[Radiance-general] RE: Lamp colour

[Greg Ward]

Hello Anthony,

Your question is a bit confusing.  I don't know what lamp .fmt file  
you're talking about, and the numbers you give don't seem to make  
sense.  The CIE (x,y) chromaticity coordinates corresponding to 2200K  
and 6000K sources are:

CT (K)    CIE x    CIE y
--------    --------    -------
2200    0.5056    0.4152
6000    0.3221    0.3318

Converting these to Radiance RGB values with Y=1.0, we get:

CT    R    G    B
-----    --    --    --
2200    1.881    0.7420    0.0636
6000    0.9078    1.032    1.049

If you use the "-t default" option and give these to the "-c" option  
of ies2rad, it will produce sources with the corresponding color  
temperature.  However, I don't really believe this is what you want.

Applying a color to your light sources will yield a color cast in  
your rendered images, which is almost never desirable.  Your eye  
tends to ignore such color casts in an environment -- this is called  
"color constancy" or "white balance" in photography.

If your true purpose is to determine how the color of your light  
source affects the visible colors of the objects in your scene, you  
need to start from the actual spectral power distribution of your  
illuminant (not just the color temperatures) and the spectral  
reflectances of your objects.  The former may be obtained from the  
manufacturer in some cases.  The latter must be obtained by  
measurements with a suitable spectroreflectometer.  Then, rather than  
dealing with CIE (x,y) chromaticities, you need to apply a technique  
like the one described in the following paper:

     http://www.anyhere.com/gward/egwr02/

This paper describes the simplest and most economical approach if you  
want to know what your colors will look like under different  
illuminants.  Any shortcut shorter than this will simply not work.

-Greg

P.S.  In future, please be careful not to include an entire mailing  
list digest in your posting.  It's an easy mistake to make with  
modern mailers -- they are ever so helpful and eager to please.

> From: "Anthony J. Farrell" <anthony.farrell at dit.ie>
> Date: May 23, 2005 1:51:10 PM PDT
>
> Hi Folks,
>
> I would be grateful if anyone would provide some insight to a  
> problem I've
> been trying to resolve for some time:
>
> I am using fluorescent lamps in an office scene, and would like to  
> model
> colour temperatures of approx. 2200K and 6000K giving a warm white and
> daylight white respectively.
>
> I have read through the Radiance manual but am a little confused  
> still as to
> how to set lamp format file (.fmt) to give me the required colours.
>
> I have generated a .dat file from the ies data for the lamp within  
> which I
> set the colour to the default set at 0.907 0.863 0.4762: I believe  
> that this
> represents 48 46 25% RGB values.
>
> However, when I run a lighting illuminance simulation the .lum file  
> refers
> only to the lamp .dat file and .fmt file (not the .rad file  
> generated by
> ies2rad). Now upond exploring the .fmt file it refers to the 'void  
> lamp
> glow' in this instance its: 4 0 3.57 14.72 860
>
> I think that this may be the line I need to change to alter the  
> colour?
> Unless it's already taken care of in the ies2rad conversion via  
> the .dat
> file in which case I need to alter the 0.907 0.863 0.4762 values  
> relative to
> desired K values. I'd be indebted to anyone that can provide the  
> relevan
> combination for the two lamp colours I'm looking for.
>
> With kind regards,
>
> Anthony Farrell
> Dublin Institute of Technology