# Modeling Lamps

Hello,

I am modeling several fluorescent troffers and am a little confused with
the "Glow" primitive commonly used to model lamps. I have figured out
that it does not pass an illum surface and cause any additional
illumination. It does, however, change the luminance of the luminaire
when viewed directly (i think only when using the "illum" primitive for
the ies data and not the "light" primitive). So, I would like to get
this "Glow" correct so that I can compare the luminance ratios of
various fixtures. In Desktop Radiance's fixture library it seems to
typically use 53.195959 as the RGB radiance value. I can't figure out
where this number came from. First of all, what units are these
Radiance values in? W/ster/m2? If this is the true than I calculate
for say a T8;

32 W / 4*pi steradians / 2*pi*0.5 sf = 0.8105 W/ster/sf

0.8105 W/ster/sf * (m2/ 10.764sf) = 0.0753 W/ster/m2

This doesn't seem right. Please correct me if there is any error in my
calculation.

If the radiance vaue is supposed to be photometric and in units of
lumens/ster/m2 than I calculate;

2900 lumens / 4*pi steradians / 2*pi*0.5 sf = 73.46 lumens/ster/sf

73.46 lumens/ster/sf * (m2/ 10.764sf) = 6.824 lumens/ster/m2

Then the Rendering with Radiance book on Page 322 and 323 gives examples
of using the "lampcolor" program and for a daylight fluorescent (? not
sure what this is) with 2900 lumen output it gives RGB values of
0.867251 1.189160 1.066163.

I am confused. And seeing as how this "Glow" primitive completely
effects the luminance of my parabolic troffers I do not want to continue
my comparison without assurance that I am modeling them correctly. Any

Regards,
Zack Rogers

Zack Rogers wrote:

Hello,

I am modeling several fluorescent troffers and am a little confused with
the "Glow" primitive commonly used to model lamps. I have figured out
that it does not pass an illum surface and cause any additional
illumination. It does, however, change the luminance of the luminaire
when viewed directly (i think only when using the "illum" primitive for
the ies data and not the "light" primitive). So, I would like to get
this "Glow" correct so that I can compare the luminance ratios of
various fixtures. In Desktop Radiance's fixture library it seems to
typically use 53.195959 as the RGB radiance value. I can't figure out
where this number came from. First of all, what units are these
Radiance values in? W/ster/m2? If this is the true than I calculate
for say a T8;

32 W / 4*pi steradians / 2*pi*0.5 sf = 0.8105 W/ster/sf

0.8105 W/ster/sf * (m2/ 10.764sf) = 0.0753 W/ster/m2

This doesn't seem right. Please correct me if there is any error in my
calculation.

If the radiance vaue is supposed to be photometric and in units of
lumens/ster/m2 than I calculate;

2900 lumens / 4*pi steradians / 2*pi*0.5 sf = 73.46 lumens/ster/sf

73.46 lumens/ster/sf * (m2/ 10.764sf) = 6.824 lumens/ster/m2

Then the Rendering with Radiance book on Page 322 and 323 gives examples
of using the "lampcolor" program and for a daylight fluorescent (? not
sure what this is) with 2900 lumen output it gives RGB values of
0.867251 1.189160 1.066163.

I am confused. And seeing as how this "Glow" primitive completely
effects the luminance of my parabolic troffers I do not want to continue
my comparison without assurance that I am modeling them correctly. Any

Reality is often much less complicated than expected.

You give a luminous flux of 2900 lm for your type of 32W T8 (this
can vary with color/quality). Given the standard efficacy used by
Radiance, this amounts to 16.2 W of light energy (which can't be
directly determined from the electrical consumption!)

The tube has a circumference of approximately 81.7 mm and a
luminous length of around 1.198 m. This results in a luminous
surface of 0.0979 m2.

The radiance of this tube is therefore:

16.2W / (0.0979m2 * Pi) = 52.7 W/m2/sr

This is what the average rgb value of your glow material should
be based on. The luminance of the tube then amounts to 7684 cd/m2.

Btw: If I feed the lampcolor program with those values, then I
get very different results than you do:

\$> lampcolor
Program to compute lamp radiance. Enter '?' for help.
Enter lamp type [WHITE]: daylight fluorescent
Enter length unit [meter]:
Enter lamp geometry [polygon]: cylinder
Cylinder length [1]: 1.198
Enter total lamp lumens [0]: 2900
Lamp color (RGB) = 35.450659 48.609340 43.581585

0.265*35.450659 + 0.670*48.609340 + 0.065*43.581585 = 44.8

This is 0.85 of the 52.7 above. Lampcolor has a depreciacion
factor for each lamp type, possibly accounting for aging (the
actual purpose of the factor is not documented).

