Welcome Desktop Radiance Users,

This is the discussion list for Desktop Radiance.  Desktop
Radiance is an AutoCAD plug-in software program which relies
upon the Radiance Synthetic Imaging System for its lighting
analyses.   Desktop Radiance is under development by Lawrence
Berkeley National Laboratory in collaboration with from Pacific
Gas and Electric company and Marinsoft, Inc.

We encourage you to use this site to enchange ideas, discuss
difficulties, and share your successes with other users of
Desktop Radiance.  For more information, please visit our web
site at:   http://radsite.lbl.gov/deskrad

Sincerely,
-Chas
Charles Ehrlich
Lawrence Berkeley National Laboratory
1 Cyclotron Road MS: 90-3111
Berkeley, CA  94720

I am trying to set up a camera that needs to be pitched.  I know how
to set up the heading angle, but is there a way to pitch the camera
in Desktop Radiance?

Thanks in advance,

Dan Katzenberger
Minneapolis, MN USA

Hi Dan,

There are several ways to manipulate camera pitch both inside
and outside of AutoCAD.

Inside AutoCAD, you can use the "DVIEW" command to adjust the
camera aiming.  Please see the AutoCAD help system for more
information on how DVIEW works...it is quite complex.  The
trick with this technique is that the modified view must be
saved "on top of" the older named view using the "DDVIEW"
command.  Select the view you have edited, click "new..." and
then specify a camera name that is identical to the camera you
just edited.  Click "Save View", then "Redefine" to overwrite
the previous definition of that camera.  The camera icon will
not update to reflect the new camera properties becuase it exists
indepently from the internal data structures which keep track
of AutoCAD views.  When you start the next simulation, be sure to
select the camera icon.

Outside of AutoCAD, you can use the the interactive rendering
program to manipulate camera settings.  There is a toolbar
slider called "pitch" which will adjust the pitch and immediately
redraw the new view.  The "Set view" dialog box can be reached
using the "Set-->view" menu command or by clicking on the "V"
toolbar icon. The "view direction" vector is what you will need
to edit.  This is by far the most difficult method.

One additional method also using the "Set view" dialog box is
to edit the "view shift" parameter.  Specify the decimal percent
above (positive) or below (negative) you want the view to be
oriented. This provides the very pleasing effect of looking up
without introducing severe 3-point perspective distortion.  This
is equivalent to using a slip-lens camera.  The vertical lines
remain vertical in the image.

Whichever way you work with the interactive rendering program,
to make the camera/view changes accessible in subsequent scenarios,
you need to use "File-->Append view to RIF file".  RIF files
are named after the scenario name.  A scenario which is duplicated
from the scenario you updated with a new view will show the
additional view(s) you created in the "Camera Simulaiton Setup"
dialog box in the "Camera" drop-down list.

We are currently working on a "Camera Properties" dialog box
which will bridge this gap between AutoCAD views and Desktop
Radiance cameras.  It will provide a real-time wireframe
representation of your view and cause the camera icon to update
itself appropriately.

-Chas

============


--- In desktopradiance@egroups.com, "Daniel Katzenberger" <dan@k...>
wrote:
> I am trying to set up a camera that needs to be pitched.  I know
> how to set up the heading angle, but is there a way to pitch the
> camera in Desktop Radiance?
>
> Thanks in advance,
>
> Dan Katzenberger
> Minneapolis, MN USA

hi,

The FAQ has been started,
http://radsite.lbl.gov/deskrad/  follow the help>FAQ link.

Ask questions and we'll try to answer them.

thanks,

judy

Hello,
I'm trying to model 3D curved surfaces.  Basically drawing an arc and
extruding it to make a 3D curved surface.  Can radiance do this?
Will
it eventually be able to do this? or do I have to physically break
the
surface up into a bunch of flat surfaces. Thanks in advance!

hi,

--- In desktopradiance@egroups.com, zrogers@a... wrote:
> Hello,
> I'm trying to model 3D curved surfaces.  Basically drawing an arc
and
> extruding it to make a 3D curved surface.  Can radiance do this?
> Will
> it eventually be able to do this? or do I have to physically break
> the
> surface up into a bunch of flat surfaces. Thanks in advance!

Not exactly.
An extruded arc is supported minimally and what you
end up with is 8 faceted surfaces (not a smooth curve),
no matter what kind of arc you have.

Eventually we plan on having a preferences box where
you can specify the refinement/number of facets for the arc.

For now, if you're modeling curves and you want more than 8 faces,
Here's the easiest solution I can think of:
1.  Change the surftab1 value to something high, say 16, by typing
at the prompt surftab1 and giving it a new value.
2.  Draw an arc.
3.  Draw a line in the direction you would like to extrude the arc.
If you want an arc to be extruded by 4 units in the Z direction,
draw a line from 0,0,0 to 0,0,4.
4.  click on the tabulated surface icon on the surface toolbar (the
3rd icon from the right).
5.  When prompted "select path curve", select the arc.
6.  When prompted "select direction vector" select the line you drew.

