Strange alignment in reflections on windows glass

[Ronnie_T]

Hi all,

 

I'm new here and relatively new to Vray.

 

I've attached an image so you can see just what the problem is. Basically the reflections in my windows are rendering at strange and unrealistic angles to the objects that create them.

 

The project itself is a college complex designed in Archicad and exported as a 3ds file in layers, then imported to Max and Vray and then each element was retextured with Vray materials.

 

The glass object was one mesh for the whole complex, and was a single sided object and I allready knew that to render the glass properly in Vray I should make it a solid object.

So I then detached various window panes and then added a shell modifier to them in order to acheive solid glass objects.

 

I've tried many different glass materials so I'm pretty sure I've eliminated that as the problem.

 

Also I have created box objects to replaced the original window meshes, just to see if they rendered with the correct reflections but they had exactly the same effect.

 

The cameras in Max were at 90-100 deg, so I wondered if it was the FOV bending the rendered image but it quite clearly is not.

 

So I'm dumbfounded with this as I've used Max for years and although I'm pretty new to Vray, reflections that are rendered so incorrectly isn't something I have come accross before in any 3d program.

 

Help most needed:confused:

 

Regards.

[AubreyM]

Try a falloff map in the glass material reflection slot.

[Tommy L]

Add an edit poly modifier, select the sub object then go to the smoothing panel and remove smoothing.

[Ronnie_T]

Hi AubreyM

I already have the material set up just like that, in fact I've also been through the entire Evermotion Archshaders libraries 1+2 glass mats as well, I've tried so many approaches to the material and it just renders the same every time:confused:

 

Hello Tommy L,

I tried out your suggestion and added the modifier, selected all poly's and removed all smoothing groups but it's still no go, it looks exactly the same as the image I provided Im afraid.

[Stephen Thomas]

Ok, I'm going to state the obvious. It seems to me that the mezzanine floor and the window aren't perpendicular to each other (judging by the angle of the wall relative to the desk and carpet texture). In which case the reflections look entirely correct to me!

[creasia]

We often have glass rendering problems with imported geometry. In our studio and have a few different solutions.

 

1. Most times replacing the glass with a max cube will solve your problems. (You can use edit mesh on your old glass object and attach the new max cube geometry if you wish to replace many instances of the glass. Depending on the bug, this may not work)

 

2. Make sure the faces are not flipped inwards on a polygon. You can use edit mesh and turn on show normals which will show blue lines pointing outwards on the direction that that the face is oriented.

 

3. Set your refraction index IOR to much lower. Default is 1.6 I think that oxygen is 1.013 keep it above 1. You will lose accurate refraction values, but your reflection and transparency remain.

Edited October 17, 2011 by creasia

[Ronnie_T]

Hi stef.thomas and Cheers for the tips creasia:)

 

Ok, I'm going to state the obvious. It seems to me that the mezzanine floor and the window aren't perpendicular to each other (judging by the angle of the wall relative to the desk and carpet texture). In which case the reflections look entirely correct to me!

 

I agree with you, however to my eyes the angle is to steep and should be in a lower position as illustrated in attachment, especially if you consider the desk and ballustrade are in line with each other. Also in second image the building in foreground and background are almost perpundicular to each other but again the reflection in the window in foreground seems incorrect but maybe I'm in need of some real life study here!

[CliveG]

For my 10 cents worth (or 6.5 pence) I agree with Stef.Thomas.

 

Why don't you just experiment with tiny adjustments to the angle of the glass and see whether this confirms or further confuses the issue?

[Ronnie_T]

For my 10 cents worth (or 6.5 pence) I agree with Stef.Thomas.

 

Why don't you just experiment with tiny adjustments to the angle of the glass and see whether this confirms or further confuses the issue?

 

Why I never thought of doing that, lord knows:o well done though complete-newbie and you're right stef.thomas you were correct as you will see below.

When I rotated the window to 90 deg to the ballustrade the reflections are indeed correct.

 

However look at the other image and see that the reflection of the revolving door, now that seems screwed up to me, but as you've proved it can't be:confused: