We are trying to do a daylight analysis that includes some frosted glass or “kalwall” style skylights which diffuse the light into the space.  We can get the VLT (Visible Light Transmittance) values easy enough.  However, is there some way to accurately (or semi-accurately) account for the rays being dispersed and spread through the frosted glass? 

Simulating frosted glazing in 3ds Max Design for lighting analysis is doable.  You however need to know how to do it properly.  Here is how:

Some useful background information:

First, as opposed to Radiance, the A&D Material has a few internal “things” going on that you need to be aware.  The most important one is that the A&D Material performs internal energy conservation as follow:

Transmissivity wins over Specular Reflectivity which wins over Diffuse Reflectance.  On top of that, the Transmissivity is weighted against a Specular / Diffuse factor.  This factor is ruled by the Translucent color / weight controls in the interface.

In contrast, in Radiance, one can specify a material that is reflecting 100% diffuse and 100% specular while transmitting 100% of the light, leading to “creating” energy.  This is why the parameters of the Radiance materials cannot be plugged “as-is” in the A&D Material.

Translucent Panels:

We compared translucent glazing simulation in mental ray against radiance and measured data and got convincing results (see image) with the following settings:

Desired Diffuse Transmittance:  0.1621  (16.21% Diffuse – Diffuse Transmittance)

• Treat surface as a single polygon in the model
• A&D Diffuse Level:  0.0
• A&D Diffuse Color:  pitch black  (so no weighting is given to the diffuse reflectance)
• A&D Reflection | Reflectivity Level:  1.0
• A&D Reflection | Reflectivity Color:  pure white (the color is a multiplier, we need it to be 1.0 1.0 1.0)
• A&D Refraction |  Transparency Level:  1.0
• A&D Refraction | Transparency Color:  pure white (the color is a multiplier, we need it to be 1.0 1.0 1.0)
•  A&D Refraction | Translucency Checkbox : ON
• A&D Refraction | Translucency Weight: 1.0  (we want it fully translucent)
• A&D Refraction | Translucency Color:  0.1621 0.1621 0.1621 (the color is a multiplier, we need it to be set to the desied transmissivity level “as-is”, equally for all RGB components)
• BRDF | Custom Reflectivity Function: ON
• BRDF | 0 Deg Refl:  0.0
• BRDF | 90 Deg Refl:  1.0
• BRDF | Curve:  ~5  (we need to approximate a typical Fresnel curve)
• Advanced Rendering Options | Thin-Walled : ON

Some important notes: 

• While the illuminances will carry through properly on light meters, the glazing appearance may not look “natural” in “pretty picture renderings”.   It seems that there is currently a limitation with the appearance of the surface when it is hit by light:  its resulting luminance won’t be correct (AFAIK) so glare analysis based on luminance measurements won’t be convincing.
• The following image shows a graph comparing 3ds max, radiance and measured data.  Ignore the “3ds max 2009 SP1 Initial Submission” curve, this is representing a case where our material settings in 3ds Max where wrong, which we corrected later on – in fact, we forgot to turn off a layer so we had 2 panes of glass on top of each other…).  The green curve is what we need to look at….

Always verify your data otherwise your project can go wrong – here are a couple advices on the topic:

IES file can contain wrong information: double check them!

I have once seen a manufacturer using the wrong values in their IES files.  I had to contact them and demonstrate that their data was wrong. It turned out that one of their engineer made a mistake with their database system where North American IES files where published using European measurements. That basically means that errors can come from anywhere in the process. Double check your work!

In the ideal scenario, you need to get a physical sample of the luminaire in your office, measure intensities with an illuminance meter and compare with the IES data provided to you in a simple test scene.

Do the numbers match up?  You are good to go. They don’t match up? Keep reading..

Check the 3ds Max scene units:

This is a typical trap: wrong units, wrong light intensities reported on light meters!

3ds Max and Revit has bugs too!

There is a issue  with the IES file reader that can affect you in rare occasions.  The bug has to do with a certain “feature” of the IES file format that some manufacturers rely on. Let me explain:

The IES file format is divided in two main sections: a header section and a data section:

In the header section, there is a special ”bit” that instruct the software to either use the “raw” data or apply a multiplier to it. For example, a manufacturer might have measured a luminaire only once with a 100W lamp in it. Knowing that the same luminaire with a 200W lamp** would emit twice as much light, the manufacturer could publish a different IES file based on the same measurements and simply state that the “raw” data should be multiplied by a factor of 2.0 by changing this special “bit” in the file as illustrated:

For some reason, 3ds Max and Revit are ignoring this multiplier and always read the “raw” data from the IES files, resulting in incorrect simulations. The work around is to use the built-in multiplier (dimmer) feature of the photometric lights:

Note about linear and area lights:

In many occurences, I am asked to perform lighting analysis reports in spaces designed with cove lights.  Unfortunately, this is still not possible to do precisely - and this problem has not been solved by the industry yet!

  There are currently two major problems with linear and area lights in lighting simulation softwares:

1. First, the IESNA standards for photometric measurements assumes that luminaires are measured from a certain distance, making any calculated points inside this distance mathematically inaccurate. If you search for “Near Field Photometry” or “Far Field Photometry”, you will find plently of information on the subject.
2. Second, 3ds max (and Revit btw, as both share the same code for rendering) still has issues to deal with linear and area lights calculation making them unreliable photometric-wise.

As I write these lines, cove lighting situations cannot be mathematically correct with photometric files in 3ds Max: you will have to do some guess work when working with theses scenarios.

**I know, a 200W lamp does not emit twice as much light than a 100W lamp that was only for keeping the example simple.