Physically accurate glass

[Pixelcon]

Hi folks,

I'm doing some facade studies and am trying to get my glass setup as close to physically accurate as possible.

We've accurately modeled the glass complete with double glazed modules etc. and I have set up the scene using a vray sun + sky with default values and have set up vray to render in an unbiased fasion with reflective and refractive caustics enabled.


The issue I have is trying to interpret the glass manufacturer's specifications.
Currently my glass is setup as below:
Diffuse: black
Reflect: white
Refract: white
Fresnel: checked
IOR: 1.51
Cutoff: .001
Reflect on back side: on (is this even needed with properly modeled glass panels?)

Now as far as I'm aware this is accurate for basic clear glass but we need to try and implement transmittance and reflectance values as per the manufacturer's specifications.

For example the glass has a maximum reflectance of 20% but there is no mention of whether this is when viewing the glass at a parallel or perpendicular angle.
I had at one stage unchecked the fresnel reflections and inserted a falloff map into the reflection slot as per the image below but I'm not sure if this is the correct approach...



The other specification for the glass is a light transmittance which is set at a value of 40%. Im wondering if there is a numerical way this can be set or if I will have to do some tests and if so, which values should I be adjusting? Theoretically this "value" will be affected by both refraction and reflection and possibly even fog...

Any help would be much appreciated...

[vlado]

For example the glass has a maximum reflectance of 20% but there is no mention of whether this is when viewing the glass at a parallel or perpendicular angle.

It might also be overall reflectance (i.e. averaged over all viewing angles), and I think this is the more likely answer (20% front reflectance seems way too high for glass, and reflectance at grazing angles seems to me to be of little practical use). What's the IOR of the glass? One can write a simple MaxScript to calculate the approximate reflectance given an IOR, and to give you a reflection value to put in the V-Ray material to get that a particular reflectance.

The other specification for the glass is a light transmittance which is set at a value of 40%. Im wondering if there is a numerical way this can be set or if I will have to do some tests and if so, which values should I be adjusting? Theoretically this "value" will be affected by both refraction and reflection and possibly even fog...

The transmittance would indeed depend on the refraction, reflection and the fog color and will change with different IORs. Is the light transmittance given for a glass sheet of particular thickness? If yes, then once the reflection color is determined from the above procedure, one can figure out how to set the refraction color too: reflectance is 20%, so 80% remain for absorption/transmission - so if you set the refraction color to medium grey, you'll get 40% transmission.

Best regards,
Vlado

[trixian]

But just trying those numbers in my head it seems this type of glass would look rather strange. Would someone be willing to try to set up a controlled comparison? That is, if one can get a real sample of the glass or a good photo in RAW format from the manufacturer. Would be interesting to see how close they get.

[vlado]

But just trying those numbers in my head it seems this type of glass would look rather strange.

It depends; it could be some sort of dark or tinted glass...

Best regards,
Vlado

[Pixelcon]

This is typical facade glass with a bit of tint. Our local laws are now restricting how reflective tower glass is now so we need to carry out these tests (it's also to show the senior designers in the studio just how far divorced from reality their vision really is!!)

I did a test yesterday afternoon with a single window module setup and measured the reflectance of a vrayplane light as well as using the vraylightmeter to measure transmitted light and I must say we were a lot closer to the specified values than I had expected.

Thanks for the tips, Vlado, these points seem like good questions to pose to the manufacturers as well as facade engineers.
Trix, I had been thinking of doing what you've mentioned and will definitely push for a sample so we can do some testing!

Thanks for the reply, guys. I will see what I can do. I think it would be good to have some solid information on this type of thing for future reference.