point calcs for verification: oddness

Hi All,

I'm having some trouble understanging something (again). I wanted to
verify that an .ies file I have is being used properly, and I'm getting
some confusing info. I have an .ies-formatted photometry file for a
PAR38. The candela at nadir is 5448. To do a simple verification that
all is well I performed the following steps:

ies2rad -o foo -di -t white -m 1 foo.ies #convert file
oconv foo.rad > foo.oct # create octree
echo "0 0 -120 0 0 1" | rtrace -h -I -ab 0 foo.oct #calculate illuminance
at 10' below fixture

... output from that is:
3.276076e+00 3.276076e+00 3.276076e+00

Taking 3.276706 and multiplying by 179 I get 586 lux, or about 10x what
I'd expect. I'd assumed that by measuring the light source in a void,
that the candela at nadir could be used in a simple point calc, unaffected
by interreflection. So, shouldn't CBCP*d^2 = illuminance ?

What am I doing wrong here?

- Rob

The d in d^w has to be meters, not feet. That could account for the 10x difference.

-G

···

From: "Rob Guglielmetti" <[email protected]>
Date: December 10, 2004 2:36:34 PM PST

Hi All,

I'm having some trouble understanging something (again). I wanted to
verify that an .ies file I have is being used properly, and I'm getting
some confusing info. I have an .ies-formatted photometry file for a
PAR38. The candela at nadir is 5448. To do a simple verification that
all is well I performed the following steps:

ies2rad -o foo -di -t white -m 1 foo.ies #convert file
oconv foo.rad > foo.oct # create octree
echo "0 0 -120 0 0 1" | rtrace -h -I -ab 0 foo.oct #calculate illuminance
at 10' below fixture

... output from that is:
3.276076e+00 3.276076e+00 3.276076e+00

Taking 3.276706 and multiplying by 179 I get 586 lux, or about 10x what
I'd expect. I'd assumed that by measuring the light source in a void,
that the candela at nadir could be used in a simple point calc, unaffected
by interreflection. So, shouldn't CBCP*d^2 = illuminance ?

What am I doing wrong here?

- Rob

Gah, that would be an hour-long brain fart. Of course that would account for the 10(.76)x difference. Thank you, Greg. How embarrassing. What's that smell?

···

On Dec 11, 2004, at 1:04 AM, Greg Ward wrote:

The d in d^w has to be meters, not feet. That could account for the 10x difference.

Rob,

When using IES files, do you think it's better to convert your lamp
color to white or use the real color values? I always use the V-lambda
multiplier (something like... rcalc -e '$1=($1*.265+$2*.67+$3*.065)*179'
) and include the color when using ies2rad. Seems like things get
complicated when you are running calculations and renderings and you
have different types of sources. Maybe white is best?

Mark

Hi Mark,

I use white for lamp color, since the renderings tend to look funky when you use actual lamp colors. RwR goes into this a bit, explaining it better than I could, but the general deal is that the eye is still better than any machine at dealing with visualizing scenes. The eye does an amazing color-balance job, on the fly, that computer displays cannot, so color temperature differences tend to get exaggerated in renderings when you use explicit lamp colors.

On the other hand, I botched a point calc that didn't even involve cosine evaluation, so you may wanna think 2x (or 10.76x) before asking me *anything* about this stuff. :wink:

Still embarrassed,
Rob Guglielmetti

www.rumblestrip.org

···

On Dec 11, 2004, at 1:48 PM, Mark de la Fuente wrote:

Rob,

When using IES files, do you think it's better to convert your lamp color to white or use the real color values? I always use the V-lambda multiplier (something like... rcalc -e '$1=($1*.265+$2*.67+$3*.065)*179' ) and include the color when using ies2rad. Seems like things get complicated when you are running calculations and renderings and you have different types of sources. Maybe white is best?

Hey, I wouldn't have caught your mistake so quickly if I hadn't made it myself already.

From: Rob Guglielmetti <[email protected]>
Date: December 11, 2004 6:53:03 AM PST

The d in d^w has to be meters, not feet. That could account for the 10x difference.

     ^^^^^^^ course, I meant to write "d^2" but didn't bother to proofread

···

On Dec 11, 2004, at 1:04 AM, Greg Ward wrote:

Gah, that would be an hour-long brain fart. Of course that would account for the 10(.76)x difference. Thank you, Greg. How embarrassing. What's that smell?