rtcontrib

radiance-general
1

Hello,

I've been trying to use daylight coefficients with rtcontrib and I have a couple of questions.
I made images with the contribution of each sky patch, but the images look quite noisy, and if I combine them I don`t get the same as if I was rendering the scene altogether.
The command I used is like this:

vwrays -ff -x 500 -y 500 -vf vint.vf | rtcontrib -ffc `vwrays -d -x 500 -y 500 -vf vint.vf` -ab 5 -ad 2048 -as 1024 -b tbin -o c_%d.pic -m skyglow -f tregenza_h.cal scene.oct

If necessary I can post some images, but probably it`s something obvious I`m missing here...

My second question is regarding the calculation of model skies (like Perez) with the rtcontrib results. Perez sky gives the luminance value at a certain point, but I guess in order to multiply the contribution of a certain section of the sky, I need to integrate the luminance values inside that section. Is this correct? Can anyone point me to any reference where this might be explained more? Any ideas about how to implement this in a script?
Thank you in advance. Best regards,

Santiago

2

Hi Santiago,

I wish I could offer you some words of encouragement, because I appreciate the difficulty you face applying rtcontrib to daylight coefficients. The truth is, you really need some scripts to help out with the task, and I haven't yet had a chance to write a general set of scripts or an executive program for DC analysis. I am working on this project currently, as a matter of fact.

All I have at this point is a few test scripts I assembled for a daylight simulation study, and I can make these available for you if you need an early start on this. If you can wait a few weeks, I will have something much better, I hope.

Just to answer your question, adding together the rtcontrib outputs directly gives you the result for a uniform sky. To get a particular sky, you must multiply each image by the average radiance of the corresponding sky patch. To compute this, a simple rtrace calculation using the desired sky model and a set of randomly distributed rays over the hemisphere will do the trick. Such a set of rays is given in the file "ray/lib/tregsamp.dat", where there are 64 rays per sky patch. The command to compute pcomb arguments from this file is:

rtrace -h skymodel.oct < tregsamp.dat \
         > total -64 -m \
         > rcalc -o '-s ${$2} c_${recno-1}.pic'

I realize this is just a clue, but if you wait a few weeks, hopefully I can provide a complete solution.

-Greg

···

From: Santiago Torres <tiago@tkh.att.ne.jp>
Date: October 30, 2005 2:21:32 PM PST

Hello,

I've been trying to use daylight coefficients with rtcontrib and I have a couple of questions.
I made images with the contribution of each sky patch, but the images look quite noisy, and if I combine them I don`t get the same as if I was rendering the scene altogether.
The command I used is like this:

vwrays -ff -x 500 -y 500 -vf vint.vf | rtcontrib -ffc `vwrays -d -x 500 -y 500 -vf vint.vf` -ab 5 -ad 2048 -as 1024 -b tbin -o c_%d.pic -m skyglow -f tregenza_h.cal scene.oct

If necessary I can post some images, but probably it`s something obvious I`m missing here...

My second question is regarding the calculation of model skies (like Perez) with the rtcontrib results. Perez sky gives the luminance value at a certain point, but I guess in order to multiply the contribution of a certain section of the sky, I need to integrate the luminance values inside that section. Is this correct? Can anyone point me to any reference where this might be explained more? Any ideas about how to implement this in a script?
Thank you in advance. Best regards,

Santiago

3

Hello experts:

If using -m option in rtcontrib, can I calculate the contribution from a
general object that does not belong to light source, such as objects having
plastic materials?

My model has two windows, and I want to calculate the contribution of each
window to the illuminance at sensor points. The "Three-Phase Daylight
Coefficient" method works. But for simple windows, can I directly calculate
the contribution using rtcontrib or others? Any suggestion is greatly
appreciated.

Thank you!

Jia

4

"Yes" is the short answer. Using the rtcontrib -V+ option, you will get the contribution from each surface or set of surfaces selected by -m.

Best,
-Greg

···

From: Jia Hu <hujia06@gmail.com>
Date: July 13, 2011 2:22:41 PM PDT

Hello experts:

If using -m option in rtcontrib, can I calculate the contribution from a general object that does not belong to light source, such as objects having plastic materials?

