PDA

View Full Version : Twist Anisotropic? Is it possible?



Schwyhart
05-12-2017, 06:17 AM
I'm adding a texture to a door handle cover. Looking at the real one, it looks like the lines twist in a spiral from the inside to outside of the cylinder. I was thinking there must be a twist node, but I can't find one.

136777

Schwyhart
05-12-2017, 08:03 AM
Here's my model

136779

MonroePoteet
05-12-2017, 09:31 AM
If no one posts a Twist-able anisotropic node for the surface itself, you might be able to use the Twist deformation to twist the doorknob's baseplate geometry. Sample scene attached.

136781 136782 136783

The scene uses a simple flattened ball with the hub triangles merged into 4-point triangles (avoids pinch artifacts at the hub), the Brushed Chrome preset applied (F8 in Surface Editor, Metal preset shelf), size of the Bump map adjusted and Reflection / Specularity / Glossiness modified. You'd use your anisotropic surface instead, since it's the geometry that's twisted by the Twist deformer.

The Twist deformation is applied in the Object's Property=>Deform tab, the Add Displacement pulldown. The TwistBase null and TwistHandle null control the amount of twist using the Bank angle and Z position of TwistHandle since the deformer's properties indicate to twist along the Z axis.

Good luck!

mTp

prometheus
05-13-2017, 05:12 AM
free dpont rman collection textures and perhaps use the http://dpont.pagesperso-orange.fr/plugins/textures/Spiral.html#2


together with the additional nodes and the deform node...
http://dpont.pagesperso-orange.fr/plugins/nodes/nodes/Deformers.html#Deformer

rustythe1
05-13-2017, 08:25 AM
just create a planar UV map project at the circle, then in the UV window grab the centre and rotate them, then when you apply a lined image it will appear to twist

Schwyhart
06-03-2017, 05:06 PM
So, I figured out it wasn't the anisotrphy, it was the scratches that are twisted. The anisotrphy isn't twisted.

Iaian7
06-05-2017, 01:53 PM
Scratches, in a general sense, are what cause the anisotropic effect. The two are inextricably linked. You'll note that the direction of anisotropic reflections run perpendicular to the direction of the scratchs. For example, if a scratch runs horizontally, it will reflect light vertically (if you think about the micro surface details created by a scratch, this makes sense; a scratch creates surfaces that face both up and down, but not side to side, as that's the direction of the scratch's trench). When a bunch of scratches all run horizontally, we see the typical stretched anisotropic reflection running vertically. Look closely at a brushed metal surface, and you'll see this effect in practice.

Of course, with enough rendering time for antialiasing, you can brute-force this effect by adding micro scratches to your surface (normal mapping preferred over bump). The anisotropic nodes simply simulate this in a much faster way. :) (if I remember correctly, ILM tried the brute force method for the first Iron Man movie and gave up due to rendering constraints)

To create a twisted anisotropic reflection, try inputing a gradient into one of the rotational channels of the anisotropic node. For example, if you have anisotropy applied using cylindrical projection on the Z axis, rotate the banking channel based on z distance of the surface, or a radial gradient if the surface is flat. This will twist the anisotropic effect. :)

Iaian7
06-05-2017, 01:57 PM
This is an extremely basic example, and I haven't even debugged it for accuracy or efficiency (just using a turbulence node for the scratches!), but it appears to be working. Maybe this helps explain?

136989

EDIT: And that's why debugging is important...the sample I posted was total nonsense. Twisting on the z axis when using cylindrical projection does nothing. My bad. :D

Iaian7
06-05-2017, 02:14 PM
There we go, that's more like it...though endless room for improvement, it's an example of how it could work. I also added a standard highlight to demonstrate a second smooth coat effect (and to better show where the scene lights would normally reflect off the surface). For this kind of surface I would typically use DPont's Bump to Normal node so I could apply surface perturbation to only specific nodes (disrupting the anisotropic highlight and reflection nodes only, for example), leaving the actual surface bump and normal inputs unaffected. That said, the attached object and scene file should only use standard Lightwave nodes. :)

136991