Radiance 5.0 installation on ubuntu 14.04

radiance-general
1

I everyone,

I was using a version of radiance from the apt-get, i guess the version we
get when we install radiance like that is 4.1, but i need some functions
from the newer version like gendaymtx, i have tried to install from source
code with the ./makeall install and i received the message "there were some
error" i have read all the info on readme and nox11help but nothing seems
to work, after doing a bunch of stuff nothing was working at all.

I needed to make a clean install of ubuntu so i did it and now i have my
computer completely clean.

Now i want to make a clean installation of radiance 5.0 to get it running
smoothly.

Should i use the binaries from NREL to LINUX 64 or compile it from source,
witch is better?

Thank to all
Humberto A.

2

I would still recommend compiling from source using the makeall script, because it’s worked well for years and years. The dreaded “there were some errors” message is usually because you did not run the makeall script as root (i.e. "sudo makeall install” instead of just “makeall install”). You also need to make sure you have the “auxiliary files" archive in addition to just the source code.

The NREL binaries for Linux were built on Ubuntu, so they would probably work ok for you, assuming you can get everything unpacked and copied to the right locations; the makeall script just does this for you a little more easily, IMHO.

···

On Mar 17, 2016, at 7:20 PM, Humberto Antunes <[email protected]> wrote:

I everyone,

I was using a version of radiance from the apt-get, i guess the version we get when we install radiance like that is 4.1, but i need some functions from the newer version like gendaymtx, i have tried to install from source code with the ./makeall install and i received the message "there were some error" i have read all the info on readme and nox11help but nothing seems to work, after doing a bunch of stuff nothing was working at all.

I needed to make a clean install of ubuntu so i did it and now i have my computer completely clean.

Now i want to make a clean installation of radiance 5.0 to get it running smoothly.

Should i use the binaries from NREL to LINUX 64 or compile it from source, witch is better?

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

3

Easiest thing to do, usually, is just download the compiled version, unpack
it, and move it to /usr/local.

···

--
Randolph M. Fritz, Lighting Design and Simulation
+1 206 659-8617 || [email protected]

On Thu, Mar 17, 2016 at 6:20 PM, Humberto Antunes < [email protected]> wrote:

I everyone,

I was using a version of radiance from the apt-get, i guess the version we
get when we install radiance like that is 4.1, but i need some functions
from the newer version like gendaymtx, i have tried to install from source
code with the ./makeall install and i received the message "there were some
error" i have read all the info on readme and nox11help but nothing seems
to work, after doing a bunch of stuff nothing was working at all.

I needed to make a clean install of ubuntu so i did it and now i have my
computer completely clean.

Now i want to make a clean installation of radiance 5.0 to get it running
smoothly.

Should i use the binaries from NREL to LINUX 64 or compile it from source,
witch is better?

Thank to all
Humberto A.

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

4

Humberto,

You mentioned X11, but didn't share an error associated with X11. So It's
tough to tell what your problem is. Can you search the text on the terminal
for the word error, and report the actual errors that occurred?

In case it is related to X11, on Raspbian, I need to install the x11
development libraries to get Radiance to compile.
The same could be true for Ubuntu?

Here's what I do:
sudo apt-get install libx11-dev

Best,
Andy

···

On Thu, Mar 17, 2016 at 6:20 PM, Humberto Antunes < [email protected]> wrote:

I everyone,

I was using a version of radiance from the apt-get, i guess the version we
get when we install radiance like that is 4.1, but i need some functions
from the newer version like gendaymtx, i have tried to install from source
code with the ./makeall install and i received the message "there were some
error" i have read all the info on readme and nox11help but nothing seems
to work, after doing a bunch of stuff nothing was working at all.

I needed to make a clean install of ubuntu so i did it and now i have my
computer completely clean.

Now i want to make a clean installation of radiance 5.0 to get it running
smoothly.

Should i use the binaries from NREL to LINUX 64 or compile it from source,
witch is better?

Thank to all
Humberto A.

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

5

Thanks for the help

I would prefer to compile from source to.
So if i download the file *"rad5R0all.tar.gz"*, unpack it, let's say, in
the desktop directory and then run *"sudo ./makeall install"* from the ray
directory everything should be installed smoothly right?

Humberto A.

6

At this point i downloaded and unpacked the "rad5R0all.tar.gz" file and now
when I try to run "sudo ./makeall install" i get the fallowing error
message:
"sudo: unable to execute ./makeall: No such file or directory".

Any hints on how to solve it?

Thank you
Humberto A.

···

2016-03-18 3:05 GMT+00:00 Humberto Antunes <[email protected]>:

Thanks for the help

I would prefer to compile from source to.
So if i download the file *"rad5R0all.tar.gz"*, unpack it, let's say, in
the desktop directory and then run *"sudo ./makeall install"* from the
ray directory everything should be installed smoothly right?

