Marble Hill

[Kadri]

Yes would be nice to hear what is going on really.

[dandelO]

I noticed a similar lattice effect on a model using GISD on only one flat Default Shader.

Also, these nasty errors around edges where GISD is visible against an image map background. Same thing?

[dandelO]

Low GI sample quality/High cache detail level. The occlusion weight is very high, which makes the effect very noticable on the teapot.

[dandelO]

Zoomed into the lattices, seems to be polygonal.

[dandelO]

What's the terrain Patch Size in the ground shots, Kadri? As small as the lattices?

[Kadri]

What's the terrain Patch Size in the ground shots, Kadri? As small as the lattices?

Martin ( hello ) i don't remember but i tried it with different values. 1 , 0.1 and i think 0.01 even.
Was mostly the same if i remember correctly. I will have a look tonight ones more.

[Kadri]

There was a problem at rendering polygons if you remember.
Maybe that is the same issue you have there and different then the pattern on the landscape?

Then the black outlines of tree objects behind smog comes to my mind.

[Kadri]

http://www.planetside.co.uk/forums/index.php/topic,17186.msg166818.html#msg166818

The same?

[dandelO]

It's definitely to do with GISD, in my images, at least. Just thought that the lattice patterns could be the same thing I found. The outline of the window renders was still visible in some other situations but was really badly noticeable against a darker area of an image map bg, as you can see. The light, sky area isn't as bad.
Sorry if this isn't related, just thought that it was very similar to the problems apparent in this thread.

[dandelO]

http://www.planetside.co.uk/forums/index.php/topic,17186.msg166818.html#msg166818

The same?

Yup.

[Kadri]

Martin here is the TGD file of the same image. You can try it for yourself too.
I tried it before with GISD or GI fully off etc but there was still the pattern.
I think it is a real displacement . 
The problem with the object smoothing was a render problem if i remember right.

[Kadri]

Even with a very small compute terrain number like 0.001 and smoothing there is a pattern.
Smoothing only makes the pattern very smaller.
It resembles kinda an image map displacement seen from very near and without smoothing

[Kadri]

For example change only the "Clumping of variation" to 0.255 in the "Fractal terrain 01" node and see what happens.
This is the same scene only that value different

[Kadri]

I would basically change the parameters or the noise flavor or Variation method etc. if i got those patterns.
Maybe with that kind of combination this is normal.
Maybe not. I don't know

[Kadri]

I have done more test. Beware that all the things below are in the "I think" category.

*  Different Noise flavor or different Variation method still does have -different ?- patterns.

*  I thought when we are closer to the ground we will get smooth-kinda surfaces.
    But it looks like needles very close with displacement and no Clumping of variation.

*  Not sure if there is a pattern (but it looks like it has) without Clumping of variation.
    But with Clumping you get a pattern with any "Variation method" or "Noise flavor" .

*  Using "Clumping of variation" and "Fractal warp shader" together does get a much apparent and different pattern.

So the joke on the previous page about the fundamentals of terrain doesn't look like a joke so much to me now.
Still curious if this was always so?

The conclusion for me for now is "Don't get too close to the ground and use different scales".