OTOY Forums

I know Octane is not really the program to make clouds, but I've always been interested in doing it... well check out what I did today!



Basically, I used the Turbulence node and hooked it up to the Scattering input of a (you guessed it) scattering medium. Normally it looks like nothing happened except for the surface, and you don't see anything, but here's the trick: You gotta adjust the ray_step parameter. Normally it's at 1000, but to see the volumetric properties of the generator, it has to be at a setting of around 0.6

Here's a picture of my setup (the model is just a really basic sphere):


The Medium's scale has to be pretty high too, like 5000, but it depends on the scale of your model of course.

I have a problem with this, it looks as though the surface is still acting weird, like it's not part of the volume, or is acting like the ray_step is still 1000. Also if the ray_step starts to go lower, the volume starts dissapearing instead of getting deeper or more volumetric. Playing around with that setting is pretty interesting. lower values will give you a closer volume, and as it increases, the volume goes into the model, but as you get higher than around 1, it starts to fade out. And one more: you can't see volumes on the "other side" of the model. So I guess normals with more than 90° difference than the camera (back facing normals?) don't have their texture generated INSIDE the volume either... strange...

Do some experiments and let's discuss them!

Here's another test


On this one, I used a custom model of a little clay cube I made a while ago... nothing special, just needed something with a surface beneath it and a different shape.

Anyway I had to turn on the caustic blur to get the light to come through and see how the shadows were rendered. I think the shadows are also only cast by what's drawn.

This shape also lets you see better the weird surface vs. volume SSS which I think needs to be addressed. Why isn't the volume being lit up like the surface? I wonder if it's because the density is far less inside the volume than it is at the surface? or maybe it's that it's being calculated incorrectly?

looks really interesting.

Impressive !

Very nice! Are there any other settings to pay attention to? I can't seem to get the fine detail within the shape like you have...(used the same numbers as you've shown in your screenshot)

Great work!

awesome.
great idea to use such a map for scattering medium... never thought of that.

Ah! magicU, the octaves parameter in the turbulence node seems to be hidden/collapsed. I put that at 16, but anything above 10 will do... 16 may be a little too rough for some things. Omega changes the contrast heavily too also how rough it is.

Also, why aren't these looking more like this?


Is it something about how the surface is calculated based on the volume? There's probably a good reason, but I don't think the surface should be treated differently than the volume, it's part of it!

Very impressive. is there a scene you can share with use to see the settings you have used to accomplish this task.
Job well done thanks for sharing your findings.