Re: [Brad] Help:Setting up a radiance system--brad

[Thomas Bleicher]

On 19 Apr 2007, at 05:40, steve michel wrote:

> I understand that not all manufacturers will provide detailed
> 3d models of their fixtures. But  ies standard does include by
> necessity at least the basic geometry of the light fixture.

Yes. However this geometry is usually reduced to the luminous area
of the fitting. This gives you in most cases a rectangle or a disk
instead of a box or cylinder. For recessed fittings there is no  
difference
anyway and a representation of the real shape is beyond the IES
data itself (that's where MGF could be used but hardly ever is).


> My checking with rcalc the lx value of the T5 confirmed to me that  
> radiance
> is doing its job (BTW it is an indirect/direct  fixture). But I  
> wonder if
> there is some bug in brad's implementation of lamps. I have loaded  
> the T5
> into blender and tried making it as a lamp, spot and area light to  
> see the
> result in radiance. In all cases the observable fixture is not even  
> shaped
> like the geometry of the lamps (bulb, t5 etc..).

What is the shape then? brad does not read the IES file itself. It
relies on ies2rad to do the magic.


> When I set the imported fixture as an area light the shape is not even
> close to the lamp geometry. Pointing the camera at any brad library  
> lamp
> fixture (I did follow the instructions) doesn't show a linear t5 as  
> linear
> at all!

The type of the Blender lamp should not have any influence on the
light source if you use photometry. You might be supposed to use
a point light source in Blender. I haven't looked at the code for
a while and managed to install it only yesterday. I will look at it
this evening. I think the lamp type only matters for light sources
without distribution data; and even there only point and spotlight
should be recognised as the brad code predates the Blender area light
feature iirc.

> Even if the light source itself is invisible, its basic shape would
> be visible if placed very close to a flat surfe (eg. ceiling)  
> therefore
> a T5 would show as linear in shape. When I did that with brad, the
> bi directional fixture did not even register on the ceiling..and as a
> spot is shows up as a circle and as an area lamp it is square!

Don't underestimate the pitfalls of Radiance! You have to set the '-ds'
parameter to show the shape of a fitting correctly. See Rob's post and
his script.

> Can anyone confirm this when placing different lamp geometry in brad?

I'll have a look at it now. Can you identify the files which are
created in the conversion process and relate to the luminaire?
If you look at them you might get an idea what brad/ies2rad does.
At least you should be able to check whether there is a box or
only one polygon and what are the dimensions.


Regards,
Thomas