense cloud varies much at all.
In this example, the density is greatest in the center of the cyclone and decreases towards the outside. This could actually be the case in reality. Any stretching problem that is due to the texture that controls the height. If you want, you can also make the lower edge of the cloud cover different heights (You can also create a sphere).

This example will show how to shape a full 3D volumetric object with a photo. Of course, this method is not the best for creating fringed cloud edges.

However, at this scale, no one can see it.

ciao
Thomas
 

prometheus

REBORN
In this case, I disagree with you. :)
The only difference to a real cloud is that the density of this cloud does not change in the Y axis. But I don't think the density in such a dense cloud varies much at all.
In this example, the density is greatest in the center of the cyclone and decreases towards the outside. This could actually be the case in reality. Any stretching problem that is due to the texture that controls the height. If you want, you can also make the lower edge of the cloud cover different heights (You can also create a sphere).

This example will show how to shape a full 3D volumetric object with a photo. Of course, this method is not the best for creating fringed cloud edges.

However, at this scale, no one can see it.

ciao
Thomas

You can´t disagree about volume 3D fractals, what I say is true about the volume. :) 2d bitmap stretches in an unatural way, clouds do not forms as volume like that, not even in a hurricane, the thing is that at that distance we do not notice that much.

what you can disagree about..that is if it is necessary to use full 3D fractals, since it looks pretty darn good in your sample, but still from my perspective..I would still like to see like more 3D volume.
But it depends also on the distance to the cyclone, hurricane etc, if you would go close up., I think your sample would fail a bit.

Don´t have much time to try myself, bedtime long time over do, the tricky part is to get the vortex rotation properly done.

Great to see someone else skilled at this though, it is needed to explore it all.(y)
 

forgot... Octane Cloud Render...

 

prometheus

REBORN
forgot... Octane Cloud Render...

Not much of new stuff there though, nor much different in terms of shading, scattering from a standard lw volumetrics as far as I can see, except for speed of the rendertimes.
That look in shading and scattering you can get with the new standard LW volumetrics, and even better noise fractals.

I don´t think ivé ever seen a great sample of octane multiscatter on a cloud sample.
If anyone has it and would like to show off Octane scattering, I still recommend we use the disney VDB cloud asset to compare.
 
Top Bottom