What you'll end up with is an arc made of 16 surfaces with height
of 4 units.  That's not too hard, right?
:-)

Hope this helps.

judy

Hi,

I'm just getting my system up and running and was wondering
which versions of AutoCAD will work with Desktop Radiance.
I know your website says AutoCAD 14, but I'm having trouble
finding anything other than AutoCAD 2000 to purchase - just
wondering if that will work.

Thanks for your help,
Melissa Koudelka
Yale University

I'm sorry you're having difficulty with Desktop Radiance.

The problems you're having a most likely due to the project setup.  We will be improving that in our next major release.

However, reinstalling the software should not be necessary.  You could try saving the drawing as a Version 12 AutoCAD file, then loading it back into a fresh AutoCAD R14 file.  This will often solve unkown rendering problems.

Also, avoid using space characters in your drawing names.

If you have a drawing that is "broken", please pkZIP it up and send it my way.

Thank you,
-Chas

===========

SVR4 nobody uid wrote:

---

Category: Bug Report
Name: Zack Rogers
UserEmail: zrogers@...
FPVersion: Version 1.0
OSVersion: Windows NT
Subject: Program stops simulating.
Date: 05 Jun 2000
Time: 12:45:42

I don't know exactly when it occurs so I can't list the steps, but several times the program has stopped simulating. In the simulation manager the scenario just quickly switches to "finished" without doing any calculation or rendering. It may be triggered from using 3d curved surfaces or having 2d lines but it has happened without these also. Either way, after this happens even the simplest boxes won't calculate. Last time the only was I could fix the problem so that I could calculate again was to completely uninstall and reinstall the program. Is there something known of that causes the simulations to stop running and is there a way to fix it without reinstalling the program?
 

The error you have encountered is probably due to a bug in the Reference Grid simulation.  Currently the software (inadvertently) only supports one reference grid in your drawing, even if only one is selected for simulation.  I have experienced this error when I use "all" to select the entire drawing contents just before Camera Simulation Setup.  Try selecting "all" then removing the reference grids which are not being used.

-Chas
 

---

FormName: Desktop Radiance Feedback Form
Category: Bug Report
Name: Zack Rogers
UserEmail: zrogers@...
FPVersion: Version 1.0
OSVersion: Windows NT
Subject: getting error - "Field 'stbdSceneFiles.Path' can't be a zero-length string"
Date: 06 Jun 2000
Time: 10:41:04

ErrorMessage

Field 'stbdSceneFiles.Path' can't be a zero-length string. Before this error it gave "Field is too small to accept the amount of data you attempted to add." Is there a limit to the size of the drawing? Both of these errors occured when starting a grid simulation.
 

Hello all,

I am wondering how the desk top radiance applies the BRTD materials that
are defined in the glazing material library to surfaces. I have tried a
few examples of using different autocad surface type but they all only
seem to have the glass material associated with them, not any BRTD
function material.
Does the glass window need to be a single surface or two surfaces (ie a
front and a back)?

Any pointers would be appreciated.

Regards,

Simon Crone.

hi Simon,

--- In desktopradiance@egroups.com, Simon Crone <crones@a...> wrote:
> Hello all,
>
> I am wondering how the desk top radiance applies the BRTD materials
that
> are defined in the glazing material library to surfaces. I have
tried a
> few examples of using different autocad surface type but they all
only
> seem to have the glass material associated with them, not any BRTD
> function material.
> Does the glass window need to be a single surface or two surfaces
(ie a
> front and a back)?
>
> Any pointers would be appreciated.
>
> Regards,
>
> Simon Crone.

In a typical glazing rafile, there is the regular glass, BRTD front,
and BRTD back definations.  You're right, currently only the
regular glass is simulated.  It's my understanding that the
BRTD information is included in the glazing radfile because we plan
to support it in later versions.

If you really want to use a BRTD now instead of glass, here's how..
After you start one simulation and all the AutoCad geometry has
been exported to radfiles, look for the glazing radfile.  It
should be named something like drawingname_g_glazingname_glass.rad.
Open that file up and where you see the modifier glazingname_glass,
replace it with glazingname_front (or glazingname_back).
So, if you had a drawing called conservatory.dwg and you used
clear3.rad, you would open conservatory_g_clear3_glass.rad, and
replace the clear3_glass with clear3_front (or clear3_back).

I think you can also get the same result by editing tables in the
database but I don't know how to explain that well.

judy

Hi Simon,

(Note: these instructions are intended for those Desktop Radiance users
not afraid to get their fingers dirty editing text files in the old
fashion Radiance for Unix way.)

The glazing industry defines the "front" surface of a glazing assembly
as the side which faces the outside of the building (toward the
Sun/sky).  This side is usually the more reflective side of the glazing
assembly.   There are many glazings where there will be no noticeable
difference between the inside and outside reflectance which may explain
why they appear the same in your simulation.

