Glass Shader

[dandelO]

On testing transparency on a built-in Rock Vs. Sphere, I'm getting the black artefacts on both now in an open to the atmosphere setup. Interesting that my indoor Rocks work fine, Spheres didn't, black artefacts again. Maybe it only occurs when atmosphere is behind the transparent surface? And then maybe only on curved surfaces? I'll try one planar Vs. Multipolygonal surface with open atmosphere behind them shortly...
Sphere left, Rock right.

[j meyer]

... each pane should have polies on one side and a hole on the other,so that they are not doublesided ...

What do you mean by this?

In your 6-sided panes example you have 2 big sides and 4 smaller ones,representing
the edges of the pane.My suggestion was removing the 4 small sides.That would leave
you with the 2 big sides and those would be doublesided polygons,but you'd want them
to be singlesided.To make them singlesided - in some apps - you'd have to assign a
hole to one side (the inner sides in this case).
I don't know how this is called in your app,sorry.
Hope it's a bit clearer now.

[Seth]

that's why i dig up that old post lately : http://www.planetside.co.uk/forums/index.php/topic,9943.msg169392.html#msg169392

[gregsandor]

I removed the outer polygons from each pane and re-rendered with no other changes to the scene posted above.  There is no difference between 6-poly and 2-poly glass in the final result. 

[fleetwood]

Don't know if this would help, but you might look at this post which includes the TGD.
What index of refraction are you using in your water shader ?

http://www.planetside.co.uk/forums/index.php/topic,11822.msg119715.html#msg119715

[gregsandor]

What index of refraction are you using in your water shader ?

1.51

[fleetwood]

You might try a test with simple 1.0 and see what the result is. In doing a car windshield I found that it was better to forget about real glass at 1.5 . This isn't a true physical object render so 1.5 may create distortion that doesn't end up looking "real".

[dandelO]

Franck, yep, since v2.1.

An IOR of 1 will actually remove any distortion, not create any, in the transparency, making a perfectly transparent surface where the back and foreview line up exactly. It's a good idea to do that. That's how I did the bee wings, mentioned earlier. Tinx increments of 0.01 are best to add above 1 to have a decent 'glass' type effect, but only when some small shift is needed, when the glass-edge is open, for example, window panes, with a frame around all edges are best left at 1, indeed, as you don't really percieve the refraction then anyway.
Would be handy if it just 'worked', like it used to, though!

[Matt]

I'll take a look at the noise problem.

Matt