[Fernando Tella]

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

Reflections and refractions (the edge effect which makes it more... solid) but also being able to see light through glass and having fast transparent shadows.

I want it all, yep.

[Bubbaloo]

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.

That's when you apply AGS to the surfaces and glass to the edges.

[Fernando Tella]

Does it look good?

[Bubbaloo]

I've used it once on a render for production, and it looked great and saved a lot of render time.



All of the glass in this rendering except the exterior window uses that method. You get good edges and quick clearing glass panels.

[Fernando Tella]

Thanks Brian. I remember that WIP. I think I thought it was too laborious. Really looks nice and now I don't think it's so much trouble.

[Fernando Tella]

Re-reading some of the answers in Tim Ellis' topic about glass I just found a new (I think) invention to workaround some of the problems with glass.

As an mxm mixing real glass with ghost layer gave strange reflections/refractions as shown in this thread I tried another approach.

First the image, then the explanations:



Looks quite good, doesn't it?

-To get this I have duplicated glass geometry;
-moved the second one just half a milimeter (if they are exactly at the same position it produces weird results);
-applied a real glass mxm to the first glass geometry and hide from GI (so it looks good but doesn't cast shadows);
-applied a ghost material (opaque bsdf 100% plus ghost layer 50% in this example) to the second geometry and hide to camera and reflections/refractions; this gives the shadow and doesn't change the look of the glass;

Done!

The shadow is not perfect as it comes from a ghost material but it's the closer I can get to the real thing and it's super-fast. The image shown reached SL7 only and looks quite nice already.

[tom]

And how does it look like with AGS alone?

[Fernando Tella]

AGS dissolves the edges of the glass:

AGS



Freak's invention:

[Bubbaloo]

It looks great!

[tom]

Very nice!

[Hervé]

very cool trick... when you say move the glass 1/2.. what do you mean exactly... move y axis...?

h.

[Fernando Tella]

very cool trick... when you say move the glass 1/2.. what do you mean exactly... move y axis...?

h.

I mean: move it just a little, because if they are exactly at the same place it produces some weird results. Maybe I should test that condition a bit further.

[Maximus3D]

That's a very clever solution, and the result looks great too. Convincing enough to fool most people.

/ Max

[Fernando Tella]

Thanks Max,

I have to test more things about this. It should distort what's behind because of refraction... pools, caustics...

[jomaga]

Very cool glass, Fernando!
thanks!