There is a way to model both the inside and outside surfaces with a
single surface.  The preferred method is with a BRTDfunc material using
the glazing.cal model.  This will be the way Desktop Radiance does it in
our next major release.  In the mean time, you might take a look at the
glazing.cal file and see if you can make sense of it.

In any case, it is never correct to use two surfaces to model a glazing
assembly with Radiance/Desktop Radiance.  Always use a single surface
whose surface normal is pointing into the building, in the direction of
energy flow, brightest to dimmest region.  The surface normal is used by
the glazing computations (mkillum) to determine the distribution of
daylight through your glass.

If you attempt to use glazing.cal and BRTDfunc to model your own
glazing, keep in mind that Radiance defines the "front" surface as the
side from which the surface normal exits according to the
right-hand-rule.  In the documentation for the glazing.cal file, you
will see fields for defining front and back reflectance...these are
relative to the surface normal.  You will want to use the properties
from the opposite side when defining the BRTDfunc material using
glazing.cal.

Good luck,
-Chas

-------- Original Message --------
Subject: Re: Modeling glass with BRTD material.
Date: Thu, 08 Jun 2000 08:58:43 +0800
From: Simon Crone <crones@...>
To: j lai <jlai@...>
References: <8hkebt+c2ai@eGroups.com>

Hi Judy,

Thanks for the prompt responce.

I have done what you suggested but I still dont get the difference
between the front and back material and how they should be applied to
what surface.

If I have a single polygon window pane that I am looking at from the
front I assume I use the glass_front material. If I now have a view
looking through the glass from behind, do i need to change the window
pane material to glass_back? Ie a seperate material definition depending
on my view. Or do I need two seperate window panes say 5mm apart?

Also what is the ettiquete for replying to the desktop radiance
discusion
list. Should I have  posted this responce there rather than directly to
you?

Regards

Simon.

Thanks for your five bug report messages this morning!  I do appreciate them being listed separately and they will probably appear in the e-mail list.

So now I see where you were seeing  Noon listed as 12 AM.  Thanks for clarifying that for me. I'll  add that bug to the hopper (and get an extra can of pesticide :-). Yes, the Simulation Manager lists the wrong time of day for the Noon hour.

As far as your MOLS are concerned, what you'll want to do is use a special type of glazing in the library called "special illum".  This glazing is a 100% transmissive (i.e., invisible) material that pre-computes the distribution of light coming from a complex fenestration system such as your MOLS's.  This will reduce the computation time for renderings of equal accuracy, but may introduce some artifacts.  Create a 3DFACE surface following the right hand rule such that the surface normal points into the room.  Attach the special illum material to it.  If there are other "panes" of glass in your system closer to the sky and sun, use the Tools-->Properties... dialog box and un-check the "as light source" box for these other glazing surfaces (keep "as light source" for the special illum).  This is necessary only if your MOLS's are "inside" your normal glazing surface.  If your MOLS's are closer to the sky/sun, then the standard glazing will take care of this automatically, i.e., no special illum is required.

As far as information on Radiance ray-tracing parameters...the online help system (Radiance-->help) has the complete Radiance reference manual.  Our User Manaual is also online and has a section dedicated to the interactive rendering program.  We will be adding help buttons to these dialog boxes in the near future.  You can also access the help system by double clicking on the "DRADHELP" file in the help directory where Desktop Radiance is installed.  You will also find the Quick-start tutorial and User Manual in PDF format in this location.

For final renderings, I encourage you to use batch mode renderings instead of the interactive renderer.  Batch mode renderings can be computed at resolutions independent of your screen size.  Also, you can eliminate "jaggies" by setting the "oversample ratio".  Both of these controls are on the "rendering" tab under the "Advanced" button of the simulation setup dialog boxes.

I'm glad Desktop Radiance is helping you solve your building related design needs.  And, thanks for sticking with the tool despite its shortcomings.

-Chas
Charles Ehrlich

==============================
 
 

Zack Rogers wrote:

Chas,

I see, it does render a noon time sky!  However, on the simulation manager it says that it is 12:00:00 AM even when it renders noon.  I always qeued my simulations and when it said the wrong time I never continued with them.  Thanks!

We are modeling MOLS(mini-optical lightshelves) for a daylighting system in an office building.  Since these surfaces are curved and specular and work off daylight Radiance is the only program capable of doing this(that I know of).  It is an awesome program and the calcs and renderings have turned out pretty neat! More information in the rview parameters box would be helpful in future additions. I guess there's plenty of technical information on what the parameter do but when your not familiar with how ray tracing works (which I'm not very familiar with) its hard to figure out which parameters cause what. Where could I find more information on how the ray-tracing works?  I've never worked with Radiance before so that explains my ignorance on these matters and I apologize.  I guess this is just a user-friendly issue.

