Classic Erosion - Lateral Displacement

Started by WAS, October 24, 2018, 04:46:18 AM

Previous topic - Next topic

WAS

Is this taken into account? I believe I asked before but didn't get a answer on the main topic.

I'm wondering if the plugin would be useful to erode on a tiny scale and mask it in with original disp to get a smoother, somewhat worn look to otherwise jagged cliffs. Reason I'm asking is testing this is going to take all night to compute the terrain and process it so if I here anything soon soon that it doens't work I can cancel

digitalguru

Don't know about lateral displacement, but small scale erosion works great.

I had a heightfield generated in World Machine and used the plugin to add some small scale rain channels that didn't peturb the terrain as a whole and it worked really well.

WAS

I'll have to play around. I did a erosion overnight at a scale of 10m and a erosion strange of 0.1 and duration of 0.05 and it still just flattened the entire shear cliff to a little rounded off mound on the ground. :\

Tried some lower resolution tests, even lowering erosions to 0.05 and duration to 0.01 and even with a split section calculation it still gives a tiny mound compared to a 800m cliff.

Oshyan

No, Classic Erosion is heightfield-based and therefore can't deal with anything that overhangs. Strictly planar.

- Oshyan

WAS

#4
Quote from: Oshyan on October 24, 2018, 04:08:14 PM
No, Classic Erosion is heightfield-based and therefore can't deal with anything that overhangs. Strictly planar.

- Oshyan

I'm wondering if there is a way to mix that planar with the original disp and at least "fake" it's effect by wrapping the terrain. Hmm..

Thanks for the clarification. I should really remember that affiliation - planar.

Edit: It seems this is sort of possible by redirecting the erosion shader by the original terrain. You do this via the Displacement to Vector and subsequent Red, Green , and Blue to Scalar functions piped into their own displacement functions and piped into the appropriate X, Y, and Z inputs of the redirect shader.

However it's so slow it's not really usable.

Dune

#5
You could try using the outputs to drive vector displacements.... but with a loose hanging erosion shader. I did that a couple of times but only as scalars, not vectors. But I guess you mean the same.

WAS

#6
Yeah I think.

I was able to tone down disp by half at 0.5 instead of 1 and it didn't take nearly as long to render, but alas, was also not a cliff, but did have some interesting shapes. Definitely something to experiment with. I think I may formulate a test file to throw up here on the forums.

I'd be interested in what you're talking about with the vector displacement, not sure I understand.

Dune

I mean the output (say stream mask) could be used to (carefully) displace vertical areas by vector displacement (or redirect of course), but you'd have to use the 'right' slope angle. In fact I've been experimenting (in the far past) with combinations of vector displacements, masked by slope angles. So South slopes get VDisp of mainly -Z, north slopes of +Z, west slopes of -X, etc.

WAS

Quote from: Dune on October 25, 2018, 05:33:48 AM
I mean the output (say stream mask) could be used to (carefully) displace vertical areas by vector displacement (or redirect of course), but you'd have to use the 'right' slope angle. In fact I've been experimenting (in the far past) with combinations of vector displacements, masked by slope angles. So South slopes get VDisp of mainly -Z, north slopes of +Z, west slopes of -X, etc.

Oh I see. That does seem it would help with the exact of the original displacement more than a redirect on XYX of just the original terrain vector broken into it's channels.