Humberto A.

7

That error is already solved, just needed to install csh.

So after i run the sudo makeall install i got the message "There were some
errors", like you asked me to I went back to see what the error are and the
error listed is:

compilation terminated.
make: *** [x11findwind.o] Error 1

So I followed your advice installed libx11-dev and got the message "Done",
so i guess the installation of radiance is finally fine.

I am now starting you tutorial "The Three-Phase Method for Simulating
Complex Fenestration with Radiance", hope everything goes smoothly.

Thank you everyone for your help.

Best regards
Humberto A.

···

2016-03-18 2:56 GMT+00:00 Andy McNeil <[email protected]>:

Humberto,

You mentioned X11, but didn't share an error associated with X11. So It's
tough to tell what your problem is. Can you search the text on the terminal
for the word error, and report the actual errors that occurred?

In case it is related to X11, on Raspbian, I need to install the x11
development libraries to get Radiance to compile.
The same could be true for Ubuntu?

Here's what I do:
sudo apt-get install libx11-dev

Best,
Andy

On Thu, Mar 17, 2016 at 6:20 PM, Humberto Antunes < > [email protected]> wrote:

I everyone,

I was using a version of radiance from the apt-get, i guess the version
we get when we install radiance like that is 4.1, but i need some functions
from the newer version like gendaymtx, i have tried to install from source
code with the ./makeall install and i received the message "there were some
error" i have read all the info on readme and nox11help but nothing seems
to work, after doing a bunch of stuff nothing was working at all.

I needed to make a clean install of ubuntu so i did it and now i have my
computer completely clean.

Now i want to make a clean installation of radiance 5.0 to get it running
smoothly.

Should i use the binaries from NREL to LINUX 64 or compile it from
source, witch is better?

Thank to all
Humberto A.

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

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

8

Hi Humberto,
I'm glad the compilation worked. I'm surprised you were able to figure out
that you needed to install csh to get makeall to run. The error "no such
file or directory" doesn't give much clue of what is wrong.

I suspect you will have problems with the commands in the three phase
tutorial. They're not up to date, and some options no longer exist or are
un-necessary. Also, the rfluxmtx command makes the process a lot simpler.
I've been working on updating the tutorial, but have made slow progress.
You might consider using the 2014 workshop presentation as a guide to the
newer methods:
https://www.radiance-online.org/community/workshops/2014-london/presentations/day1/McNeil_BSDFsandPhases.pdf

Andy

···

On Fri, Mar 18, 2016 at 6:55 AM, Humberto Antunes < [email protected]> wrote:

That error is already solved, just needed to install csh.

So after i run the sudo makeall install i got the message "There were some
errors", like you asked me to I went back to see what the error are and the
error listed is:

compilation terminated.
make: *** [x11findwind.o] Error 1

So I followed your advice installed libx11-dev and got the message
"Done", so i guess the installation of radiance is finally fine.

I am now starting you tutorial "The Three-Phase Method for Simulating
Complex Fenestration with Radiance", hope everything goes smoothly.

Thank you everyone for your help.

Best regards
Humberto A.

2016-03-18 2:56 GMT+00:00 Andy McNeil <[email protected]>:

Humberto,

You mentioned X11, but didn't share an error associated with X11. So It's
tough to tell what your problem is. Can you search the text on the terminal
for the word error, and report the actual errors that occurred?

In case it is related to X11, on Raspbian, I need to install the x11
development libraries to get Radiance to compile.
The same could be true for Ubuntu?

Here's what I do:
sudo apt-get install libx11-dev

Best,
Andy

On Thu, Mar 17, 2016 at 6:20 PM, Humberto Antunes < >> [email protected]> wrote:

I everyone,

I was using a version of radiance from the apt-get, i guess the version
we get when we install radiance like that is 4.1, but i need some functions
from the newer version like gendaymtx, i have tried to install from source
code with the ./makeall install and i received the message "there were some
error" i have read all the info on readme and nox11help but nothing seems
to work, after doing a bunch of stuff nothing was working at all.

I needed to make a clean install of ubuntu so i did it and now i have my
computer completely clean.

Now i want to make a clean installation of radiance 5.0 to get it
running smoothly.

Should i use the binaries from NREL to LINUX 64 or compile it from
source, witch is better?

Thank to all
Humberto A.

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

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

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

9

I figured it out when i opened the makeall file from the ray directory i
saw the first line.

I am already having trouble with the commands in the tutorial, so after a
quick reading of the presentation I see that basically the rfluxmtx commamd
comes in to replace the rcontrib command, right?

I am using radiance to run annual dynamic daylight simulation of dynamic
shading devices for my master's thesis, hope i will be able to do it
because my master's thesis advisor is not much of a help in that topic.

