# integral of radiation in one point -spherical illuminance

Dear Martin, Minki,

Thanks for the replies, I guess what I was trying to calculate (and to
whom dr. Martin refers to) is better called cubical illuminance; the
script Greg shared with Minki (thanks as always, Greg!) allows to sample
a full sphere in one go.

I think I have my ideas a lot clearer now, I'll just need to implement
either one of the approaches.

Thanks,

Giovanni

···

From: Moeck, Dr. Martin [mailto:[email protected]]
Sent: 08 May 2012 12:52
-spherical illuminance

Hi Giovanni,

that is called spherical illuminance. As an approximation, you could
calculate 6 illuminance values (up, down, East, West, North, South) and
average them. Christopher Cuttle wrote a few papers on spherical
illuminance.

Regards

Martin Moeck

OSRAM

________________________________

From: Giovanni Betti [mailto:[email protected]]
Sent: Tuesday, May 08, 2012 1:42 PM

Dear list,

I have question that I hope you'll help get my head around.

I want to calculate the overall illuminance on a point in space that is,
regardless of directionality.

I have made some simplified 2d sketches for clarity.

As I understand a radiance sensor point in rtrace will have cosine
related sensitivity (image01)

If I am to place two coincident with opposing normals (image2) I'll miss
on contributions from the sides.

Rotating the normals by 90 degrees at a time (figure 3) and summing
contributions might not work either because will overestimate diagonal
contributions (figure 4 ).

So I'm not getting too much closer to the solution...

Is there something that I am missing here?

Any light on this will be appreciated,

Best,

Giovanni Betti

Giovanni,

I think there is another way, which requires some more fiddling, but that can give you a lot more flexibility.
Especially with visualisations.
I would start by taking n. 6 x 90º wide angular images of luminance from the observer position.
You could use vwrays to work with rtrace.
Then derive the illuminance at the view point, as you know the solid angle of each pixel / direction and the luminance of it.
See the IESNA book for details. You could write a little script with rcalc, python or octave.
I guess Andy or Greg would do all in one line with sed / awk

Once done you could have polar maps of your illuminance component, directionality of lighting in the space, etc etc.

To start playing with the idea you could do as Mark is suggesting, 6 illuminance values are good as you could easily plot the vector illumiannce to visualise the directionality of lighting.
This is an useful metrics for museums.

Either way, have fun!
G

···

On 8 May 2012, at 14:16, Giovanni Betti wrote:

Dear Martin, Minki,

Thanks for the replies, I guess what I was trying to calculate (and to whom dr. Martin refers to) is better called cubical illuminance; the script Greg shared with Minki (thanks as always, Greg!) allows to sample a full sphere in one go.

I think I have my ideas a lot clearer now, I’ll just need to implement either one of the approaches.

Thanks,

Giovanni

From: Moeck, Dr. Martin [mailto:[email protected]]
Sent: 08 May 2012 12:52

Hi Giovanni,

that is called spherical illuminance. As an approximation, you could calculate 6 illuminance values (up, down, East, West, North, South) and average them. Christopher Cuttle wrote a few papers on spherical illuminance.

Regards

Martin Moeck
OSRAM

From: Giovanni Betti [mailto:[email protected]]
Sent: Tuesday, May 08, 2012 1:42 PM

Dear list,

I have question that I hope you’ll help get my head around.
I want to calculate the overall illuminance on a point in space that is, regardless of directionality.
I have made some simplified 2d sketches for clarity.
As I understand a radiance sensor point in rtrace will have cosine related sensitivity (image01)
If I am to place two coincident with opposing normals (image2) I’ll miss on contributions from the sides.
Rotating the normals by 90 degrees at a time (figure 3) and summing contributions might not work either because will overestimate diagonal contributions (figure 4 ).
So I’m not getting too much closer to the solution…

Is there something that I am missing here?
Any light on this will be appreciated,

Best,
Giovanni Betti
_______________________________________________
[email protected]

Hi Giovanni,

did you try using rsensor?
Using a simple sensor file with all sensor values equal to 1 such as

···

*
degrees 0 90 180 270
0 1 1 1 1
120 1 1 1 1
*
should do the work.

Best,
David

2012/5/9 giulio antonutto <[email protected]>

Giovanni,

I think there is another way, which requires some more fiddling, but that
can give you a lot more flexibility.
Especially with visualisations.
I would start by taking n. 6 x 90º wide angular images of luminance from
the observer position.
You could use vwrays to work with rtrace.
Then derive the illuminance at the view point, as you know the solid angle
of each pixel / direction and the luminance of it.
See the IESNA book for details. You could write a little script with
rcalc, python or octave.
I guess Andy or Greg would do all in one line with sed / awk

Once done you could have polar maps of your illuminance component,
directionality of lighting in the space, etc etc.

To start playing with the idea you could do as Mark is suggesting, 6
illuminance values are good as you could easily plot the vector
illumiannce to visualise the directionality of lighting.
This is an useful metrics for museums.

