direct/indirect photometric

Some improvements have been made to ies2rad since the book was
published, and direct/indirect fixtures should produce better
renderings now, I believe.

However, if you want to insert your own visible luminaire geometry, the
-i option is the way, but do not replace the geometry that ies2rad
produces. Instead, make sure that the illum sphere you specify
encloses the geometry you add for your visible fixture, and create this
fixture using "glow" modifiers for the emitting surfaces rather than
"light" or "illum" materials.

I hope this helps.
-Greg

Greg,

Wouldn't I need to modify the illum sphere when I'm creating a linear fluorescent luminaire so that the illum geometry fits tightly around the long luminaire? Is it ok to create a sphere with a radius of 2 feet for a 4 feet long luminaire, and then just leave it alone?

John

Hi John,

You have a point, especially since the illum sphere will never be subdivided during source sampling -- i.e., its distribution will always eminate from a point. However, you must do one thing if you substitute the geometry for a fixture; you must multiply the A1 scalefactor in the brightdata primitive by the ratio of the illum sphere area divided by the illum replacement geometry area, or the total lumens will not be maintained during the switch. Alternatively, you can specify an illum radius whose area (4*PI*radius^2) exactly equals the area of the replacement geometry.

-Greg

···

From: John An <[email protected]>

Wouldn't I need to modify the illum sphere when I'm creating a linear fluorescent luminaire so that the illum geometry fits tightly around the long luminaire? Is it ok to create a sphere with a radius of 2 feet for a 4 feet long luminaire, and then just leave it alone?

Greg and John An, Unless the new version of Radiance is significantly different that when I wrote chapter 5, I do not think that you need to change A1 if you follow the method described in the book. The method was specifically designed to address linear sources. It is also the method employed in the luminaire editor in Desktop Radiance. Is it still true that the "lboxcorr" function automatically corrects for the surface area of the emitter? If not, then we have a major re-write on our hands for the new edition. -Chas

···

[email protected] wrote:
Date: Mon, 26 May 2003 09:02:47 -0700
Subject: Re: [Radiance-general] Re: direct/indirect photometric
From: Greg Ward
To: [email protected]
Reply-To: [email protected]

Hi John,

You have a point, especially since the illum sphere will never be
subdivided during source sampling -- i.e., its distribution will always
eminate from a point. However, you must do one thing if you substitute
the geometry for a fixture; you must multiply the A1 scalefactor in the
brightdata primitive by the ratio of the illum sphere area divided by
the illum replacement geometry area, or the total lumens will not be
maintained during the switch. Alternatively, you can specify an illum
radius whose area (4*PI*radius^2) exactly equals the area of the
replacement geometry.

-Greg

From: John An

Wouldn't I need to modify the illum sphere when I'm creating a linear
fluorescent luminaire so that the illum geometry fits tightly around
the long luminaire? Is it ok to create a sphere with a radius of 2
feet for a 4 feet long luminaire, and then just leave it alone?

--__--__--

_______________________________________________
Radiance-general mailing list
[email protected]
http://www.radiance-online.org/mailman/listinfo/radiance-general

End of Radiance-general Digest

Hi Chas,

You are right -- I had forgotten this feature of lboxcorr (and boxcorr). There is a caveat, though, which is using the -i value of 0.56419 requires that you be working in meters, and likewise the box dimensions must also be specified in meters for the units to work out correctly. In this case, the A1 parameter gets set to 1 by ies2rad and the lboxcorr applies the correction. (Remember that the box must be centered at the origin for lboxcorr to work properly.)

-Greg

···

From: Charles Ehrlich <[email protected]>

Greg and John An,

Unless the new version of Radiance is significantly different that when I wrote chapter 5, I do not think that you need to change A1 if you follow the method described in the book.� The method was specifically designed to address linear sources.� It is also the method employed in the luminaire editor in Desktop Radiance.

Is it still true that the "lboxcorr" function automatically corrects for the surface area of the emitter?� If not, then we have a major re-write on our hands for the new edition.

-Chas

Date: Tue, 27 May 2003 09:43:33 -0700 (PDT)
From: Charles Ehrlich <[email protected]>

Greg and John An, Unless the new version of Radiance is significantly different that when I wrote chapter 5, I do not think that you need to change A1 if you follow the method described in the book. The method was specifically designed to address linear sources. It is also the method employed in the luminaire editor in Desktop Radiance. Is it still true that the "lboxcorr" function automatically corrects for the surface area of the emitter? If not, then we have a major re-write on our hands for the new edition. -Chas

Chas and Greg,

Thanks for the info. I was away on a short vacation, and just got back. I'm still more than a bit confused about accurately modeling direct/indirect luminaires. Could you tell me if I have this right?

Background info:
My space is modeled in inches.

The photometric file is in feet (as reported by ies_read):

Luminaire Dimensions

···

--------------------
Measurement units = Feet
Width = 0.46
Length = 4.00
Height = 0.00

Question 1: I have noticed that the height for all of the direct/indirect luminaires is 0. How is that physically possible? More importantly, how would a dimension of 0 ever "enclose" the geometry of the luminaire?

Question 2: The actual luminaire dimensions are larger than the dimensions given by ies_read. For instance, the length of the luminaire is listed in the cut sheet as being 53-5/8". Which dimension do I use to create the illum "enclosure?"

