Need help for lateral displacement

[Alfamike]

Hello guys,

Keep having to bother you with lateral displacement over here.

Please have a look at the screen shot below.

Without the Compute Terrain the Displacement offset along normal does not work, not a problem. This displacement generously changes the terrain normals (creating 90° slopes) so I figure a (re) Compute Normal 01 is needed for further displacement. This completely screws up the terrain, though.

With and without this Compute Normal 01 my lateral displacement for the 90° slopes never works.

I'd really like to get this displacement on the 90° slopes to work but cannot get my head around it.

Help/hints/ideas will be much appreciated.

Have a nice holiday and good 2008!

Thanks.

AM

[rcallicotte]

I'm not sure how much help specifically I can give to you, but I'm wondering if you did a search for the Compute Terrain / Compute Normal that Matt wrote about six months ago.  He explained about the Final Normal and Final Terrain and how these work in a node tree.

[Cyber-Angel]

I have been thinking since Vertical Only, Lateral Only and  Lateral Normalized displacement types all require a computed normal node, could not TG2 create one automatically if one of these types is selected as the fractal breakup node is created automatically when you create a surface layer node? This would save time I think then having to create the computed normal node by hand.

Regards to you.

Cyber-Angel     

[Alfamike]

Hey!

Thanks.

Just searched the site for all "compute", "fine", "normal". Just this thread showing up though. Nothing much showing up searching for "terrain", even...??

I guess a lot of displacements make only minor changes to the terrain, not needing a recompute of the normals slowing down the rendering. Maybe a tickbox on the displacement tab to either recompute terrain/normals or not?

Warm winds!
AM.

PS Just found something written by Matt 9 July 2007, called "child layer inconsistancy", but it did not help very much. 

[j meyer]

Maybe you could find it by clicking on Matt's username
and then on show last posts.

[Alfamike]

Maybe you could find it by clicking on Matt's username
and then on show last posts.

Yep, got it. Restricted my search to a little too narrow a time frame around half a year ago. Search for all Matt's posts and it came up.

This is it for those interested : http://forums.planetside.co.uk/index.php?topic=1249.15

Thanks, got something to read now.

AM



Back later.

[Alfamike]

Back.......

Read Matt's article, is a good read.

But it didn't help me.

Anyone be so kind to look at attached tgd and see if they can get some lateral displacement on the vertical "walls"?

Much appreciated.

AM.

[Oshyan]

I think there are a number of problems here. First, you're using Surface Layers in the Terrain group which, although not "illegal", doesn't have much purpose in the way that you appear to be using them. Just use Power Fractals directly to provide the displacement. The first Compute Normal also should not be necessary, but you've likely kept it because it has produced the terrain effect you wanted with the vertical faces. As far as I can tell this is an anomaly, not an appropriate method to create that kind of effect, and I suspect this is why later use of lateral displacement isn't working. I would suggest looking at the Procedural Canyons thread for some inspiration and reference on how to properly create these kinds of terrain shapes: http://forums.planetside.co.uk/index.php?topic=1266.0

So essentially you want to find a more appropriate method for creating your initial canyon/mesa-type displacement (with vertical walls), and then try normal displacement methods on it with a single Compute Normal. Post again if you still have problems after that.

- Oshyan

[Alfamike]

Hello Oshyan,

Thanks for your time and for shedding some light on this.

I looked at the function to generate canyons you mentioned and tried to plug that output -after compute normal- into a power fractal to generate some lateral displacement.
That works, but it looks like the displacement amplitude has to be many times bigger than what you'd expect. In other words, an amplitude of -say- 5m did not look like max 5m features on the near vertical walls. Not a problem 'cause you can just bump up that value.

I was originally working on creating some good snow and found Lucio's "embossed snow" clip file, amongst others. Struggled with that lots but could not get the lateral displacement on the edges of the "snow" or "glacier" to show.

This post is meant to get over that. I just removed all other shaders and only left the one from the top of Lucio's called "Snow_vertical emboss". He uses a surface layer for this under the shaders node group.
According Lucio one needs to play a lot with the fuzziness of the slope restrictions and the displacement offset in that vertical emboss shader to get a good result, but I don't get anything, neither lateral nor normal.

If you have some more time I would really appreciate another comment on this, and thanks in advance.

I'll attach Lucio's clip file.

Have a good 2008.

AM.

[Oshyan]

I'm afraid I'm not really sure what you're trying to achieve at this point, or what problem you're having, so I can't offer much more specific advice. But I know several people have had success with Lucio's techniques, so you may just need to experiment some more, start simpler (from the default scene).

- Oshyan

[Alfamike]

Thanks Oshyan.

Hello again everybody,

My problem is really quite simple I think.

1. I select a bit of terrain using slope constraints.
2. Then I displace that terrain along the normal using an offset only, no multiplier.
3. I can't get lateral displacement on the "walls", the sides of the displacement.

I have tried many things and am now thinking about re-installing TG2 since I have had too many crashes trying to get this to work.

If anybody can throw together a small TGD in which the above works, I would be grateful if they could let me know.

Thanks.

AM.