Either way, have fun!
G

On 8 May 2012, at 14:16, Giovanni Betti wrote:

Dear Martin, Minki,****
** **
Thanks for the replies, I guess what I was trying to calculate (and to
whom dr. Martin refers to) is better called cubical illuminance; the script
Greg shared with Minki (thanks as always, Greg!) allows to sample a full
sphere in one go.****
** **
I think I have my ideas a lot clearer now, I’ll just need to implement
either one of the approaches.****
** **
Thanks,****
** **
Giovanni****
** **
** **
* *
*From:* Moeck, Dr. Martin [mailto:[email protected]]
*Sent:* 08 May 2012 12:52
-spherical illuminance****
** **
Hi Giovanni,****
****
that is called spherical illuminance. As an approximation, you could
calculate 6 illuminance values (up, down, East, West, North, South) and
average them. Christopher Cuttle wrote a few papers on spherical
illuminance.****
****
Regards****
****
Martin Moeck****
OSRAM****
****
****
****
------------------------------
*From:* Giovanni Betti [mailto:[email protected]<[email protected]>
]
*Sent:* Tuesday, May 08, 2012 1:42 PM
****
Dear list,****
****
I have question that I hope you’ll help get my head around.****
I want to calculate the overall illuminance on a point in space that is,
regardless of directionality.****
I have made some simplified 2d sketches for clarity.****
As I understand a radiance sensor point in rtrace will have cosine related
sensitivity (image01)****
If I am to place two coincident with opposing normals (image2) I’ll miss
on contributions from the sides.****
Rotating the normals by 90 degrees at a time (figure 3) and summing
contributions might not work either because will overestimate diagonal
contributions (figure 4 ).****
So I’m not getting too much closer to the solution…****
****
Is there something that I am missing here?****
Any light on this will be appreciated,****
****
Best,****
Giovanni Betti****
_______________________________________________
[email protected]

_______________________________________________
[email protected]

That's a really smart suggestion, David. This didn't even occur to me, and it avoids the issue with source sampling Lars referred to earlier (April 20 post). Guilio's idea would also work, although it might take longer to achieve similar accuracy.

Cheers,
-Greg

···

From: David Geisler-Moroder <[email protected]>
Date: May 9, 2012 2:23:17 AM PDT

Hi Giovanni,

did you try using rsensor?
Using a simple sensor file with all sensor values equal to 1 such as

degrees 0 90 180 270
0 1 1 1 1
120 1 1 1 1

should do the work.

Best,
David

2012/5/9 giulio antonutto <[email protected]>
Giovanni,

I think there is another way, which requires some more fiddling, but that can give you a lot more flexibility.
Especially with visualisations.
I would start by taking n. 6 x 90º wide angular images of luminance from the observer position.
You could use vwrays to work with rtrace.
Then derive the illuminance at the view point, as you know the solid angle of each pixel / direction and the luminance of it.
See the IESNA book for details. You could write a little script with rcalc, python or octave.
I guess Andy or Greg would do all in one line with sed / awk

Once done you could have polar maps of your illuminance component, directionality of lighting in the space, etc etc.

To start playing with the idea you could do as Mark is suggesting, 6 illuminance values are good as you could easily plot the vector illumiannce to visualise the directionality of lighting.
This is an useful metrics for museums.

Either way, have fun!
G

On 8 May 2012, at 14:16, Giovanni Betti wrote:

Dear Martin, Minki,

Thanks for the replies, I guess what I was trying to calculate (and to whom dr. Martin refers to) is better called cubical illuminance; the script Greg shared with Minki (thanks as always, Greg!) allows to sample a full sphere in one go.

I think I have my ideas a lot clearer now, I’ll just need to implement either one of the approaches.

Thanks,

Giovanni

From: Moeck, Dr. Martin [mailto:[email protected]]
Sent: 08 May 2012 12:52

Hi Giovanni,

that is called spherical illuminance. As an approximation, you could calculate 6 illuminance values (up, down, East, West, North, South) and average them. Christopher Cuttle wrote a few papers on spherical illuminance.

Regards

Martin Moeck
OSRAM

From: Giovanni Betti [mailto:[email protected]]
Sent: Tuesday, May 08, 2012 1:42 PM

Dear list,

I have question that I hope you’ll help get my head around.
I want to calculate the overall illuminance on a point in space that is, regardless of directionality.
I have made some simplified 2d sketches for clarity.
As I understand a radiance sensor point in rtrace will have cosine related sensitivity (image01)
If I am to place two coincident with opposing normals (image2) I’ll miss on contributions from the sides.
Rotating the normals by 90 degrees at a time (figure 3) and summing contributions might not work either because will overestimate diagonal contributions (figure 4 ).
So I’m not getting too much closer to the solution…

Is there something that I am missing here?
Any light on this will be appreciated,

Best,
Giovanni Betti