I used the command:
ies2rad -di -t white -i .56419 -o light2 d21998.ies

and the resulting light2.rad file reads:

# ies2rad -di -t white -i .56419 -o light2
# Dimensions in inches
#<IESNA:LM-63-1995
#<[TEST] 21998
#<[DATE] 05/31/2001
#<[LUMCAT] AGDMWU232-F3-120-1/2-EB
#<[LUMINAIRE] DAY-BRITE ADAGIO DIRECT/INDIRECT W/33-CELL SPECULAR LOUVER
#<[LAMP] F32T8
#<[BALLAST] ADVANCE REL-2P32-SC
#<[MANUFAC] DayBrite-Capri-Omega(Genlyte Thomas Group)
#<[_TIFF_FILE_NAME] AGD2
#<[_VERSION] fo2ies 2.01
#<Reflection factor 0.92, Test distance = 26 ft.
# 60 watt luminaire, lamp*ballast factor = 0.88

void brightdata light2_dist
5 corr light2.dat source.cal src_phi4 src_theta
0
1 1364

light2_dist illum light2_light
0
3 1 1 1

light2_light sphere light2.s
0
4 0 0 0 0.56419

____

I edit this file to use lboxcorr. I convert inches to meters (multiply inches by .0254).
...
void brightdata light2_dist
5 lboxcorr light2.dat source.cal src_phi4 src_theta
0
4 .815 1.363 .305 .0921
...
Question 3: Because I used the -di option in the ies2rad command, I thought that converting inches to meters was the way to go. Should I be converting feet to meters?
____

Question 4: When I replace the illum sphere with a box, should I be using the units of my model (inches) or sticking with meters? In other words, should I be using:

!genbox light2_light light2.s 1.363 .305 .0921 \
  > xform -t -.6815 -.1525 -.04605

or

!genbox light2_light light2.s 53.625 12 3.625 \
  > xform -t -26.8125 -6 -1.8125

I have tried both, and when I used meters, my image is black. When I use inches, I get a reasonable image of my space.

Question 5: Can the dimensions of the illum geometry be slightly different than the dimensions defined for the lboxcorr function? The rounding required in converting back and forth between units would inevitably result in some differences.

Thanks again.

John

I don't really have time for this, so my answers are going to be short.

From: John An <[email protected]>

Chas and Greg,

Thanks for the info. I was away on a short vacation, and just got back. I'm still more than a bit confused about accurately modeling direct/indirect luminaires. Could you tell me if I have this right?

Background info:
My space is modeled in inches.

The photometric file is in feet (as reported by ies_read):

Luminaire Dimensions
--------------------
Measurement units = Feet
Width = 0.46
Length = 4.00
Height = 0.00

Question 1: I have noticed that the height for all of the direct/indirect luminaires is 0. How is that physically possible? More importantly, how would a dimension of 0 ever "enclose" the geometry of the luminaire?

The heights are set to zero by ies2rad to avoid a "dark band" at the level of the luminaire when it's hung from the ceiling. It's a hack to avoid artifacts. You're right that it cannot enclose any geometry this way.

Question 2: The actual luminaire dimensions are larger than the dimensions given by ies_read. For instance, the length of the luminaire is listed in the cut sheet as being 53-5/8". Which dimension do I use to create the illum "enclosure?"

Whatever you need to enclose the geometry you create.

I used the command:
ies2rad -di -t white -i .56419 -o light2 d21998.ies

and the resulting light2.rad file reads:

# ies2rad -di -t white -i .56419 -o light2
# Dimensions in inches
#<IESNA:LM-63-1995
#<[TEST] 21998
#<[DATE] 05/31/2001
#<[LUMCAT] AGDMWU232-F3-120-1/2-EB
#<[LUMINAIRE] DAY-BRITE ADAGIO DIRECT/INDIRECT W/33-CELL SPECULAR LOUVER
#<[LAMP] F32T8
#<[BALLAST] ADVANCE REL-2P32-SC
#<[MANUFAC] DayBrite-Capri-Omega(Genlyte Thomas Group)
#<[_TIFF_FILE_NAME] AGD2
#<[_VERSION] fo2ies 2.01
#<Reflection factor 0.92, Test distance = 26 ft.
# 60 watt luminaire, lamp*ballast factor = 0.88

void brightdata light2_dist
5 corr light2.dat source.cal src_phi4 src_theta
0
1 1364

light2_dist illum light2_light
0
3 1 1 1

light2_light sphere light2.s
0
4 0 0 0 0.56419

____

I edit this file to use lboxcorr. I convert inches to meters (multiply inches by .0254).
...
void brightdata light2_dist
5 lboxcorr light2.dat source.cal src_phi4 src_theta
0
4 .815 1.363 .305 .0921
...
Question 3: Because I used the -di option in the ies2rad command, I thought that converting inches to meters was the way to go. Should I be converting feet to meters?

The A2, A3, and A4 arguments should be in meters. Where did you get the 0.815 value? I assume this is a lumen depreciation factor.

____

Question 4: When I replace the illum sphere with a box, should I be using the units of my model (inches) or sticking with meters? In other words, should I be using:

!genbox light2_light light2.s 1.363 .305 .0921 \
  > xform -t -.6815 -.1525 -.04605

or

!genbox light2_light light2.s 53.625 12 3.625 \
  > xform -t -26.8125 -6 -1.8125

I have tried both, and when I used meters, my image is black. When I use inches, I get a reasonable image of my space.

You should use whatever corresponds to your -d? option -- inches in this case. I would recommend generating the box in meters then adding -s 39.37 to your xform command to make it clear:

!genbox light2_light light2.s 1.363 .305 .0921 \
  > xform -s 39.37 -t -26.8125 -6 -1.8125

Question 5: Can the dimensions of the illum geometry be slightly different than the dimensions defined for the lboxcorr function? The rounding required in converting back and forth between units would inevitably result in some differences.

They should be close -- they don't have to be exact.

-Greg