I will try to get the three phase method running with the help of that
presentation.

Thank you very much Andy
Humberto A.

···

2016-03-18 17:18 GMT+00:00 Andy McNeil <[email protected]>:

Hi Humberto,
I'm glad the compilation worked. I'm surprised you were able to figure out
that you needed to install csh to get makeall to run. The error "no such
file or directory" doesn't give much clue of what is wrong.

I suspect you will have problems with the commands in the three phase
tutorial. They're not up to date, and some options no longer exist or are
un-necessary. Also, the rfluxmtx command makes the process a lot simpler.
I've been working on updating the tutorial, but have made slow progress.
You might consider using the 2014 workshop presentation as a guide to the
newer methods:

https://www.radiance-online.org/community/workshops/2014-london/presentations/day1/McNeil_BSDFsandPhases.pdf

Andy

On Fri, Mar 18, 2016 at 6:55 AM, Humberto Antunes < > [email protected]> wrote:

That error is already solved, just needed to install csh.

So after i run the sudo makeall install i got the message "There were
some errors", like you asked me to I went back to see what the error are
and the error listed is:

compilation terminated.
make: *** [x11findwind.o] Error 1

So I followed your advice installed libx11-dev and got the message
"Done", so i guess the installation of radiance is finally fine.

I am now starting you tutorial "The Three-Phase Method for Simulating
Complex Fenestration with Radiance", hope everything goes smoothly.

Thank you everyone for your help.

Best regards
Humberto A.

2016-03-18 2:56 GMT+00:00 Andy McNeil <[email protected]>:

Humberto,

You mentioned X11, but didn't share an error associated with X11. So
It's tough to tell what your problem is. Can you search the text on the
terminal for the word error, and report the actual errors that occurred?

In case it is related to X11, on Raspbian, I need to install the x11
development libraries to get Radiance to compile.
The same could be true for Ubuntu?

Here's what I do:
sudo apt-get install libx11-dev

Best,
Andy

On Thu, Mar 17, 2016 at 6:20 PM, Humberto Antunes < >>> [email protected]> wrote:

I everyone,

I was using a version of radiance from the apt-get, i guess the version
we get when we install radiance like that is 4.1, but i need some functions
from the newer version like gendaymtx, i have tried to install from source
code with the ./makeall install and i received the message "there were some
error" i have read all the info on readme and nox11help but nothing seems
to work, after doing a bunch of stuff nothing was working at all.

I needed to make a clean install of ubuntu so i did it and now i have
my computer completely clean.

Now i want to make a clean installation of radiance 5.0 to get it
running smoothly.

Should i use the binaries from NREL to LINUX 64 or compile it from
source, witch is better?

Thank to all
Humberto A.

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

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

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

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

10

Hello again everybody,

I have a question related to the command rfluxmtx.
While using rcontrib to generate the view matrix we had to set the window
orientation with the command -b kbinS or (N,E or W), how do we do that in
the rfluxmtx? Is it set by the "u" variable or it is not needed to do this
with the rfluxmtx command?

Also, as reinhart.cal file been replaced for the reinhartb.cal?

Best regards
Humberto A.

···

2016-03-18 18:27 GMT+00:00 Humberto Antunes <[email protected]>:

I figured it out when i opened the makeall file from the ray directory i
saw the first line.

I am already having trouble with the commands in the tutorial, so after a
quick reading of the presentation I see that basically the rfluxmtx commamd
comes in to replace the rcontrib command, right?

I am using radiance to run annual dynamic daylight simulation of dynamic
shading devices for my master's thesis, hope i will be able to do it
because my master's thesis advisor is not much of a help in that topic.

I will try to get the three phase method running with the help of that
presentation.

Thank you very much Andy
Humberto A.

2016-03-18 17:18 GMT+00:00 Andy McNeil <[email protected]>:

Hi Humberto,
I'm glad the compilation worked. I'm surprised you were able to figure
out that you needed to install csh to get makeall to run. The error "no
such file or directory" doesn't give much clue of what is wrong.

I suspect you will have problems with the commands in the three phase
tutorial. They're not up to date, and some options no longer exist or are
un-necessary. Also, the rfluxmtx command makes the process a lot simpler.
I've been working on updating the tutorial, but have made slow progress.
You might consider using the 2014 workshop presentation as a guide to the
newer methods:

https://www.radiance-online.org/community/workshops/2014-london/presentations/day1/McNeil_BSDFsandPhases.pdf

Andy

On Fri, Mar 18, 2016 at 6:55 AM, Humberto Antunes < >> [email protected]> wrote:

That error is already solved, just needed to install csh.

So after i run the sudo makeall install i got the message "There were
some errors", like you asked me to I went back to see what the error are
and the error listed is:

compilation terminated.
make: *** [x11findwind.o] Error 1

So I followed your advice installed libx11-dev and got the message
"Done", so i guess the installation of radiance is finally fine.