-schorsch

···

--
Georg Mischler -- simulations developer -- schorsch at schorsch.com
+schorsch.com+ -- lighting design tools -- http://www.schorsch.com/

Zack Rogers wrote:

Hello,

I am modeling several fluorescent troffers and am a little confused with
the "Glow" primitive commonly used to model lamps. I have figured out
that it does not pass an illum surface and cause any additional
illumination. It does, however, change the luminance of the luminaire
when viewed directly (i think only when using the "illum" primitive for
the ies data and not the "light" primitive). So, I would like to get
this "Glow" correct so that I can compare the luminance ratios of
various fixtures. In Desktop Radiance's fixture library it seems to
typically use 53.195959 as the RGB radiance value. I can't figure out
where this number came from. First of all, what units are these
Radiance values in? W/ster/m2? If this is the true than I calculate
for say a T8;

<snip>

Then the Rendering with Radiance book on Page 322 and 323 gives examples
of using the "lampcolor" program and for a daylight fluorescent (? not
sure what this is) with 2900 lumen output it gives RGB values of
0.867251 1.189160 1.066163.

A "daylight fluorescent" is a lamp with a color temperature of 5000 K or above, like that of daylight.

I am confused. And seeing as how this "Glow" primitive completely
effects the luminance of my parabolic troffers I do not want to continue
my comparison without assurance that I am modeling them correctly. Any

Reality is often much less complicated than expected.

Ah, but modeling and simulating reality is a different matter.

You give a luminous flux of 2900 lm for your type of 32W T8 (this
can vary with color/quality). Given the standard efficacy used by
Radiance, this amounts to 16.2 W of light energy (which can't be
directly determined from the electrical consumption!)

The tube has a circumference of approximately 81.7 mm and a
luminous length of around 1.198 m. This results in a luminous
surface of 0.0979 m2.

The radiance of this tube is therefore:

16.2W / (0.0979m2 * Pi) = 52.7 W/m2/sr

This is what the average rgb value of your glow material should
be based on. The luminance of the tube then amounts to 7684 cd/m2.

Btw: If I feed the lampcolor program with those values, then I
get very different results than you do:

\$> lampcolor
Program to compute lamp radiance. Enter '?' for help.
Enter lamp type [WHITE]: daylight fluorescent
Enter length unit [meter]:
Enter lamp geometry [polygon]: cylinder
Cylinder length [1]: 1.198
Enter total lamp lumens [0]: 2900
Lamp color (RGB) = 35.450659 48.609340 43.581585

0.265*35.450659 + 0.670*48.609340 + 0.065*43.581585 = 44.8

This is 0.85 of the 52.7 above. Lampcolor has a depreciacion
factor for each lamp type, possibly accounting for aging (the
actual purpose of the factor is not documented).

-schorsch

Thank you, Georg, for the excellent summary of the steps in modeling a lamp. My only suggestion would be to consider modeling the lamp as "WHITE" if there is only going to be one type (color temperature) lamp in the scene. This will avoid hassles with color balancing the image with pfilt.

The depreciation factor associated with each lamp type is defined in the lamp.tab file. The depreciation factor is intended to account for the difference between the rated "initial lumens" of the lamp and the effective "mean lumens" that one would expect to get over the rated life of the lamp. This is according to a discussion with Greg many years ago.

-Chas

Charles Ehrlich
Project Manager
Heschong Mahone Group
11626 Fair Oaks Blvd. #302
Fair Oaks, CA 95628
Phone: (916) 962-7001
Fax: (916) 962-0101
Email: [email protected]
Web: http://www.h-m-g.com

···

---------------------------------
Do You Yahoo!?
Check out Yahoo! Shopping and Yahoo! Auctionsfor all of your holiday gifts!

Zack Rogers wrote:

If the radiance vaue is supposed to be photometric and in units of
lumens/ster/m2 than I calculate;

2900 lumens / 4*pi steradians / 2*pi*0.5 sf = 73.46 lumens/ster/sf

73.46 lumens/ster/sf * (m2/ 10.764sf) = 6.824 lumens/ster/m2

Then the Rendering with Radiance book on Page 322 and 323 gives examples
of using the "lampcolor" program and for a daylight fluorescent (? not
sure what this is) with 2900 lumen output it gives RGB values of
0.867251 1.189160 1.066163.

Howdy Zack!

The total output of a source is found by integrating its radiance L over
the directions of emission and over its entire surface area A. This
effectively gives you an outer integral over the area and an inner
integral over the directions per surface element. The latter is weighted
by the cosine of the angle to the surface normal and the differential
solid angle. Lacking ascii art skills I'll refrain from cramming the
formula in here.

In the very simplistic situation where the emission is invariant with
respect to direction and position on the luminaire (i.e. diffuse), this
integral simply becomes:

Power = L * pi * A,

The RGB radiance values in the example above are derived from these
rules (lampcolor assumes a diffuse emitted), and they also include a
depreciation factor for the luminaire.

I am confused.

Me too! :^)

···

--
"Life is too boring for short videos" (TechVision)