Mantra displacement issue

[Norad]

Hello guys

I'm encountering a following problem:

I create a turbulent noise, that I plug into Displace Along Normals to create noisy surface, however displacement happens in relation to camera space for some reason.

As you can see I'm displacing only in selected places - this is defined by vertex colour map (extracted from Vornoi Fracture depth)

The SOP in short - Vornoi Fracture, faceting, Explode View to see different parts clearly, Delete to concentrate on one piece, Point to convert __depth attribute into $CG, VOP SOP to remap the $CG slightly, Rest for shader's Rest Position (enabled on 'if' for both normal and quadric transform )

And finally the shader network

I'm taking Rest Position as Position attribute for Turbulent Noise, to prevent sliding. I accumulate two noises and revert them, and then add to the $P. A mix is defining where to use original $P and where to use noisy, based on bias fed through $CG from the vertex colour (surfaceColour node). Then result is fed to Displace Along Normals and through to the Displacement Output.

As you can see the setup is pretty basic, however I'm having the issue with rendering the displacement, where it happens along the camera coordinates, not along the surface normals. I tried feeding $N position to Displace Along Normals, tried plugging in Transform nodes everywhere, only result being blowing the displacement out of the camera viewing capacity.

Can anyone help me with this issue? Or have any thoughts on the problem? I'd very appreciate any help, its for a deadline on 8th February, so please share any knowledge you have in this area

shatteredMaterial04.hipnc

Edited January 23, 2013 by Norad

[Akabane]

Tried using "Object Space" inside the Displace Along Normal node?

[Norad]

Yes, no difference

[sam.h]

If you want to displace along the normal, your noise should be float and then you can plug that into the displace along normal node "amount" input. P and N in current space can go straight into the displace along normal node.

If you are just using the disp along normal node to recalculate the normal (never tried that before ... ) you should transform your P global into object space before the add, then back into current before it goes into the disp along normal node .... that should work.

[Norad]

sam.h thank you so much for that reply! It worked, both suggestions were correct. I see now how and why it was wrong before, you helped me big time here!

Take care mate