Also these MOLS make the files huge. So we would like to eventually create a luminaire that models these MOLS for different days.  I know this is in the plan for upcoming additions and it would be nice if you could also create the 'look' of the luminaires too.

I have a few more questions,errors, suggestions but I'll send them separately in case you want to post them on the newsgroup.  Thanks again for your help!

Zack

Charles Ehrlich wrote:

 
Zach,

I can't seem to replicate this error.  When I set the time in the Camera Simulation Setup to 12:00:00 PM, it renders a noontime sky.

Is this just a matter of adjusting the time widget appropriately?  Fist click on the hour and use the arrows to adjust to 12, then click on the minutes and again use the arrows, then click on seconds...., lastly click on the "AM" and click the arrow to switch it to PM.

-Chas

---

FormName: Desktop Radiance Feedback Form
Category: Bug Report
Name: Zack Rogers
UserEmail: zrogers@...
FPVersion: Version 1.0
OSVersion: Windows NT
Subject: When the simulation time is set on 12:00pm (noon) it switches to midnight when the calculations run?  When you set it on midnight it doesn't switch to noon.  So I can't seem to get noon for a simulation time.
Date: 05 Jun 2000
Time: 08:25:42

-------- Original Message --------
From: Zack Rogers <zrogers@...>
Subject: Multiple north arrows
To: deskrad@...

Hello,

I was running a simulation with two north arrows and just switching
between them in the simulation setup-->advanced-->geometry tab.
However, the results for the two were the same(it didn't switch north
arrows).  I had to rotate the arrow and resimulate which takes much
longer than if you could just switch it in the simulation setup.

Zack Rogers


==========================

Zack,

Sounds like you've found another bug.  We'll address it as soon as
possible.

-Chas

Zack Rogers wrote:

> Hello,
>
> It seems you can't have grids directly on another surface. Is this
> correct? I had a grid on the ceiling and it gave me quadruple the light
> levels.  I lowered it an inch and it worked.  Along the same lines it
> would be nice in future versions to be able to specify the normal of the
> grid when you enter in. To do the ceiling or walls you have to 3d rotate
> with AutoCAD.  Thanks!
>
> Zack Rogers

Zack,

Currently, Desktop Radiance 1.0 only supports reference grid calculations on
a horizontal workplane with the sensors pointing "up".  A reference grid on
the ceiling will actually be pointing at the sky, not down toward the floor
as you might expect.

This will be addressed in our next major release (1.5), but will continue to
be a limitation in our upcoming sub-release, 1.01.

-Chas

Zack,

Unlike radiosity-based lighting analysis software, Radiance is much more
flexible about surface geometry, both in terms of surface normals and
overlapping surfaces.

Surface normal orientation does not matter for anything except glazings (if
they are treated "as light source") and for other light emitting surfaces
(luminaire luminous surfaces).   If you can attach a material to it, then
you can safely ignore the surface normal orientation.  You're also supposed
to be able to attach a material to a closed PLINE (without thickness) and
have it treated as an N-Sided polygon for Radiance, but this was
inadvertently disabled for the V1.0 build.  This makes floor plates much
easier to deal with.

In general you can ignore overlapping and intersecting surfaces.  One
important exception is the case of surfaces which protrude outside of the
building and are exposed to extremely high levels of daylight
illumination.  Sometimes this condition can cause bright colored "blotches"
to appear on the floor near the junction between inside and outside.  This
problem is easily solved by creating walls with an inside and outside
surface with the floor terminating somewhere in between the two faces,
preferably at the inside face itself.  But baseboard and door trim
overlapping with floor and wall surfaces will never cause a problem of the
inverse nature, i.e., I've never seen dark blotches.  This problem can
also be solved by increasing the accuracy parameters with the "Advanced"
dialog box.

-Chas

====================

Zack Rogers wrote:

> Hello,
>
> What if any problems will occur if your surfaces don't have normals?  I
> am wondering because it is much easier to build your models in the 3d
> solids and then explode into 3d surfaces.  However, when you explode
> these surfaces have no normals but you can still attach surfaces to
> them.  Does this cause any problems?  I know for windows a normal is
> needed but for surfaces it would save alot of time if you didn't need
> normals.  It just seems like I get better results (renderings) with
> 'normaled' surface.
>
> Along the same lines, does overlapping surfaces cause any problems.  ie
> having one large floor with columns and interior walls covering some of
> it up or putting say baseboards on the base of the walls?  I know some
> programs need to be fully 'closed' to work properly.  Thanks.
>
> Zack Rogers

Zack,

Again, thanks for your help.  I have responded to your bug reports and
feature suggestions in-line...

