distorted effects in rounded glass facade

[max montana]

hi there,
i face the problem that i get distorted effects when i see through the rounded glass facade. would be nice if i could get some suggestions.
thx

[Sascha Selent]

Is it single-sided geometry? Refracted material needs volume.

If it´s single-sided then the whole thing is treated like a big glas-solid.

[re:FORM]

Firstly, reduce your refraction IOR to something close to 1.0.
Also, check your smoothing groups, the top edges of the glass should be on a seperate smoothing group to the inner and outer faces.

[max montana]

thx for your replies, refraction ior 1,0 solved my problem.

[stevesideas]

IOR 1 may have solved your problem but I dont think that would be correct for glass. As Sascha said, you need to make sure your glass has thickness...use shell modifier and then you can use the correct glass IOR. Think its about 1.53.

[tricky]

...surely if it looks alright it'll do?

[Stoo]

you can use the correct glass IOR. Think its about 1.53.

As an aside, I've used the proper glass IOR of 1.5-odd and the refraction looks totally wrong, far too much distortion (for want of better) between glazed and adjoining open areas. I've had to ramp it down to 1.05 to get what looks, at least visually, correct. Is there perhaps something I've not done which has made the refraction so wild?

[cgrant3d]

As an aside, I've used the proper glass IOR of 1.5-odd and the refraction looks totally wrong, far too much distortion (for want of better) between glazed and adjoining open areas. I've had to ramp it down to 1.05 to get what looks, at least visually, correct. Is there perhaps something I've not done which has made the refraction so wild?

Just a wild guess but did you model your glass curved? In real life the glass between the mullions would be a single flat piece and in your model its probably curving hence the distortion you're getting. If it looks right with your mesh and 1.05 IOR thats fine and you can call it finished. If you want to keep working on it, then I'd make sure the extruded spline that you probably used to create the glass was perfectly flat between the mullions and you'll be able to render it with 1.5 IOR.

[Stoo]

No, this was perfectly flat glass. As you say though, if it looks right that'll do; I just find it curious that using the correct physical values produces an effect which is not desirable.

[Sascha Selent]

Thickness

Again: be sure to use thickness! If you have single-faces and the raytracer hits only one surface it will look wrong! The ray enters a volume, is refracted once, but never get out, thus pointing into a completely wrong direction. It has to refract once more back into the air to get the right direction.

I´m curious: max montana and Stoo: Did you use single-faces? Please check it.

With the right mesh VRay will render proper results!

[Stoo]

Oooh, I'm very Catholic with my glass: Extruded lines to create a mesh with volume and all that, thus everything should be right from the off. When I have the time perhaps I'll review my workflow to check it's not a flaw in my modelling or similar. As I say, by ramping the IOR to 1.05 or thereabouts produces a much more satisfactory result whilst my past and present colleagues also use similar values.

[wxyz]

Nurbs

Here's an example with nurbs and 1.6 IOR

[stevesideas]

That was my point earlier...IOR 1.5 should work with glass that has the correct thickness as Vray is physically correct. Strange..

[max montana]

as i face the problem with curved objects for glass again i refound this post.
i solved the problem last time by reducing the refraction ior to 1,0. not physically correct but it was ok for me. no i remembered the hint to use thickness, and added a shell odifier to the glass, and it works just fine.
now the question: the last object was an spilne, i made the outline 1cm to get "thickness" and extruded the object. but it still gives strange results on the curved areas. so now, what is meant with "thickness"?

[stevesideas]

as i face the problem with curved objects for glass again i refound this post.
i solved the problem last time by reducing the refraction ior to 1,0. not physically correct but it was ok for me. no i remembered the hint to use thickness, and added a shell odifier to the glass, and it works just fine.
now the question: the last object was an spilne, i made the outline 1cm to get "thickness" and extruded the object. but it still gives strange results on the curved areas. so now, what is meant with "thickness"?

Can you post a screengrab of the strange results...