Page 1 of 1

Real CarGlass? What parametrs?

Posted: Thu Oct 05, 2006 2:48 am
by cooler_inc
Hello.
I need a real carglass (windshield, headlights) but i can't setup the materials. I'm not satisfied with the result.
I use Blender and Indigo 0.5
I tryed IOR 1.33,1.5 but there is not refraction at the windshield in real world.
RayTransp button enabled, Fresnel 4.0, Alpha 0.3

Does anybody help me with materials?

Posted: Thu Oct 05, 2006 2:57 am
by playmesumch00ns
Have you modelled the windshield as it is in the real world? i.e. with thickness? I believe it's the double refractio that gives you a "normal" image through the glass.

Posted: Thu Oct 05, 2006 3:06 am
by eman7613
i used frensal 1.2 for a glass table i made in cinema for a table and it worked well, try that maybe?

Posted: Thu Oct 05, 2006 8:53 am
by leope
Blender alpha and fresnel parameters are not used by the exporter.

Code: Select all

rayMirr ==> <normal_reflectance>
ior     ==> <ior>
filter  ==> <cauchy_b_coeff>
r,g,b   ==> <rgb_absorptivity>

Posted: Thu Oct 05, 2006 8:13 pm
by cooler_inc
2playmesumch00ns: the windshield is modelled like in the real world like a flattened out cube.
2eman7613: thnx, I'll try.
2leope: if alpha and fresnel are not used by the exporter it means that we have to write in the xml file manualy?

Posted: Fri Oct 06, 2006 8:14 pm
by cooler_inc
I used various parametrs of IOR, Spec, Fresnel and can't to approach to desirable result.
Image
Done 1776031500.000000 mutations (5781.352539 mutations per pixel)
Time elapsed: 14h, 33m, 22s
Mutations per second: 33892.072011
Image
Any suggestions?

Posted: Fri Oct 06, 2006 9:49 pm
by Elroc
Is that the IOR you used in that image? 1.01?
I assume you only have a windsheild modelled on the car? Your right it does look weird.

At the risk of being completely wrong and way of base i'm going to take a shot in the dark :) Is your windshield modelled as a flat plane, or does it have some thickness?

What I'm thinking is that it is only a plane, and therefore the light only bends once - when it enters the windshield, therefore the funny effect. If give it a thickness, light should bend twice, once entering, and once leaving the glass, giving you a better image.

Image

Or I could be completely wrong of course :)

Edit: Note to self, read rest of thread before posting - see you have modelled the windshield with thickness. In which case I'm stumped, sorry.

Posted: Fri Oct 06, 2006 9:57 pm
by 8bstudio
Elroc wrote:At the risk of being completely wrong and way of base i'm going to take a shot in the dark :) Is your windshield modelled as a flat plane, or does it have some thickness?
Right Elroc!!

I had same problem: the glass mustn't be a plane but must be a double plane with the normals out or a closed object with normals out!

And maybe IOR value around 1.5 or higher!

Posted: Fri Oct 06, 2006 10:26 pm
by cooler_inc
My windshield has a thickness (as shown right at Erloc's image)
I'll post as soon as I make an experiment with material settingns.

Posted: Sun Oct 08, 2006 8:32 pm
by cooler_inc
I made several tests but the windshield still looks weird and the transparency is disappeared. Take a look.
Image
Done 1889752700.000000 mutations (3936.984792 mutations per pixel)
Time elapsed: 19h, 54m, 54s
Mutations per second: 26358.267701
Image

I just need transparent object with a reflection and without IOR. But if I change the IOR to 1.0 the windshield looks like on my previous picture.
Shaders->Spec 0.0 or 0.5 doesn't make the windshield real (no reflection)
:cry: :cry: :cry: :cry:

Posted: Mon Oct 09, 2006 11:13 pm
by kuba
I think the problem could be in normals. Normals at the bottom of the window points downwards and the ones on the top edge into the sky. They are affected by the bottom/top face of the window.

Sorry for my english.

Posted: Mon Oct 09, 2006 11:16 pm
by manitwo
normal-smoothing could also be the bad guy :?

Posted: Sat Oct 14, 2006 7:57 pm
by cooler_inc
Finally I approached the desirable result, I removed polygons from bottom and top.
Image

Posted: Sat Oct 14, 2006 8:32 pm
by Elroc
Cool, looks good.

If I understand your solution correctly, I think Manitwo was correct n saying normal smoothing was causing your problem.