===
Zack Rogers wrote:

> Hello,
>
> Don't worry this is my last e-mail!  I've been working with this program
> all week so lots of stuff came up.
>
> I installed the program at home last night and on my first model the
> simulation wouldn't run. The same problem I wrote about before, I tried
> several times with 'fresh' AutoCAD14 drawings to no avail.  I do think
> that it has something to do with the project setup.  I'm not sure but it
> may have happened all three times when I didn't save the drawing and it
> made a 'Drawing' folder for the simulations.  Along the same lines
> changing the names of drawings and the resimulating sometimes lets you
> setup a new folder and sometimes doesn't and also doesn't just let you
> just open on old simulation folder etc..  I agree these are all problems
> with the way the projects are setup, where files get saved and
> referenced etc.

The project setup situation is abysmal.  But, we won't be able to fix the
problem until our next major release, V1.5.

>
> You are probably aware of this stuff(bugs and suggestions for future
> releases);
>
> The percent complete doesn't work in the simulation manager

Known problem.

>
> Scale of camera, reference grid, and reference point is way out of
> proportion with the north arrow.  If you want to see the north arrow the
> camera gets gigantic!

Known problem.

>
> Doesn't remember locations from one project to another.

Overlooked problem and duly noted.  Thank you.  I imagine that if the
software had more default locations, this would be less of an issue,
correct?

>
> It would be nice to adjust the suspension height on the luminaires.
> Also when you can create new luminaires it would be nice to be able to
> draw how you wanbt them to look.

Our luminaire editor will initially allow for custom distribution (IES)
files, and later will allow editing the geometry.

>
> Is there anyway to use the old Radiance functions like gensky in these
> versions.  I want to get sky and clouds outside my windows.

You can create your own sky by editing the sky.rad file, but be careful to
rename it because each time the simulation is started, the sky selected in
the simulation setup dialog is re-created over-writing the old sky file.

>
> More friendly interactive grid output.  I've been importing it from
> Editplus into excel.  Maybe a more graphical way to see the data points.

Known problem. The filename listed at the top of the display grid contains
the calculated values.

>
> This has to do with the size limit in batch mode question I sent
> earlier.  I also got the error "(40) scanline::scanline(): illegal
> length" when trying to do batch mode with a camera.

This is a new one.  Can you ZIP up the drawing and send it to me?  I need to
test this one further.

>
> I know this is a lot of *&%$ I sent you guys this morning and please
> don't feel any need to respond quickly or even at all.  I have enough of
> a grasp of it to continue with the work I'm doing with it.  Mostly I
> geuss its been stuff that might make the program more user friendly.

All of your suggestions and bug reports are very useful.  Again, thank you.

>
> Thanks for the help and good luck!
>
> Zack Rogers
> Architectural Energy Corporation
> Boulder, CO
> (303)444-4149

-Chas

"Steele, Kelly" wrote:

> Charles,
>
> Does the Radiance software use any weather data for the lighting
> calculations?
>
> Kelly Steele
>
> Hallam Engineering
> A Division of Hallam Associates, Inc.
> 60 Farrell Street, Suite 300
> South Burlington, VT 05403
> Tel:  802.658.4891
> Fax:  802.658.1457
> www.hallam.com

=====================

Hi Kelly,

Desktop Radiance 1.0 does not use weather data for the daylighting
calculations.

In an upcoming version, we plan to support "custom skies" which will
have the ability to tie the calculations to weather data.

If an annual summary of energy use is your primary need, then I suggest
that you take a look at the ADELINE software package:

http://radsite.lbl.gov/adeline

-Chas

Hello,

I know you can't create new luminaires yet but is there anyway I can
replace the IES files of an existing luminaire manually.

Also, it appears that the window distributions due to daylight are
coming from a point at the center of the window.  It seems that the
direct (sun) component comes through the window with the correct
distribution but that the diffuse (sky) component comes from a point
located at the center of the window.  The sky component still seems like
it has the correct distribution further away from the window but not up
against it because it is coming from a point source.  Am I
misinterpretting the calculations or is this how the windows are
modeled?

Additionally, when rendering a window you get 'scallops' on the adjacent
walls due to the window.  Why is this and can you avoid the scallops
somehow?

Thanks in advance,
Zack

Hi Zack,

It is true that the current version of Desktop Radiance does not support the creation or editing of new luminaires.  However, there is a document by Richard Mistrick which describes the process of replacing the IES distribution by manually editing the Radiance input file within Desktop Radiance.

    Desktop Radiance Overview

Unfortunately, we can not provide support for the methods described in the above document.

A luminaire editor is a high priority for our next major release.

-Chas
 

Zack Rogers wrote:

<snip>

Also, it appears that the window distributions due to daylight are
coming from a point at the center of the window.  It seems that the
direct (sun) component comes through the window with the correct
distribution but that the diffuse (sky) component comes from a point
located at the center of the window.  The sky component still seems like
it has the correct distribution further away from the window but not up
against it because it is coming from a point source.  Am I
misinterpretting the calculations or is this how the windows are
modeled?

