View Full Version : Strange IoR Behavior

08-31-2008, 03:57 PM
I ran across a strange behavior with a glass material in an architectural rendering I'm working on. Take a look at the Before Photo. This is a shot of an upstairs interior balcony. The camera is looking at the balcony. On the far side of the balcony are three windows. On the near side of the balcony is a railing with glass panel inserts. Normally the scene is lit by a single distant light to represent the sun, a bluish background to represent blue sky, and polygons at each window with a luminosity factor to represent the light coming in from the windows. (I could instead use an area light at each window, but it renders much slower that way) In Layout I set these illuminating window polys to not cast or receive shadows and to not be sceen by the camera, so in effect they're invisible except for the light they cast. The scene is rendered with Monte Carlo radiosity. In the attached examples the only lights are the window lights, no sunlight and no background light. This explains why the sky outside the windows is black.

If you take a look at the before photo you'll notice that where the glass from the balcony railing overlays the glass from the window it appears completely white. It looks as though it's a reflection, but it is not. It is an anomoly and undesirable. In the after photo the effect is gone. I got the effect to go away by changing the Index of Refraction (IoR) of the glass from 1.5, where it should be, to 1.0. I'd like to maintain the accuracy of my glass by getting this rendering to work with the proper IoR setting. Even an IoR setting of 1.002 will cause the same problem.

Any ideas?

Before Photo (IoR = 1.5)

After Photo (IoR = 1.0)