My model has two windows, and I want to calculate the contribution of each window to the illuminance at sensor points. The "Three-Phase Daylight Coefficient" method works. But for simple windows, can I directly calculate the contribution using rtcontrib or others? Any suggestion is greatly appreciated.

Thank you!

Jia

5

Thank you! I have other questions related to this program.

(1) Question 1:

In the case that both of "-f", "-b" and "-m" are used, the .cal algorithm
and "-b" option are applied to objects with "-m" materials, for example,

vwrays -ff $vw |rtcontrib @image.opt -ffc $(vwrays -d $vw) -f tregenza.cal
-b tbin -o images/patches/p%03d%s.hdr \
-m sky_glow -m ground_glow -w testroom_whitesky.oct

If I replace "ground_glow" with "ceiling_mat", the .cal will applied to
"ceiling" geometry and divide ceiling into many patches according to the ray
direction and intersection with ceiling?

(2) Question 2:

Assume my model has two windows in the south wall and one sensor in the
interior building, and I wish to calculate the daylighting coefficient of
each patch of sky and each window to the illuminance at the sensor point.
The result is like a matrix as follows:

                    SkyPatch001, SkyPatch001,
Window01 coefficient (RGB) coefficient (RGB)
Window02 coefficient (RGB) coefficient (RGB)

Each coefficient shows the contribution of each sky patch and each window to
the sensor illuminance. Can I achieve this in one rtcontrib command?
Is the following expression right?

[input ray] | rtcontrib -f tregenza.cal -o window_%s.data -b tbin -m
window01_mat -b -b tbin -m window02_mat [other necessary options]

Thank you,

Jia

···

On Wed, Jul 13, 2011 at 7:05 PM, Greg Ward <gregoryjward@gmail.com> wrote:

"Yes" is the short answer. Using the rtcontrib -V+ option, you will get
the contribution from each surface or set of surfaces selected by -m.

Best,
-Greg

> From: Jia Hu <hujia06@gmail.com>
> Date: July 13, 2011 2:22:41 PM PDT
>
> Hello experts:
>
> If using -m option in rtcontrib, can I calculate the contribution from a
general object that does not belong to light source, such as objects having
plastic materials?
>
> My model has two windows, and I want to calculate the contribution of
each window to the illuminance at sensor points. The "Three-Phase Daylight
Coefficient" method works. But for simple windows, can I directly calculate
the contribution using rtcontrib or others? Any suggestion is greatly
appreciated.
>
> Thank you!
>
> Jia

_______________________________________________
Radiance-general mailing list
Radiance-general@radiance-online.org
http://www.radiance-online.org/mailman/listinfo/radiance-general

6

Hi Jia,

I will answer your questions the best I can inline...

From: Jia Hu <hujia06@gmail.com>
Date: July 14, 2011 9:12:12 AM PDT

Thank you! I have other questions related to this program.

(1) Question 1:

In the case that both of "-f", "-b" and "-m" are used, the .cal algorithm and "-b" option are applied to objects with "-m" materials, for example,

vwrays -ff $vw |rtcontrib @image.opt -ffc $(vwrays -d $vw) -f tregenza.cal -b tbin -o images/patches/p%03d%s.hdr \
-m sky_glow -m ground_glow -w testroom_whitesky.oct

If I replace "ground_glow" with "ceiling_mat", the .cal will applied to "ceiling" geometry and divide ceiling into many patches according to the ray direction and intersection with ceiling?

Yes, the bins will still be based on ray direction due to the "tbin" definition in "tregenza.cal", but will apply to rays striking surfaces using "ceiling_mat" instead of "ground_glow". The intersection position on the ceiling will not affect "tbin" since it doesn't use this information, so the same direction at one corner of the ceiling will go to the same bin as a ray hitting the opposite corner in that direction.