Additionally, when rendering a window you get 'scallops' on the adjacent
walls due to the window.  Why is this and can you avoid the scallops
somehow?

Thanks in advance,
Zack

...regarding the other issues you raised in this e-mail...

Desktop Radiance glazings are created as virtual light sources.  This means that the light from those glazing surfaces that are treated "as light source" are pre-calculated (using one of the Radiance programs called mkillum) to determine the distribution of light coming through that surface.  This takes into consideration the exterior geometry and any fenestration details that may be present.  In other words, blinds with their slats oriented "up" will result in light being cast onto the ceiling more than the floor.

As a light source, the centroid of the glazing surface serves as the "origin" of the light from that surface.  To improve the accuracy of the sampling of the glazing surface, you can add "-ds .1" to the Advanced dialog box, rendering tab, override options field.  This will cause all light sources in your scene to be subdivided such that the longest dimension of the light source divided by distance between each point in the scene to the centroid of the light source is less this parameter.  The smaller the "-ds" parameter, the greater the subdivisions and the greater the computation time.  A similar effect can be achieved by changing the "Quality" parameter from medium to high in the rendering tab of the Advanced Dialog box of Camera Simulation Setup, but this will have other significant rending time implications.

The maximum number of subdivisions for any light source surface is 64 patches, i.e., 8 horizontal by 8 vertical patches.  This is the reason you will see "scallops" along the edges of your fenestration and on nearby walls.  The points along the surface that are closer to the centroid of the sub-divided patch will be brighter than the ones closer to the edges between the patches.

A partial but expensive (much longer rendering time) solution to this problem is to turn on "soft shadows" on the rendering tab of the Advanced dialog box.  This causes each patch of the glazing to be additionally sampled by an additional number of randomly distributed rays.

Sometimes, a less expensive solution to this problem is to turn off the "as light source" feature from the glazing properties dialog box for each glazing where this is a problem.  This will make it necessary to increase the number of ambient bounces for your simulation by one or two in the lighting tab of the advanced dialog box.  Use the command Radiance-->tools-->show properties..., then select the glazing surface.

Well...that's how it's supposed to work, but unfortunately, there is a bug in the program that prevents this feature from working.  Un-checking the "as light source" box and clicking apply does not in fact apply this change.

A work-around is to edit the .RIF file associated with the current scenario.  If the scenario is called "test2" then the .RIF file would be called "test2.rif".  It will look something like:

============== test2.rif ================
## The Desktop Radiance Input File (rif) for model: Drawing_glz.dwg
## Created: Wed Jun 14 15:29:04 2000
mkillum= -ab 1
materials= glz/lbnl/gry3.rad
illum= Drawing_glz_g_gry3_glass.rad
scene= sky.rad
#@northarrow= nonedefined xform -rz 0
ZONE= I 4.27268 9.32739 1.24856 6.30327 -0.01 0.01
QUALITY= M
UP= Z
INDIRECT= 1
# AMBFILE= Drawing_glz.amb
VARIABILITY= M
REPORT= .1667
render= -st .001 -dj .02
view= current -vf current.vf
================= end of test2.rif ===========

Move the contents of the "illum=" line (in boldface above) to the "scene=", or just change "illum=" to "scene=".   The resulting file will look like:

============ test2.rif  as edited ==============
## The Desktop Radiance Input File (rif) for model: Drawing_glz.dwg
## Created: Wed Jun 14 15:29:04 2000
mkillum= -ab 1
materials= glz/lbnl/gry3.rad
scene= Drawing_glz_g_gry3_glass.rad
scene= sky.rad
#@northarrow= nonedefined xform -rz 0
ZONE= I 4.27268 9.32739 1.24856 6.30327 -0.01 0.01
QUALITY= M
UP= Z
INDIRECT= 1
# AMBFILE= Drawing_glz.amb
VARIABILITY= M
REPORT= .1667
render= -st .001 -dj .02
view= current -vf current.vf
===============end of test2.rif as edited ============

Save this modified file as its original name, test2.rif.  Be sure it is saved as an ASCII Text file, not as a wordpad or other non-ASCII file.  You can then start this simulation from the Simulation Manager.

My apologies for this bug.  It will be fixed in an upcoming maintenance release.

-Chas
 
 
 
 

Zack Rogers wrote:

> Hello,
>
> Why does illuminance levels go down with additional ambient bounces
> considered?  For example,  on the comparison I did, the illuminance
> levels went from 920 lux for one ambient bounce to 820 for 9 ambient
> bounces.  Not a big difference, 10 fc, but it seems like the more
> ambient bounces you add that the illuminance should go up.  Thanks!
>
> -Zack

================

Zack,

You are probably getting this anti-intuitive behavior because the
ambient value is set too high.  The "ambient value" is used by the
Radiance rendering engine to approximate the average amount of light
present in the scene at all locations.  If the rendering is computed
with no ambient bounces, then the ambient value is the amount used in
all places of the scene that are in shadow.  The ambient value is
"added" to the very last bounce of ambient light.  As you progressively
add more ambient bounces of light, the potential contribution (and
therefore error) of the ambient value setting is reduced because this
term is multiplied by the surface reflectance of each intervening bounce
of light.  If the ambient value is set too low, the illuminance
calculation will converge upward to the correct value, and inversely
will converge downward if set too high.