I am now starting you tutorial "The Three-Phase Method for Simulating
Complex Fenestration with Radiance", hope everything goes smoothly.

Thank you everyone for your help.

Best regards
Humberto A.

2016-03-18 2:56 GMT+00:00 Andy McNeil <[email protected]>:

Humberto,

You mentioned X11, but didn't share an error associated with X11. So
It's tough to tell what your problem is. Can you search the text on the
terminal for the word error, and report the actual errors that occurred?

In case it is related to X11, on Raspbian, I need to install the x11
development libraries to get Radiance to compile.
The same could be true for Ubuntu?

Here's what I do:
sudo apt-get install libx11-dev

Best,
Andy

On Thu, Mar 17, 2016 at 6:20 PM, Humberto Antunes < >>>> [email protected]> wrote:

I everyone,

I was using a version of radiance from the apt-get, i guess the
version we get when we install radiance like that is 4.1, but i need some
functions from the newer version like gendaymtx, i have tried to install
from source code with the ./makeall install and i received the message
"there were some error" i have read all the info on readme and nox11help
but nothing seems to work, after doing a bunch of stuff nothing was working
at all.

I needed to make a clean install of ubuntu so i did it and now i have
my computer completely clean.

Now i want to make a clean installation of radiance 5.0 to get it
running smoothly.

Should i use the binaries from NREL to LINUX 64 or compile it from
source, witch is better?

Thank to all
Humberto A.

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

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

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

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

11

Hi Humberto,

The rfluxmtx command takes over for genklemsamp and calls rcontrib with the options needed for correct operation. You don't need to specify the bin or the .cal file(s), as rfluxmtx takes care of that part. You do need to set both the "u=Z" (or whatever) up-direction and the desired hemispherical sampling type (e.g., "h=r4") in your Radiance scene description handed to rfluxmtx as the source or receiver(s).

The reinhartb.cal file is more general than the original reinhart.cal file, and uses slightly different mechanisms (hence the need for a new name).

I hope this helps, and thanks to everyone for pitching in. I realize that the changes made post-tutorial have caused a good deal of confusion, but we hope to get it all sorted out in the end.

Cheers,
-Greg

···

From: Humberto Antunes <[email protected]>
Subject: Re: [Radiance-general] Radiance 5.0 installation on ubuntu 14.04
Date: March 18, 2016 2:49:55 PM PDT

Hello again everybody,

I have a question related to the command rfluxmtx.
While using rcontrib to generate the view matrix we had to set the window orientation with the command -b kbinS or (N,E or W), how do we do that in the rfluxmtx? Is it set by the "u" variable or it is not needed to do this with the rfluxmtx command?

Also, as reinhart.cal file been replaced for the reinhartb.cal?

Best regards
Humberto A.

2016-03-18 18:27 GMT+00:00 Humberto Antunes <[email protected]>:
I figured it out when i opened the makeall file from the ray directory i saw the first line.

I am already having trouble with the commands in the tutorial, so after a quick reading of the presentation I see that basically the rfluxmtx commamd comes in to replace the rcontrib command, right?

I am using radiance to run annual dynamic daylight simulation of dynamic shading devices for my master's thesis, hope i will be able to do it because my master's thesis advisor is not much of a help in that topic.

I will try to get the three phase method running with the help of that presentation.

Thank you very much Andy
Humberto A.

2016-03-18 17:18 GMT+00:00 Andy McNeil <[email protected]>:
Hi Humberto,
I'm glad the compilation worked. I'm surprised you were able to figure out that you needed to install csh to get makeall to run. The error "no such file or directory" doesn't give much clue of what is wrong.

I suspect you will have problems with the commands in the three phase tutorial. They're not up to date, and some options no longer exist or are un-necessary. Also, the rfluxmtx command makes the process a lot simpler. I've been working on updating the tutorial, but have made slow progress. You might consider using the 2014 workshop presentation as a guide to the newer methods:
https://www.radiance-online.org/community/workshops/2014-london/presentations/day1/McNeil_BSDFsandPhases.pdf

Andy

On Fri, Mar 18, 2016 at 6:55 AM, Humberto Antunes <[email protected]> wrote:
That error is already solved, just needed to install csh.

So after i run the sudo makeall install i got the message "There were some errors", like you asked me to I went back to see what the error are and the error listed is:

compilation terminated.
make: *** [x11findwind.o] Error 1

So I followed your advice installed libx11-dev and got the message "Done", so i guess the installation of radiance is finally fine.

I am now starting you tutorial "The Three-Phase Method for Simulating Complex Fenestration with Radiance", hope everything goes smoothly.

Thank you everyone for your help.

Best regards
Humberto A.

