[Radiance-general] translucent glass

[Greg Ward]

Hi Arvinder,

> From: Arvinder Dang <ashudang@yahoo.com>
> Date: Fri Oct 24, 2003  5:01:34  PM US/Pacific
>
> Hello Everybody,
> I am now modelling my skylight case, a 36' x30' room
> with four 4' x 4' skylights distributed evenly.(well
> depth 3'). In my Doe 2 model I am using a flat glass
> which has properties similar to that of a
> 2 dome fiberglass(translucent) with Visible
> transmittance = 0.5, SC = 0.45
> I need to create a similar glass type in Radiance.
> I am using the trans material type and based a on
> previous discussion have created
>
> void trans skylight
> 0
> 0
> 7 0.54 0.54 0.54 0 0 1 0
>
> Is this the correct material definition.
> What would be the definition for a similar white
> translucent material.(Would the color effect the
> illuminance values inside)

How did you come up with 0.54 for the diffuse?  I would have expected 
this to be 0.50 if all you wanted was a 50% visible transmittance.  
However, there is no material that transmits 50% of the visible 
diffusely without also reflecting a fair amount.  Given that you only 
care about the interior values, I suppose it doesn't matter in this 
case, but a more realistic model would be:

void trans skylight
0
0
7 1 1 1 0 0 0.5 0

This would be an ideal scattering surface, spreading the incident 
radiation equally in all directions, 50% transmitted and 50% reflected. 
  There is no such material in the real world, but some materials come 
reasonably close.  The diffuser is white because all of the components 
(red green and blue) are equal.  A colored surface would affect the 
illuminances somewhat.

> Also I know that I need to use mkillum for this.Do I
> need to define both the window surface as well as a
> polygon below the well as illums or simply the polygon
> below the well so as to obtain accurate results.
> Please Guide me with this.

In this special diffuse case, you can hand the skylight polygons to 
mkillum directly, setting

#@mkillum d=0 c=n l-

In the scene file to compute a total average diffuse light source for 
each surface using mkillum.  There is no sense in creating illum's (as 
opposed to light sources) because there is nothing to see through such 
a diffuser, and no sunlight will penetrate it.

-Greg