I can't seem to figure out what is wrong here... I want to put some white color to align nicely on a hedge seen in the following image. Sadly, I keep getting quite visible vertical lines going across it. I tried putting "tex coords" in. I tried putting "transform input shader" in (which works with the trees) and none of these work...
Please help me diagnose:
tex_probl.png
Have you tried checking 'Use undisplaced space (position in geometry)' ?
Same. :( Didn't even budge.
(I removed the top cover to make sure it's not interfering.)
tex_probl.png
Maybe adjust the Y scale in your PF to compensate? 0.1 should tell you if it works.
Nope. ;) I changed the scale, trying it with both "world space" and "undisplaced space". Even added a 4D noise. Something's very stubborn here...
I'm starting to conclude that the fractal itself might be the driving factor here. I made another test on a clean tgd but applied significantly smaller fractal scales (0.35 / 2.2 / 0.01) for the mask. PF seems to wrap up nicely around the hedge, but I imagined it should not work like this:
niby ok.png
Interesting that the scale needed to go higher in Y too get it to pull in. Looks like you've got it working. 8)
Quote from: Hetzen on December 06, 2019, 02:40:43 PMInteresting that the scale needed to go higher in Y too get it to pull in. 8)
Not quite! The second image has 1/1/1 scales in each dimension, only the fractal scale being smaller.
Actually, stretching the X scale helped as features have just stormed through vertical fractal parts, effectively, covering them up. But it is a nuisance all the same - this behavior has really caught me off-guard.
I even tried to add one more "transform input" after the surface shader... And have discovered a "voxel tree". :D vox.png
Are you sure you aren't seeing shadow/image pass effects?
When I look at the hedge in the foreground there doesn't seem to be any stretching, what I see is the object shape being repeated and shadowing/image pass effects applied darkening the area.
Even in your reworked PF mask version, I see the same vertical differences between object instances.
It'd be cool to see the hedge in a row on X facing the camera. That would rid the angle layering of instances to see them clearly as individual objects blending in flat against each other.
Yes, I think it's just the darker areas where the objects bend inwards, so to speak, that make the 'stripes'.
I dare to disagree. Because these stripes are mostly white. ::) And not quite far apart enough to account for the shading differences between neighboring objects.
White, because that's what you see best in the light, away from the depths between instances. If you'd render without sun or use a different pov for testing, you may see the white in the 'grooves' between instances better.
Acknowledged. But I still, respectfully, disagree since reducing fractal size and stretching it x-way seem to alleviate the problem.
And like I said - the "centers" of instances are too far away from each other to be considered as the root of the problem. If that was it, the vertical stretching would have been visible at regular intervals, not random. Randomness, please note, being one of the fractal's inherent properties.
So what if you apply the same fractal setup to a pop of planes, mimicking the hedge instances row. Then you can see how the fractal behaves on flat surfaces. As I can't believe (theoretically) this is happening. You did set to world space, I presume.
Regularity of the instances may be in spacing, but you also have rotation, giving each instance a 'unique' bulge. Resulting in irregular whites.
With respect also, of course ;)
Is warping enabled in the PF?
Quote from: cyphyr on December 07, 2019, 07:20:45 AMIs warping enabled in the PF?
Nope. It wasn't in either case.
I would still love to see a straight facing shot of the original scenario .
Quote from: N-drju on December 06, 2019, 05:39:10 PMQuote from: Hetzen on December 06, 2019, 02:40:43 PMInteresting that the scale needed to go higher in Y too get it to pull in. 8)
Not quite! The second image has 1/1/1 scales in each dimension, only the fractal scale being smaller.
Actually, stretching the X scale helped as features have just stormed through vertical fractal parts, effectively, covering them up. But it is a nuisance all the same - this behavior has really caught me off-guard.
I even tried to add one more "transform input" after the surface shader... And have discovered a "voxel tree". :D
vox.png
hee hee
Quote from: WAS on December 08, 2019, 01:06:02 PMI would still love to see a straight facing shot of the original scenario .
I'll provide it if you guys like. No problem. But I'll make the final render first, then we can experiment.