(2) Question 2:

Assume my model has two windows in the south wall and one sensor in the interior building, and I wish to calculate the daylighting coefficient of each patch of sky and each window to the illuminance at the sensor point. The result is like a matrix as follows:

                    SkyPatch001, SkyPatch001,
Window01 coefficient (RGB) coefficient (RGB)
Window02 coefficient (RGB) coefficient (RGB)

Each coefficient shows the contribution of each sky patch and each window to the sensor illuminance. Can I achieve this in one rtcontrib command?
Is the following expression right?

[input ray] | rtcontrib -f tregenza.cal -o window_%s.data -b tbin -m window01_mat -b -b tbin -m window02_mat [other necessary options]

You don't need to give the "-b tbin" option twice, and the "-b" option by itself will generate an error, but otherwise it is OK. Note that this command neglects any exterior obstructions that might interfere with the travel of a sample ray from the exterior of the window to the sky. If your building has external shades or obstructions outside, the above technique will not produce accurate results.

The general ability to choose combined interactions in rtcontrib is lacking. E.g., you can't ask for all rays that pass through "window1" then go on to intersect "building2" and reflect to "skypatch17". It would be a useful feature to add, but such accounting is problematic.

Best,
-Greg

7

Many thanks!!!

Hi Jia,

I will answer your questions the best I can inline...

> From: Jia Hu <hujia06@gmail.com>
> Date: July 14, 2011 9:12:12 AM PDT
>
> Thank you! I have other questions related to this program.
>
> (1) Question 1:
>
> In the case that both of "-f", "-b" and "-m" are used, the .cal algorithm
and "-b" option are applied to objects with "-m" materials, for example,
>
> vwrays -ff $vw |rtcontrib @image.opt -ffc $(vwrays -d $vw) -f
tregenza.cal -b tbin -o images/patches/p%03d%s.hdr \
> -m sky_glow -m ground_glow -w testroom_whitesky.oct
>
> If I replace "ground_glow" with "ceiling_mat", the .cal will applied to
"ceiling" geometry and divide ceiling into many patches according to the ray
direction and intersection with ceiling?

Yes, the bins will still be based on ray direction due to the "tbin"
definition in "tregenza.cal", but will apply to rays striking surfaces using
"ceiling_mat" instead of "ground_glow". The intersection position on the
ceiling will not affect "tbin" since it doesn't use this information, so the
same direction at one corner of the ceiling will go to the same bin as a ray
hitting the opposite corner in that direction.

My understanding of rtcontrib is that it calls rtrace to trace ray as the

normal process but rtrace onlys reports the ray coefficient modifer in
modifer list specified by "-ti" in rtrace.

>
> (2) Question 2:
>
> Assume my model has two windows in the south wall and one sensor in the
interior building, and I wish to calculate the daylighting coefficient of
each patch of sky and each window to the illuminance at the sensor point.
The result is like a matrix as follows:
>
> SkyPatch001, SkyPatch001,
> Window01 coefficient (RGB) coefficient (RGB)
> Window02 coefficient (RGB) coefficient (RGB)
>
> Each coefficient shows the contribution of each sky patch and each window
to the sensor illuminance. Can I achieve this in one rtcontrib command?
> Is the following expression right?
>
> [input ray] | rtcontrib -f tregenza.cal -o window_%s.data -b tbin -m
window01_mat -b -b tbin -m window02_mat [other necessary options]

You don't need to give the "-b tbin" option twice, and the "-b" option by
itself will generate an error, but otherwise it is OK. Note that this
command neglects any exterior obstructions that might interfere with the
travel of a sample ray from the exterior of the window to the sky. If your
building has external shades or obstructions outside, the above technique
will not produce accurate results.

The general ability to choose combined interactions in rtcontrib is
lacking. E.g., you can't ask for all rays that pass through "window1" then
go on to intersect "building2" and reflect to "skypatch17". It would be a
useful feature to add, but such accounting is problematic.

You are right! it is problematic when external shades or obstructions

exist. Unfortunately my case has external shades. Hope you could add this
feature in the near future.

Thanks,

Jia

···

On Thu, Jul 14, 2011 at 12:52 PM, Greg Ward <gregoryjward@gmail.com> wrote: