The Elusive Billowy Clouds

[moodflow]

As we all know, creating those perfect fluffy "billowy" clouds is a difficult endeavor.   However, I feel TG2 is capable of producing them - we just need to find the right way.

I wanted to post some of the progress I've made, as well as some theories/ideas as to how they might be created.  Hopefully, this will generate some ideas on your part, and we'll finally be able to have those elusive clouds one day.

Personally, I feel generating these clouds with a simple power fractal will not fully do it.  There are too many variables to content with.  So I've been trying to use multiple power fractals as "masks" to help shape the clouds for each variable for a final result (similar to the fractal stack methods I've posted about in the past). 

Below are some attached images of how I've been trying to pull it off.  In the below examples, I've used a large master fractal to generate the basic "macro" cloud shape, followed by another medium scale fractal to add some additional random detail.  Finally, there is a "micro" cloud shape, and this gives the final detail, but is "clamped" as to not have random splotches all over the place, which seem to ruin the effect.

Hopefully this is going in the right direction and will help you carve out some billowy clouds of your own.

I am still not satisfied with the below result...there is still much to do.

[Hetzen]

So are you layering up cloud generators at slightly higher altitudes?

I don't see how you can apply a distortion along a Y axis.

The example you've shown looks rather good.

[moodflow]

So are you layering up cloud generators at slightly higher altitudes?

I don't see how you can apply a distortion along a Y axis.

The example you've shown looks rather good.

Actually, these masks are to build up a 2D procedural image for a single cloud layer.  The axis are only X and Z.  So the example .gif files were of an overhead view (and very simple, just to show the approach I was using).  Of course the fractal would build up similarly in the Y direction (in the same way a heightfield would work). 

There was a discussion recently about how the cloud shader presents 2D inputs as 3D clouds.  You can see it here:  http://forums.planetside.co.uk/index.php?topic=4649.0  though no discussion was given by Matt or Oshyan yet at the time of this writing.

[Hetzen]

Funny, I was part of that thread.

I'd like to know a little more on how a cloud shader interprets a 2D image to form a 3D.

At the moment I'm working on some low orbit clouds seen from above, and am using an image map to control where they form, adjusting the coverage in the density fractal on two subsequently higher altitude duplicates. It's looking good, and isn't at the same camera angle as you are playing with, so is a different solution. But I do have a reference image which does have a storm stack, and apart from using several altitude layers on a focused set of X/Z masks, I'm not sure how else this is achievable.

I'm not convinced a cloud shader allows external distortion along the Y.

Please prove me wrong.

[RArcher]

Not entirely sure what you are looking for as I haven't followed along to much with this, but I just did a quick test about redirecting vertically and it seems to work fine for me.

First Test - Standard cloud no redirect:

Node network:



Sample:



Second Test - Redirect vertically:

Node network:



Sample:




As you can see it is clearly pulling up and stretching the cloud vertically.  Whether this is what you were looking for or not I am not sure.  I probably also should not have posted this here but in the other thread...  So back on topic.

Moodflow:  How are you organizing the node structure and masks?  Is it similar to the two examples I posted in the cloud sharing thread or a different technique altogether?  Would you mind posting a node network screenshot?

[Tangled-Universe]

The only distortion along the Y-axis is possible using a warp-shader.
Following Mohawk's tread I tried to use the redirect-shader but to no avail.
The warp-shader can curve clouds upwards to extreme extends if necessary.
However, like Moodflow said I'm also not sure how TG2 builds up the fractal.

Martin

[moodflow]

I haven't done any tests with the warp shader, but you guys have me intrigued!  I'm gonna go see how that works out with the current setup, but I have a feeling it will be warpy/stretchy (which is cool in its own right).

RArcher, actually, its just 3 fractals, one above the other that is a mask for the fractal below it.  The shape of the clouds is determined by the fractal's parameters.  I'm still trying to figure out whats going on and what does what within the fractal (and especially what settings create a billowy look). 

I'll post a shot of the nodes soon.

[Tangled-Universe]

I posted my message almost simultaneously with Ryan...

The reason your redirect works is because you connect the output to the warpshader.
If you disable the warpshader the effect will be lost.
Maybe the warpshader makes it possible to make use of the redirect shaders selective axis' inputs. However, a redirect alone won't work?

Martin

[old_blaggard]

Interesting discussion here.  I was always under the impression that the density fractal sent a 3D perlin noise to the cloud shader, and then some of the tweaks such as Base wispiness and Invert profile would modify how the cloud interpreted the noise.

[Hetzen]

RArcher, that's very useful. I'm flicking back and forth between the two images to see how the original has been effected through the warp shader. It's difficult to tell, but something is clearly happening. OK, there's a little bit of stretching going on, (cutting in parts if you like, as the clouds don't seem to be taller for the effect).

I would seriously like a description on how the Cloud Shader builds itself up from it's input. And how I can fuck with it.

Is it a matter of animated/progressive noise through Y?