2016-03-18 2:56 GMT+00:00 Andy McNeil <[email protected]>:
Humberto,

You mentioned X11, but didn't share an error associated with X11. So It's tough to tell what your problem is. Can you search the text on the terminal for the word error, and report the actual errors that occurred?

In case it is related to X11, on Raspbian, I need to install the x11 development libraries to get Radiance to compile.
The same could be true for Ubuntu?

Here's what I do:
sudo apt-get install libx11-dev

Best,
Andy

On Thu, Mar 17, 2016 at 6:20 PM, Humberto Antunes <[email protected]> wrote:
I everyone,

I was using a version of radiance from the apt-get, i guess the version we get when we install radiance like that is 4.1, but i need some functions from the newer version like gendaymtx, i have tried to install from source code with the ./makeall install and i received the message "there were some error" i have read all the info on readme and nox11help but nothing seems to work, after doing a bunch of stuff nothing was working at all.

I needed to make a clean install of ubuntu so i did it and now i have my computer completely clean.

Now i want to make a clean installation of radiance 5.0 to get it running smoothly.

Should i use the binaries from NREL to LINUX 64 or compile it from source, witch is better?

Thank to all
Humberto A.

12

Oh, but I did neglect to mention a critical point, which is that rfluxmtx figures out the hemisphere orientation(s) from the surface or surfaces you specify in your source and receiver files. So, it's very important that surface normals are directed in a certain way -- specifically, receiver normals point in the direction of the sampled hemisphere, and source normals point *away* from the hemisphere we want to sample.

Though this may seem counter-intuitive, it follows naturally from the typical orientation of windows in Radiance, whose normals face the interior. When the windows are acting as the receivers in a matrix calculation, such as for a view or illuminance value matrix, their normal is looking towards the interior hemisphere being sampled. When their role changes to make them the source surfaces, as when you calculate the exterior daylight matrix, they are still facing the correct direction to have their opposite-side hemisphere send out samples to the exterior.

In this way, rfluxmtx allows you to re-use your receiver description in one run as the source description in another run.

I'm hoping this makes sense, because I mention it in the man page, but it's not quite obvious and I did it the wrong way around the last time I used the tools, myself(!)

Best,
-Greg

rfluxmtx.pdf

···

From: "Gregory J. Ward" <[email protected]>
Subject: Re: [Radiance-general] Radiance 5.0 installation on ubuntu 14.04
Date: March 18, 2016 4:40:09 PM PDT

Hi Humberto,

The rfluxmtx command takes over for genklemsamp and calls rcontrib with the options needed for correct operation. You don't need to specify the bin or the .cal file(s), as rfluxmtx takes care of that part. You do need to set both the "u=Z" (or whatever) up-direction and the desired hemispherical sampling type (e.g., "h=r4") in your Radiance scene description handed to rfluxmtx as the source or receiver(s).

The reinhartb.cal file is more general than the original reinhart.cal file, and uses slightly different mechanisms (hence the need for a new name).

I hope this helps, and thanks to everyone for pitching in. I realize that the changes made post-tutorial have caused a good deal of confusion, but we hope to get it all sorted out in the end.

Cheers,
-Greg

From: Humberto Antunes <[email protected]>
Subject: Re: [Radiance-general] Radiance 5.0 installation on ubuntu 14.04
Date: March 18, 2016 2:49:55 PM PDT

Hello again everybody,

I have a question related to the command rfluxmtx.
While using rcontrib to generate the view matrix we had to set the window orientation with the command -b kbinS or (N,E or W), how do we do that in the rfluxmtx? Is it set by the "u" variable or it is not needed to do this with the rfluxmtx command?

Also, as reinhart.cal file been replaced for the reinhartb.cal?

Best regards
Humberto A.

13

Thanks Greg, that definitely helped.

So we really need to be careful about the surface normal, if the objective
is to get illuminance values on the work plane inside a room the normal of
the window should face the interior, is that it?

There is just something i cant't understand, defining the up-direction, in
the windows located in vertical surfaces the up-direction is allways "u=Z"
no matter the orientation of the window, while in the ground or sky surface
in the daylight matrix receiver file up-direction is "u=Y", right?

It may be import to say that my model is a simple office room with only a
window, a lot like Axel's model on the rcontrib lesson, and i will have to
analyse the illuminance at the work plane for the four different
orientations of the window with different control parameters for the
shading device applied to that same window.

Best regards
Humberto A.

···

2016-03-18 23:50 GMT+00:00 Greg Ward <[email protected]>:

Oh, but I did neglect to mention a critical point, which is that rfluxmtx
figures out the hemisphere orientation(s) from the surface or surfaces you
specify in your source and receiver files. So, it's very important that
surface normals are directed in a certain way -- specifically, receiver
normals point in the direction of the sampled hemisphere, and source
normals point *away* from the hemisphere we want to sample.