To change the ambient value setting, go the "Advanced Calculation
Parameters", "Lighting" tab and input the value in the box next to
"Ambient Value".  The default ambient value assumed by Desktop Radiance
depends upon the settings for the Zone specified under the "Geometry"
tab.  If the zone is an interior zone, the default ambient value is
0.01.  If the zone is exterior, the ambient value is set to 10.0"  The
setting also depends upon the Exposure setting under the "Analysis"
tab.  If an exposure setting is specified, the ambient value will be
computed according to a formula that I don't have at my disposal.  If an
ambient value is specified explicitly, it will over-ride the computed
and default values.

There is a procedure for estimating the ambient value described in the
book "Rendering with Radiance" by Greg Ward Larson and Robert
Shakespeare.  This procedure or something similar will be incorporated
into a future version of Desktop Radiance.  Where the book talks about
the rview "exposure =" command to report the current exposure setting,
the Radiance for Windows version of rview would use the "Set-->Exposure"
command.  The Current Exposure is indicated at the top of the dialog
box.

-Chas

--- In desktopradiance@egroups.com, melissa.koudelka@y... wrote:
> Hi,
>
> I'm just getting my system up and running and was wondering
> which versions of AutoCAD will work with Desktop Radiance.
> I know your website says AutoCAD 14, but I'm having trouble
> finding anything other than AutoCAD 2000 to purchase - just
> wondering if that will work.
>
> Thanks for your help,
> Melissa Koudelka
> Yale University

Hi Melissa,

As of June 20, 2000, only AutoCAD R14.01 is compatible with
Desktop Radiance.  Problems have been reported with AutoCAD
MAP on top of AutoCAD R14 as well as with AutoCAD Mechanical
Desktop.  So, to be safe, I would say that these are also not
currently compatible.  No plans are in place to support MAP
or Mechanical Desktop.

Support for AutoCAD 2000 and Architectural Desktop are currently
in progress and will be available in an upcoming sub-version.

-Chas

I've drawn a box 99'x50' and 27' high with a window on the east wall
in order to simulate the daylight. The problem is that though in
Autocad the camera view seems right in RVIEW window the geometry
seems to be very distorted and you can't even make out the shape of
the room. I've tried several camera positions and several lens
lengths but with no luck so far. Do i do something wrong or is there
something that i miss in the simulation procedure?

Thanks,

Nikos

While it is reasonable to expect that Desktop Radiance could reproduce the on-screen AutoCAD representation of the view accurately, the software currently handles horizontal and vertical view sizes for "current" views (a view created by pressing return instead of explicitly selecting a camera) as if it is a parallel projection view.  This causes extremely distorted view parameters when the current view is supposed to be a perspective.  The horizontal and vertical view size are clamped to the maximum allowable value of 179 degrees.

The work around is to always explicitly select the camera even if the AutoCAD view on the screen is the one you want to render.  The "lens" of the camera should be clearly visible and selectable from within a perspective view of your model.

This problem will be solved in an upcoming release.  Our approach will be to ask the user for all "current" views if it should be a parallel projection or a perspective view.

Thanks for your feedback,
-Chas

=============
 

 
--------------------
FROM: varopoulos@...
DATE: Wed, 21 Jun 2000 07:46:46 -0000
SUBJECT: Rview geometry

I've drawn a box 99'x50' and 27' high with a window on the
east wall in order to simulate the daylight. The problem is
that though in Autocad the camera view seems right in RVIEW
window the geometry seems to be very distorted and you can't
even make out the shape of the room. I've tried several
camera positions and several lens lengths but with no luck
so far. Do i do something wrong or is there something that i
miss in the simulation procedure?

Thanks,

Nikos

Hello,

I can't seem to find the IESROT90 program that Rick Mistrick talks about
in the overview.  However, I used the same luminaire as he in his
example and it seemed to work correctly for both y-axis and x-axis
orientations without manually rotating the photometry.

-Zack

============

Zack,

Thanks for pointing out that the Overview document is out-of-date.  I will let all of you know when an updated version is available.

Rick and I figured out a better way to achieve the same result without using the IESROT90 program.  This requires that you put the string "-rz 90" in the luminaire ".fmt" file on the same line that has the string "${TWIST}" right after the "src_theta" argument.  You also have to add "2" to the number of arguments on that line.

