Applying Displacement

[j meyer]

Ok,you guys were absolute right.It was definitely my fault.
It was a missing Transform shader that caused the difference
in the other file.As soon as that was added it was indeed the
same result as with the CT with a patch size of 1.
Still interesting enough to me.

My apologies to anyone thinking I've wasted their time.

In the meantime I read your post with TG open Jon and now
I can say that what you suggested is pretty similar to what
I like to do.Only that I connect the PF directly to the displ
shader and control the height with that.Of course I would have
to use a Displacement to scalar,if I'm not mistaken, to use it
with a Build vector.
The Build vector thing is indeed a cool variation I didn't think
of myself before.
Your suggestion is a good example for the differences between
color based displacement and the built in displ of a PF.
This difference is what I don't understand and one of the reasons
I decided to work with color based displ rather.
Again thanks for your input,I do appreciate.


Well,errh,I had another idea in regard to the use of a Compute normal,
which I wanted to test.And so I did.
I wanted to find out,if it would be possible to apply that method to just
a part of a terrain.So that there would be the main portion of the terrain
with the CT (default patch size) only and a small area with the CN to get
a better resolution or whatever you call that.
I made a separate node chain for the SSS-block,the CN and the surface
layer for the displ and connected it to another surface layers child input.
The main node chain has an identical SSS-block that is connected as
usual.
And it seems to work, only the SSS-block of the child chain is getting
the "higher resolution" displacement.

[attach=1]

Then I wanted to apply what I just learned and replaced the CN with
a CT with the patch size set to 1. I did nothing else,at least nothing
I'm aware of.
This is the result.

[attach=2]

Thereafter I tried it with World space disabled,same result.
Any ideas as to what may cause the difference here?


Edit: forgot to say the SSS-block with the CN/CT is on the right hand side.

[Tangled-Universe]

It would help a lot if you can accompany both images with their respective node-network screenshots

[j meyer]

When you're right you're right.
One should be sufficent in this case,though.
Here it is.
[attach=1]

[Dune]

I used that method of a localised small patch CT/CN for making walls. The only problem is that if you have a lot of displacement in the line before and add that as child, it will 'double' the whole terrain displacement, so you have to be careful.
The difference between CT and CN is not visible in your examples, which is quite logical, you only effectively use the normal calculation. Only if you'd apply color, you would need the XYZ information, hence an added XYZ shader, or a replacing small patch CT.

[j meyer]

Thanks Ulco,I get the first paragraph of your post,but for the second
I'm not sure that I understand what you're saying,'cause to me the
difference is clearly visible in the images.
And the xyz will be needed for terrain/displ matching colors,did I get
that right?

[Dune]

You mean the difference between upper and lower image (left-left/right-right)? I don't see them next to eachother, but at first glance there's not a huge difference. And yes, XYZ is for texture location.
If you want horizontal strata colors for instance a default compute terrain will often warp the strata, so if you add a XYZ shader, the strata will be absolutely straight (if stretched XZ).

[j meyer]

Ah ok then I think I know what you're refering to with that xyz.

Yes the difference is not that huge,it's just on some parts.
They don't get displaced as much as with the CN.

[j meyer]

....
        There was only one Ct in all my setups,so we must have misunderstood
        each other somewhere.
....

Hey j meyer excuse me please,but this is hilarious man.
The most shocking display of mathematical insufficiency
ever.Obviously you can't even count to 2.
C'mon now,really...of course there are 2 CTs in some of
the files.

[j meyer]

Thanks for pointing that out J!
Indeed a mathematical record performance.
Will be hard to top that one!


Martin,I apologize for being that dense,unbelieveable.
Of course the first CT wasn't there to compute the flat terrain.
While experimenting I put the noise over the whole of the
default scene,even without the default SSS.As you can see
in the version posted to the Rectangular noise thread btw..
One setup to test more than one possibility.
Hope that makes it somewhat clearer.

[j meyer]

Didn't think of posting an animated gif either.
Would have been better,too.

[attach=1]

[Dune]

It's like a pumping heart... fascinating.