Though this may seem counter-intuitive, it follows naturally from the
typical orientation of windows in Radiance, whose normals face the
interior. When the windows are acting as the receivers in a matrix
calculation, such as for a view or illuminance value matrix, their normal
is looking towards the interior hemisphere being sampled. When their role
changes to make them the source surfaces, as when you calculate the
exterior daylight matrix, they are still facing the correct direction to
have their opposite-side hemisphere send out samples to the exterior.

In this way, rfluxmtx allows you to re-use your receiver description in
one run as the source description in another run.

I'm hoping this makes sense, because I mention it in the man page, but
it's not quite obvious and I did it the wrong way around the last time I
used the tools, myself(!)

Best,
-Greg

rfluxmtx.pdf
<http://www.radiance-online.org/learning/documentation/manual-pages/pdfs/rfluxmtx.pdf/at_download/file&gt;

*From: *"Gregory J. Ward" <[email protected]>

*Subject: *Re: [Radiance-general] Radiance 5.0 installation on ubuntu
14.04

*Date: *March 18, 2016 4:40:09 PM PDT

Hi Humberto,

The rfluxmtx command takes over for genklemsamp and calls rcontrib with
the options needed for correct operation. You don't need to specify the
bin or the .cal file(s), as rfluxmtx takes care of that part. You do need
to set both the "u=Z" (or whatever) up-direction and the desired
hemispherical sampling type (e.g., "h=r4") in your Radiance scene
description handed to rfluxmtx as the source or receiver(s).

The reinhartb.cal file is more general than the original reinhart.cal
file, and uses slightly different mechanisms (hence the need for a new
name).

I hope this helps, and thanks to everyone for pitching in. I realize that
the changes made post-tutorial have caused a good deal of confusion, but we
hope to get it all sorted out in the end.

Cheers,
-Greg

*From: *Humberto Antunes <[email protected]>

*Subject: *Re: [Radiance-general] Radiance 5.0 installation on ubuntu
14.04

*Date: *March 18, 2016 2:49:55 PM PDT

Hello again everybody,

I have a question related to the command rfluxmtx.
While using rcontrib to generate the view matrix we had to set the window
orientation with the command -b kbinS or (N,E or W), how do we do that in
the rfluxmtx? Is it set by the "u" variable or it is not needed to do this
with the rfluxmtx command?

Also, as reinhart.cal file been replaced for the reinhartb.cal?

Best regards
Humberto A.

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

14

Hi Humberto,

It does help to know a little context about your model. Yes, using "u=Z" for vertical windows in any orientation and "u=Y" for the sky or skylights is the standard approach. This is in fact the default behavior in rfluxmtx if no up-direction is specified, but it's best to be explicit about it. The "u=Y" for horizontal relates to the standard choice of Y for "north" in gensky and gendaylit. You can of course rotate the sky during your dctimestep phase, and this is much faster than rotating your surface orientations and re-running rfluxmtx.

In other words, I recommend running rfluxmtx once for the interior illuminances and their relation to your inward-facing window. Then, you will run it again to connect the still inward-facing window polygon as a source this time and its flux relationship to the ground and sky hemisphere. This brings up another important point, which is how best to specify the ground+sky in rfluxmtx. I discussed this durning my 2014 Workshop presentation on new Radiance features:

  Radiance Code Update 2014 http://www.radiance-online.org/community/workshops/2014-london/presentations/day1/Ward_WhatIsNew.pdf

The discussion of rfluxmtx starts at slide 38. Pay special attention to the example on slide 42, where I show how to combine a uniform ground hemisphere with a Reinhart sky description.

Once you have your two flux matrices, you will pipe the output of gensky or (more likely) gendaylit to xform -rz to rotate in the opposite direction of your model's rotation, then through genskyvec into dctimestep, where you will combine the matrices with the desired window glazing+shading BTDF to compute the irradiance values for your interior points. Alternatively, you may run gendaymtx for a whole year, using its built-in -rz option rather than xform.

Cheers,
-Greg

···

From: Humberto Antunes <[email protected]>
Subject: Re: [Radiance-general] Radiance 5.0 installation on ubuntu 14.04
Date: March 19, 2016 5:30:51 AM PDT

Thanks Greg, that definitely helped.

So we really need to be careful about the surface normal, if the objective is to get illuminance values on the work plane inside a room the normal of the window should face the interior, is that it?

There is just something i cant't understand, defining the up-direction, in the windows located in vertical surfaces the up-direction is allways "u=Z" no matter the orientation of the window, while in the ground or sky surface in the daylight matrix receiver file up-direction is "u=Y", right?

It may be import to say that my model is a simple office room with only a window, a lot like Axel's model on the rcontrib lesson, and i will have to analyse the illuminance at the work plane for the four different orientations of the window with different control parameters for the shading device applied to that same window.