Therefore, a luminaire that looks like:
 

void brightdata 10c38f4k_dist
10 flatcorr lum/philips/10c38f4k/10c38f4k.dat source.cal src_theta -t 0 0 -.241399 -rz ${TWIST}
0
1 ${DIM}
#Original ies2rad value: 56.5884

would become:

void brightdata 10c38f4k_dist
12 flatcorr lum/philips/10c38f4k/10c38f4k.dat source.cal src_theta -rz 90 -t 0 0 -.241399 -rz ${TWIST}
0
1 ${DIM}
#Original ies2rad value: 56.5884

Those of you confounded by what all of this means, rest assured, a luminaire editor is in progress.

-Chas
 

I recently installed Desktop Radiance to perform daylight calculations.  Is
there a way to put adjacent buildings next the the building that I want to
do calculations for?

Thanks,
Melissa Goren

The Quick-start Tutorial might give the impression that Radiance can only handle small, isolated room simulations, but in fact, one of the most powerful features of Desktop Radiance is that it can simulate just about anything that you model for it.

This means that you simply need to construct the neighboring buildings using any convenient 3D geometry and attach the appropriate materials to these surfaces.  If the ground is flat, then a circle with a large radius is a good way construct a ground plane to receive the building shadows.  When you start the simulation, select all of the applicable geometry you'd like to include in the rendering or other reference point calculations.

It occurs to me that some Desktop Radiance users may not be familiar with the heritage of the underlying Radiance rendering engine.  The gallery of pictures on the Radiance web site demonstrates the range of building projects that can be modeled and simulated with Desktop Radiance.

-Chas

Hi,

I wonder if there any place where I can find a complete list of all the
differences between Desktopradiance 1.0 and Radiance for Unix,
for example the fact that Desktopradiance support only one zone
and it has only one level of accuracy (medium)?

Thank you

Silvia

Dear Chas,

I would like to create a new location (the city that i live in Athens-
  Greece) in the "Add new location" window. How do i have to enter the
Latitude, Longitude, Altitude, Time zone and Stand. Meridian
Parameters? I guess some kind of special strings are needed.

Have a nice vacation!!

Best regards,

Nikos

hi Silvia,

Silvia Pezzana wrote:

> I wonder if there any place where I can find a complete list of all
the
> differences between Desktopradiance 1.0 and Radiance for Unix,
> for example the fact that Desktopradiance support only one zone
> and it has only one level of accuracy (medium)?

No, such a list does not exist, but I'll try to answer your question
here.

In Radiance for Unix, to start a simulation you should have
-the material and geometry files,
-the location/sky file,
-the view file(s), and
-the RIF file.
Within the RIF file is where you can specify the simulation
parameters such as zone, quality, accuracy, resolution, view(s), etc.

In Desktop Radiance, to start a simulation you need to have the
equivalents of those listed above.

-the materials and geometry files are associated together by using the
Radiance->Materials->attach materials and selecting the AutoCad
geometry.
-the location/sky file is created by using the pulldown menus on the
Camera Simulation setup box.  You can specify a new location and pick
a CIE sky from the list.
-the views you've created and saved using the camera icon in Desktop
Radiance are listed under the field "Camera Name".
-the RIF file is created when you are prompted to define a scenerio
name before starting a simulation.
If you don't want to take the default simulation parameters and want
to specify your own, you should click on the "Advanced" button near
the bottom of the Camera Simulation Setup dialog box.  In there you
have the option of changing the defaults.
For example, to overwrite "medium" accuracy, click the "rendering"
tab and select "high" for rendering quality.  Much of what you see
under the "Advanced" is what would be in the RIF file if you use
Radiance.

Hope this helps.

judy

You may have experienced problems using the AutoCAD "RENDER" command with Desktop Radiance drawings.  Two unrelated issues have been diagnosed.  This message should help you work around these problems.

Firstly, we have discovered an AutoCAD bug that causes problems with the Desktop Radiance CAMERA entity.  The symptom of this bug is that AutoCAD crashes without warning after typing the render command, then clicking the "render" button.  But, it only happens for the very first camera that you attach to the drawing, and only if it is the very first block of any type that has been inserted into the drawing..  If this camera is deleted, the bug goes away.  If more cameras are subsequently attached to the drawing, the bug does not reappear.  So, the way to work around this problem is to delete the first camera inserted into the drawing.  The fix for this problem will be included in our next maintenance release, but unfortunately will result in incompatibilities with drawings created in older versions of Desktop Radiance.

Secondly, we discovered that the render command creates empty block entities.  When a simulation is initiated, these empty blocks are erroneously determined to have valid Desktop Radiance attachments and are exported for the simulation.  To get around this issue, you can remove the reference to the "preprocess" batch file that creates the octrees and instances for these blocks.  To do this, prepare to start a simulation as usual, then click the "advanced" button on the "Camera Simulation Setup" dialog box.  Go to the "geometry"
tab and remove any text found in the box next to "Preprocessing file".  We will have a fix for this bug in our next maintenance release as well.

-Chas