Cobble Pattern

[René]

Thanks for sharing.
I played a bit with this file and I must say that it gives good results for making rocks when more variation is added. Of course there is room for improvement, as always.

[Dune]

That is pretty cool. Must be played carefully by the looks of some. I like the first test best, actually.

[René]

Exploding geometry is always lurking. Sooner or later I will have to add a compute normal but I will only do that at the very last if there is no other way.

[Dune]

Yes, that's right, but the sooner actually the better, or it will have to compute a lot.

[Hetzen]

I've always loved your rock work Rene. Thanks for trying this out.

Yeah you're right Ulco, you have to be careful with any displacement, those inside-out normals can be a real pain. I was playing around with a 'minecraft' type of function, which sort of works, but as you can see from the two pics below have issues.

This uses the modulo function to break up an input terrain into cubes. It's more of a proof of concept. It's a bit clunky as you can see by the texture 'flags' you can see on the tops.
[attachimg=1]

Here's the same terrain and function, but with a random expansion per zone. Different light position too. Again more issues to work out. Neither needs a Compute at this stage.
[attachimg=2]

[Kadri]

Interesting. Is this only with one surface (last image especially) ? It is probably, but i just want to be sure

[bobbystahr]

Even more brilliant...keep. on. tweaking.

[Dune]

Interesting indeed, too bad about the stretched parts. A PF over it and a lot will be far less visible, I guess. This would (could) work well on volumetrics, if/where only the outside visible areas were rendered.

[Hannes]

Wow, this is cool, Jon! I hope you can tweak this further.
And René, your rock stuff looks brillant as well.

[René]

I've always loved your rock work Rene. Thanks for trying this out.

Yeah you're right Ulco, you have to be careful with any displacement, those inside-out normals can be a real pain. I was playing around with a 'minecraft' type of function, which sort of works, but as you can see from the two pics below have issues.

This uses the modulo function to break up an input terrain into cubes. It's more of a proof of concept. It's a bit clunky as you can see by the texture 'flags' you can see on the tops.
[attachimg=1]

Here's the same terrain and function, but with a random expansion per zone. Different light position too. Again more issues to work out. Neither needs a Compute at this stage.
[attachimg=2]

The second is very interesting. It almost looks like there are overhangs. Would that be possible?

[René]

Another iteration. It's still looks a bit wobbly.
I'm starting to incline more and more towards the idea that in order to make good rocks/terrains you have to know the processes that have led to them looking as they do. Not to imitate the natural processes in any way, but to be able to analyze what you see exactly when looking at reference photos. Often there is so much going on that you can't see the forest from the trees.

[Hannes]

That's awesome, René!!!

[archonforest]

Two thumbs up! Wish to know how to create something like this.

[Dune]

Very cool!

[mhaze]

Wow! René, that last one is amazing!