Best regards
Humberto A.

2016-03-18 23:50 GMT+00:00 Greg Ward <[email protected]>:
Oh, but I did neglect to mention a critical point, which is that rfluxmtx figures out the hemisphere orientation(s) from the surface or surfaces you specify in your source and receiver files. So, it's very important that surface normals are directed in a certain way -- specifically, receiver normals point in the direction of the sampled hemisphere, and source normals point *away* from the hemisphere we want to sample.

Though this may seem counter-intuitive, it follows naturally from the typical orientation of windows in Radiance, whose normals face the interior. When the windows are acting as the receivers in a matrix calculation, such as for a view or illuminance value matrix, their normal is looking towards the interior hemisphere being sampled. When their role changes to make them the source surfaces, as when you calculate the exterior daylight matrix, they are still facing the correct direction to have their opposite-side hemisphere send out samples to the exterior.

In this way, rfluxmtx allows you to re-use your receiver description in one run as the source description in another run.

I'm hoping this makes sense, because I mention it in the man page, but it's not quite obvious and I did it the wrong way around the last time I used the tools, myself(!)

Best,
-Greg

rfluxmtx.pdf

From: "Gregory J. Ward" <[email protected]>
Subject: Re: [Radiance-general] Radiance 5.0 installation on ubuntu 14.04
Date: March 18, 2016 4:40:09 PM PDT

Hi Humberto,

The rfluxmtx command takes over for genklemsamp and calls rcontrib with the options needed for correct operation. You don't need to specify the bin or the .cal file(s), as rfluxmtx takes care of that part. You do need to set both the "u=Z" (or whatever) up-direction and the desired hemispherical sampling type (e.g., "h=r4") in your Radiance scene description handed to rfluxmtx as the source or receiver(s).

The reinhartb.cal file is more general than the original reinhart.cal file, and uses slightly different mechanisms (hence the need for a new name).

I hope this helps, and thanks to everyone for pitching in. I realize that the changes made post-tutorial have caused a good deal of confusion, but we hope to get it all sorted out in the end.

Cheers,
-Greg

From: Humberto Antunes <[email protected]>
Subject: Re: [Radiance-general] Radiance 5.0 installation on ubuntu 14.04
Date: March 18, 2016 2:49:55 PM PDT

Hello again everybody,

I have a question related to the command rfluxmtx.
While using rcontrib to generate the view matrix we had to set the window orientation with the command -b kbinS or (N,E or W), how do we do that in the rfluxmtx? Is it set by the "u" variable or it is not needed to do this with the rfluxmtx command?

Also, as reinhart.cal file been replaced for the reinhartb.cal?

Best regards
Humberto A.

15

Ok I get it now, you can't imagine how helpful all that info had been.

I had been running some tests and i saw something that is quite strange,
when i am creating my view matrix i use the following command "rfluxmtx -I+
-ab 12 -ad 50000 -lw 2e-5 -y 6 - < data/photocells.pts objects/window.rad
materials/testroom.mat objects/testroom.rad > results/photocells.vmx", if I
run the this command several times the output is always different, without
changing anything in the files or in the command, and this ends up creating
different work plane illuminances, do you have any idea on what is causing
this to happen?

Best regards
Humberto A.

16

Well, Radiance is a Monte Carlo simulation, which by default initializes to a new point each time. You can average together your results for greater accuracy, but how much difference are you seeing?

Cheers,
-Greg

···

From: Humberto Antunes <[email protected]>
Subject: Re: [Radiance-general] Radiance 5.0 installation on ubuntu 14.04
Date: March 19, 2016 1:47:06 PM PDT

Ok I get it now, you can't imagine how helpful all that info had been.

I had been running some tests and i saw something that is quite strange, when i am creating my view matrix i use the following command "rfluxmtx -I+ -ab 12 -ad 50000 -lw 2e-5 -y 6 - < data/photocells.pts objects/window.rad materials/testroom.mat objects/testroom.rad > results/photocells.vmx", if I run the this command several times the output is always different, without changing anything in the files or in the command, and this ends up creating different work plane illuminances, do you have any idea on what is causing this to happen?

Best regards
Humberto A.

17

I analyzed the data more carefully and the difference is not that big, I'm
going to follow your advice and I'm going to create a script to run the
simulation N times until the average of the final value starts to converge,
this way i think I will get a very accurate value, do you think this is a
good idea?

Best regards
Humberto A.

···

2016-03-20 1:30 GMT+00:00 Greg Ward <[email protected]>:

Well, Radiance is a Monte Carlo simulation, which by default initializes
to a new point each time. You can average together your results for
greater accuracy, but how much difference are you seeing?

