Procedural dunes - problem with the render

Started by wiwine, January 09, 2010, 01:23:53 PM

Previous topic - Next topic

wiwine

Hello everybody...
I'm currently working on a procedural dunes generation, and I'm not far to get what I wanted,
using the functions.
But I've got a big problem with the renders : multiple thin lines appears everywhere and I
don't know how to avoid them. See the renders below.
When the camera position is near the ground, it seems to be worse.
Have you an idea to solve this problem ?

inkydigit

not sure, but is this caused by some smallest scale displacements being too small?

FrankB

very likely, if you ask me. Try a bigger smallest scale (1 cm for testing) and see if they go away.

wiwine

I deleted the displacement and the functions that made the small sandwaves.
The main displacement (that makes the dunes) as a value = 7.
There's no improvement... I would even say it's worse !

See the attached files.

FrankB

do your functions potentially produce divide by 0, or coming close to dividing by zero?

wiwine

Not much divisions in my functions. When I use a "divide scalar" it is with a constant value.
There's no "division by 0".
I sent my .tgd file in the previous post, if you want to make a test by yourself.

TheBlackHole

They just issued a tornado warning and said to stay away from windows. Does that mean I can't use my computer?

cyphyr

#7
I added a "compute terrain" node (default), between the last multiply scalar and the surface layer,  seems to be working ... :) I'm on a slow laptop so it'll be an hour or so till the whole frame is complete.
:)
Richard
ps damn that's some fine functions you got there ;D
pps: I'm rendering your "quick render" settings
www.richardfraservfx.com
https://www.facebook.com/RichardFraserVFX/
/|\

Ryzen 9 3900X @3.79Ghz, 64Gb (TG4 benchmark 6:20)
Ryzen 9 5950X @3.4Ghz, 16Gb (TG4 benchmark 4:28)

wiwine

I just tried your compute terrain (I've got a quad-core, it's fast). Does not work, unfortunatly...

cyphyr

#9
Interesting, I've only got 1 error so far similar to yours and I'm halfway through the render.
Attached file in case I didn't explain something right.
:)
Rrichard
ps when I first did this I copied and pasted the "compute terrain" from your set up and used that between between the last multiply scalar and the surface layer. Unfortunately I forgot to tick "enable" lol
www.richardfraservfx.com
https://www.facebook.com/RichardFraserVFX/
/|\

Ryzen 9 3900X @3.79Ghz, 64Gb (TG4 benchmark 6:20)
Ryzen 9 5950X @3.4Ghz, 16Gb (TG4 benchmark 4:28)

wiwine

I've still got the thin lines with your file !
Could you make another test : a render with my first file, without modification ?

cyphyr

Sure :)
Rendering again with your original unmodified file.
I just uploaded the modified render, there is one line, but that's all.

I've wondered for a while if different processors might produce slightly different results at extremes.
Richard
www.richardfraservfx.com
https://www.facebook.com/RichardFraserVFX/
/|\

Ryzen 9 3900X @3.79Ghz, 64Gb (TG4 benchmark 6:20)
Ryzen 9 5950X @3.4Ghz, 16Gb (TG4 benchmark 4:28)

Matt

I think it's a multi-threading bug. With max threads at 4 threads I got the spikes, with max threads at 0 I got no spikes. It's probably one of the function nodes or shaders, but I haven't figured out which one yet. If you discover which node is causing the problem, please let me know.
Just because milk is white doesn't mean that clouds are made of milk.

cyphyr

For the sake of completeness here's the render done with two threads at its original settings. Still one error.
Richard
www.richardfraservfx.com
https://www.facebook.com/RichardFraserVFX/
/|\

Ryzen 9 3900X @3.79Ghz, 64Gb (TG4 benchmark 6:20)
Ryzen 9 5950X @3.4Ghz, 16Gb (TG4 benchmark 4:28)

wiwine

I was just rendering the same view with min/max threads = 1 and there's no spike !
I'm going to reduce the whole process, function after function, until the spikes disappear. Then I tell you.