crevices?

Started by René, August 16, 2019, 04:08:36 AM

Previous topic - Next topic

René

Is there any way to detect crevices? Normally I do this with masks that I extract from the rock textures, but now I have a setup where this is hardly possible. I tried it with the merging of two distribution shaders, limit maximum and limit minimum, with not much success.
At the risk of talking complete nonsense: is it possible to use the information from GI surface details to do this? The only way I can think of is to use a render with exaggerated GI surface details via camera projection but then I inevitably get stretching.

Tangled-Universe

I have just sent you a PM with a potential solution or something which might help you move forward. Or none of both lol :P

Dune

Why not post here?

Anyway, I have been working with several layers of displacement and use displacement to scalar each layer and add/subtract, etc those. You could try that maybe...

Tangled-Universe

Fair question Ulco.
I chose to do that since it took me time to figure it out as well as that I do intend to share it at some point, but not now and not through a forum post. So, not for now ;)

Hannes

What about intersect underlying - favour depressions.. It's a bit tough to get the right values, but it may be a solution.

Dune

I can imagine that, Martin. A lot of shared stuff is quickly down the drain, so to speak.

René

Quote from: Dune on August 16, 2019, 06:36:01 AMAnyway, I have been working with several layers of displacement and use displacement to scalar each layer and add/subtract, etc those. You could try that maybe...
Thanks Dune, I am going to experiment with that.

René

Quote from: Hannes on August 16, 2019, 07:30:40 AMWhat about intersect underlying - favour depressions.. It's a bit tough to get the right values, but it may be a solution.
I've never had much success with intersect underlying, but I can try again. Does that work on vertical planes as well? Because then technically they are not depressions but horizontal notches.

Hannes

Quote from: René on August 16, 2019, 10:18:01 AM
Quote from: Hannes on August 16, 2019, 07:30:40 AMWhat about intersect underlying - favour depressions.. It's a bit tough to get the right values, but it may be a solution.
I've never had much success with intersect underlying, but I can try again. Does that work on vertical planes as well? Because then technically they are not depressions but horizontal notches.
Hmm, good point. I'm not sure atm...

WAS

It's weird to talk about things you aren't going to share on this public forum. Just PM them next time TU.

And I'm very surprised something so simple is a such a issue for this community, let alone TG. I've asked twice with no real help. I asked in another community and got immediate comprehensive explained help and a video tutorial. Amazing how much this community doesn't get it.

Difference Shaders guys. The only issue is mapping 2D Y maps to X/Z don't seem to work, even with world space.

WAS

#10
Here are two revised versions which are a little more straight forward to what they need to do.

Lateral is still broken.

This method is best for large crevices that are differences between your main smoother displacement, and your detail displacement.

Very curious why even when you have a defined map from Y it's not showing on the shape right, and just solid when on lateral surfaces.

René

Quote from: WASasquatch on August 16, 2019, 03:13:46 PMHere are two revised versions which are a little more straight forward to what they need to do.

Lateral is still broken.

This method is best for large crevices that are differences between your main smoother displacement, and your detail displacement.

Very curious why even when you have a defined map from Y it's not showing on the shape right, and just solid when on lateral surfaces.
Quote from: WASasquatch on August 16, 2019, 01:11:25 PMIt's weird to talk about things you aren't going to share on this public forum. Just PM them next time TU.

And I'm very surprised something so simple is a such a issue for this community, let alone TG. I've asked twice with no real help. I asked in another community and got immediate comprehensive explained help and a video tutorial. Amazing how much this community doesn't get it.

Difference Shaders guys. The only issue is mapping 2D Y maps to X/Z don't seem to work, even with world space.
Thanks for your input. The tgd's you posted are in fact a way of masking; that's the most logical way. If I started from scratch I would do it that way. However, I'm working on an older document with an already existing shader structure where I didn't take into account that I would like to edit the crevices later on. I didn't really expect GI surface details to play a role in this but decided to ask anyway; I just had to think about it because GI surface details also detects indentations in a certain way. What I had hoped for was that there would be a way to find cracks afterwards, something like the way the distribution shader calculates the slope.

WAS

#12
Quote from: René on August 17, 2019, 10:58:16 AM
Quote from: WASasquatch on August 16, 2019, 03:13:46 PMHere are two revised versions which are a little more straight forward to what they need to do.

Lateral is still broken.

This method is best for large crevices that are differences between your main smoother displacement, and your detail displacement.

Very curious why even when you have a defined map from Y it's not showing on the shape right, and just solid when on lateral surfaces.
Quote from: WASasquatch on August 16, 2019, 01:11:25 PMIt's weird to talk about things you aren't going to share on this public forum. Just PM them next time TU.

And I'm very surprised something so simple is a such a issue for this community, let alone TG. I've asked twice with no real help. I asked in another community and got immediate comprehensive explained help and a video tutorial. Amazing how much this community doesn't get it.

Difference Shaders guys. The only issue is mapping 2D Y maps to X/Z don't seem to work, even with world space.
Thanks for your input. The tgd's you posted are in fact a way of masking; that's the most logical way. If I started from scratch I would do it that way. However, I'm working on an older document with an already existing shader structure where I didn't take into account that I would like to edit the crevices later on. I didn't really expect GI surface details to play a role in this but decided to ask anyway; I just had to think about it because GI surface details also detects indentations in a certain way. What I had hoped for was that there would be a way to find cracks afterwards, something like the way the distribution shader calculates the slope.

That would be nice, and there must be a way about it if the GI can accomplish it. Unless it's more image manipulation using raws, but I doubt it.

I feel there may even be a way to get the broken share to work via some trickery. The definition is there, it's just not read on lateral.

Maybe @Matt can chime in on it.

WAS

I wonder if APs setup for lateral texturing would work here in a sense.

https://planetside.co.uk/forums/index.php/topic,24625.msg250245.html#msg250245

Applying the texture from crevices as the main input or some other fashion before displacement, so the colour is with the displacement. Didn't think of this until now. I'll have to test in a bit when the CPU is free.

Dune

It's best if you can use greyscales piled up and mixed for one final lateral displacement, which is the method I often use. Otherwise you have to convert displacements off your layers into colors again for masking purposes.