Cheers,
-Greg

*From: *Humberto Antunes <[email protected]>

*Subject: *Re: [Radiance-general] Radiance 5.0 installation on ubuntu
14.04

*Date: *March 19, 2016 1:47:06 PM PDT

Ok I get it now, you can't imagine how helpful all that info had been.

I had been running some tests and i saw something that is quite strange,
when i am creating my view matrix i use the following command "rfluxmtx -I+
-ab 12 -ad 50000 -lw 2e-5 -y 6 - < data/photocells.pts objects/window.rad
materials/testroom.mat objects/testroom.rad > results/photocells.vmx", if I
run the this command several times the output is always different, without
changing anything in the files or in the command, and this ends up creating
different work plane illuminances, do you have any idea on what is causing
this to happen?

Best regards
Humberto A.

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

18

You can achieve the same result as averaging output by repeating the points on your input and adding a corresponding -c N for rcontrib.

If you already have the matrices, though, you can average them using rmtxop. Let's say you have 3 matrices, you can use:

  rmtxop -s .33333 res1.mtx + -s .33333 res2.mtx + -s .33333 res3.mtx > avg.mtx

Note that you would use a different scaling factor for more or fewer matrices.

Cheers,
-Greg

···

From: Humberto Antunes <[email protected]>
Subject: Re: [Radiance-general] Radiance 5.0 installation on ubuntu 14.04
Date: March 20, 2016 2:36:12 PM PDT

I analyzed the data more carefully and the difference is not that big, I'm going to follow your advice and I'm going to create a script to run the simulation N times until the average of the final value starts to converge, this way i think I will get a very accurate value, do you think this is a good idea?

Best regards
Humberto A.

2016-03-20 1:30 GMT+00:00 Greg Ward <[email protected]>:
Well, Radiance is a Monte Carlo simulation, which by default initializes to a new point each time. You can average together your results for greater accuracy, but how much difference are you seeing?

Cheers,
-Greg

From: Humberto Antunes <[email protected]>
Subject: Re: [Radiance-general] Radiance 5.0 installation on ubuntu 14.04
Date: March 19, 2016 1:47:06 PM PDT

Ok I get it now, you can't imagine how helpful all that info had been.

I had been running some tests and i saw something that is quite strange, when i am creating my view matrix i use the following command "rfluxmtx -I+ -ab 12 -ad 50000 -lw 2e-5 -y 6 - < data/photocells.pts objects/window.rad materials/testroom.mat objects/testroom.rad > results/photocells.vmx", if I run the this command several times the output is always different, without changing anything in the files or in the command, and this ends up creating different work plane illuminances, do you have any idea on what is causing this to happen?

Best regards
Humberto A.

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

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

19

That process seems to be much easier and more time efficient i will do it
like that.

Thank you very much for the help.

Best regards
Humberto A.

···

2016-03-20 21:57 GMT+00:00 Greg Ward <[email protected]>:

You can achieve the same result as averaging output by repeating the
points on your input and adding a corresponding -c N for rcontrib.

If you already have the matrices, though, you can average them using
rmtxop. Let's say you have 3 matrices, you can use:

rmtxop -s .33333 res1.mtx + -s .33333 res2.mtx + -s .33333 res3.mtx >
avg.mtx

Note that you would use a different scaling factor for more or fewer
matrices.

Cheers,
-Greg

*From: *Humberto Antunes <[email protected]>

*Subject: *Re: [Radiance-general] Radiance 5.0 installation on ubuntu
14.04

*Date: *March 20, 2016 2:36:12 PM PDT

I analyzed the data more carefully and the difference is not that big, I'm
going to follow your advice and I'm going to create a script to run the
simulation N times until the average of the final value starts to converge,
this way i think I will get a very accurate value, do you think this is a
good idea?

Best regards
Humberto A.

2016-03-20 1:30 GMT+00:00 Greg Ward <[email protected]>:

Well, Radiance is a Monte Carlo simulation, which by default initializes
to a new point each time. You can average together your results for
greater accuracy, but how much difference are you seeing?

Cheers,
-Greg

*From: *Humberto Antunes <[email protected]>

*Subject: *Re: [Radiance-general] Radiance 5.0 installation on ubuntu
14.04

*Date: *March 19, 2016 1:47:06 PM PDT

Ok I get it now, you can't imagine how helpful all that info had been.

I had been running some tests and i saw something that is quite strange,
when i am creating my view matrix i use the following command "rfluxmtx -I+
-ab 12 -ad 50000 -lw 2e-5 -y 6 - < data/photocells.pts objects/window.rad
materials/testroom.mat objects/testroom.rad > results/photocells.vmx", if I
run the this command several times the output is always different, without
changing anything in the files or in the command, and this ends up creating
different work plane illuminances, do you have any idea on what is causing
this to happen?

Best regards
Humberto A.

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

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

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