[Fernando Tella]

Hi there!

I'm experiencing some strange behaviour in this scene.

Take a look at the images:

1- In this one the glass wall has a standard glass material applyed:



2- This one has a material with a only reflective layer (opaque) plus a ghost layer in additive mode; no refraction expected:



3- Here comes the strange part; in this one the glass material is made by a standard glass material plus a ghost layer in additive mode:



Where does that refracted image of the padel court comes from? It looks like a lens effect as the image is kind of curved, but if that was the case it should also show in any of the other images.
The glass wall is made by chamfer boxes without smoothing and are not intersecting any other geometry.

I've tryed unchecking multilight, use instancing, deactivating image based environment, making the material from scratch... Same result.

The scene is always launched from 3dsmax 2008 plugin but the same happens from studio.

Any ideas?

Thanks!

[Fernando Tella]

Nobody?

[Tim Ellis]

This is the result of mixing a glass mxm with a ghost layer. You can expect strange refractions and sunlight through the material.

If you remove the ghost layer, the strange refraction will disappear.

I found this during my sunlight through glass tests. http://www.maxwellrender.com/forum/view ... ight+glass

Tim.

[Fernando Tella]

Thanks for replying Tim, I was feeling a bit alone with this.

It is quite odd; I cannot understand where a curved duplicated refraction could come from.

If this things will get solved with 2.0 I would already pay for the upgrade. :/

[tom]

Hi Fernando,

Could you please send me the scene? Before to send, please try to eliminate everything possible while keeping the problem alive. Or you can also make something simple from scratch which has same problem.

Thanks!
tom

[Fernando Tella]

I think I never sent you anything about this problem, did I?

Maybe you still want it or it's already solved... anyway it's easily reproduced.

Refractive + ghost layer > strange result.

[tom]

No, you haven't but, I have to admit I couldn't understand what the problem is exactly. Could you please demonstrate it in the simplest way you can? A few primitive objects with a few materials should be enough to demonstrate it.

[Fernando Tella]

Here it goes:

This scene is made with primitives; two boxes for glasses and some more objects to provide reflections.

The problematic material is composed by two layers: one is quite like a glass from the wizard (I increased the Nd to make reflections more noticeable) and the other one is a ghost layer.
This is the material:

http://www.ftella.com/links/odd%20glass.mxm

I made the test with the problematic material and then another 2 tests with only one of the layers of the material:

Here is the result only with the glass layer:



Looks like expected.

Here is with just the ghost layer:



Just like expected.

Here is (glassy layer + ghost):





Here is the 3dsmax scene (2009):

http://www.ftella.com/links/odd%20glass.max

I hope I explained better this time.

[Fernando Tella]

Any clue?

[tom]

Sorry, I can't see a bug here. What do you expect from blending ghost with glass? Btw, your Nd for glass is 10 which is not correct. It should be around 1.5 instead. Ghost means fully clipped and as you have 25% ghost blended into glass you let %25 of the rays excluded from glass calculation so the final absurd look is quite normal.

[JDHill]

I think it's just what Tim said, since you have a material with two indices of refraction; by the weight of those layers, you will have a result where a certain percent of the light is refracted while the remainder is not. I don't see this as strange, since it is the result you're 'asking' for when building the material:

[Fernando Tella]

@ Tom: I expect just that, a blend of ghost and glass, but this kind of mix is adding some strange curved reflections/refractions which shouldn't be there; reflections/refractions which are not seen neither in glass, neither in ghost. Those are more clearly seen in the third image of my first post. Look at top part of the glass wall where you can clearly see a duplicated and curved refracted image of the paddle court behind.
I know about glass Nd, just raised it for the test to make reflections clearer.

@ JDHill: As I'm using this material for very thin glass pieces I can live with that duplicated refraction. Lower a bit your camera so you can see the border of the base plane through the glass object and maybe you start seeing the weird reflections or refractions (I still don't know where those come from, I think the latter).

I made this kind of material to make building glass faster with natural fresnel reflections and it works besides this oddities.

Animation: ghost > glass > mix

[JDHill]

Well, I couldn't say exactly what happens with refraction when you combine those two things, but if you want the fresnel reflection to be correct, then just use the correct Nd in the reflective layer of an ags material. I'm not sure what happens internally when the material has a split Nd, since you have to consider the internal reflection and such; regardless, that the result can be strange is not surprising to me. If you look in the materials database, there are a few ags materials I made that use the Nd in the reflective layer to give the material a glass-like reflection curve - there are some test renderings with a round structure that show how the materials respond at different viewing angles.

[Fernando Tella]

Yep, that works for glass surrounded by a frame; that's what I'm using right now in most cases: opaque layer plus ghost. That's something that doesn't look good when the glass doesn't have a frame, though; when refraction gives some color to the edge of the glass. I created this material for those cases.

I thought I could add a ghost layer just to make something more transparent without interfering the rest of the layers. That changes a bit the way I think my materials...

This glass thing is not yet completely solved. We need RS2...

[tom]

I thought I could add a ghost layer just to make something more transparent without interfering the rest of the layers. That changes a bit the way I think my materials...

This glass thing is not yet completely solved. We need RS2...

Oh, now I see. You're after dissolving the glass while keeping its reflections intact. Well yes, this is not possible with ghost.