Mesh smoothing

radiance-general
1

Hi all

I have a question about smoothing. If you add the -s switch in gensurf you can get
a smooth shaded surface. Is there any way I can insert this smoothing to a surface
that I have imported from AutoCAD? I do not want to increase the polygon count but
i do want to smooth out the faceted glare on the surface.

Any ideas about this?

On a side note I am also looking at modelling light redirection systems (horizantal/vertical
glazing) and refractive systems. I have looked at some of the designs on www.schorsch.com
and attempted to model them using AutoCAD or a gen program then assigning a dielectric
material. Does any else have experience with creating these and would it be possible
to use a different modifier? I am looking for the most accurate calculation.

Thanks

Russell Maunder

BBSc (hons)
Victoria University of Wellington
School of Architecture and Design
New Zealand

···

__________________________________________________
Advertise with ZFREE - to find out more click below
http://www.zfree.co.nz/about-us/advert.html

2

Russell Maunder wrote:

I have a question about smoothing. If you add the -s switch in gensurf you
can get a smooth shaded surface. Is there any way I can insert this smoothing
to a surface that I have imported from AutoCAD? I do not want to increase the
polygon count but i do want to smooth out the faceted glare on the surface.

The "correct" solution would check if any normals are defined
already in Autocad/DXF, and use those to generate texfunc
modifiers similar to the ones gensurf creates from that
information. I have some unproven code to add this feature to
Radout and dxf2rad, but have never had the time to look into it
more deeply, in order to understand how and if it actually works.

Theoretically, you could also try to analyze the polygons in a
radiance scene file, find out which ones have common edges, and
if the difference between their normals is below a certain
margin, apply a texfunc to them, smoothing out that difference.
Not very elegant, but possible. Note that Rayfront removes
all modifiers when it imports geometry data, so that you will
have to edit the imported files in place.

If anyone can point me to really solid and understandable
information about how to work with the "barycentric coordinates"
that gensurf uses for this purpose, then I'll try to get it to
work with the next releases of Radout and dxf2rad.

On a side note I am also looking at modelling light redirection systems
(horizantal/vertical glazing) and refractive systems. I have looked at some
of the designs on www.schorsch.com and attempted to model them using AutoCAD
or a gen program then assigning a dielectric material. Does any else have
experience with creating these and would it be possible to use a different
modifier? I am looking for the most accurate calculation.

Those illustrations are on my site because I'm currently working
on a plug-in to Rayfront that will allow a full and correct
simulation of (most of) those systems. Simulating them with
Radiance alone is very exhaustive, and likely inaccurate in most
cases. Radiance is not designed to handle multiple mirror
reflections in tight spaces effectively. Of course, this doesn't
have to keep you from trying, just that you shouldn't put too
much trust into the results.

-schorsch

PS: Didn't forget your other mail...

···

--
Georg Mischler -- simulations developer -- schorsch at schorsch.com
+schorsch.com+ -- lighting design tools -- http://www.schorsch.com/

3

Hi Russel,

Hi all

I have a question about smoothing. If you add the -s switch in gensurf
you can get
a smooth shaded surface. Is there any way I can insert this smoothing
to a surface
that I have imported from AutoCAD? I do not want to increase the
polygon count but
i do want to smooth out the faceted glare on the surface.

Any ideas about this?

surface smooting is done by gensurf with a cal file that manipulates the
surface normal across
a generated polygon. The result is one flat polygon, with a surface
normal that varies across the polygon
(aka Phong shading). However, the true surface has to be known to the
generator to figure out
the intended normal at each point.
With an already tesselated surface that information is lost. One would
have to fit a smooth
(e.g. spline based) new surface to the vertices and recalculate normals
from that, but neither is
the surface what the original surface might have been, nor do I know a
program to do that
(infos, anyone ?).

On a side note I am also looking at modelling light redirection
systems (horizantal/vertical
glazing) and refractive systems. I have looked at some of the designs
on www.schorsch.com
and attempted to model them using AutoCAD or a gen program then
assigning a dielectric
material. Does any else have experience with creating these and would
it be possible
to use a different modifier? I am looking for the most accurate
calculation.

You may find that classic Radiance doesn't do forward raytracing and
hence won't handle light
redirection with dielectric volume materials. George Mischler, I think,
uses a method developped by
Thomas Schmidt ([email protected]), with does trick classic Radiance to do
this for certain type of
elements. Roland Schregle ([email protected]) is implementing a
photon-map add-on to Radiance
that handles the general case. That's currently in beta and validation
is pending. Thomas work has
been much more crosschecked, but is implicitly not as general.

-Peter

···

--
pab-opto, Freiburg, Germany, www.pab-opto.de

4

Russell,

We have successfully modelled the refractive light redirection system
known as the ‘laser cut panel’. It is simply a sheet of acrylic
with laser cuts in it. The cuts redirect incident illumination by
total internal reflection, giving the panel very powerful light
redirecting properties. The algorithm to model the device is based
on some simple geometrical calculations of directions of deflected and
undeflected rays, and relative intensities of the rays. The
modifier used is the prism2 material. For more info, check the
paper in Lighting Research and Technology:

Greenup, P.J., Edmonds, I.R. and Compagnon, R. (2000), Radiance
Algorithm to Simulate Laser-Cut Panel Light-Redirecting Elements,Lighting Research and Technology, 32 (2), 49-54.

I have also developed some algorithms to model reflecting light
redirecting devices, known as light guiding shade panels. The
process is:

model the flat reflective surfaces as polygons, the curved reflective
surfaces using gensurf

run oconv with the appropriate room and sky definitions

set a void output surface adjacent to the inside of the light redirecting
device

run mkillum with appropriately high settings of the -s and -d parameters
in particular

add the resulting geometry using oconv -i …

get your pictures, illuminances, etc as usual.

the results for each algorithm have been compared against measurements in
labs and in test buildings, and compare well. if you would like
more info, or the calculation file for the laser cut panels, let me know
and i will be happy to send it to you.

Regards,

Phil Greenup

On a side note I am also looking at modelling
light redirection systems (horizantal/vertical

glazing) and refractive systems. I have looked at some of the designs on
www.schorsch.com

and attempted to model them using AutoCAD or a gen program then assigning
a dielectric

material. Does any else have experience with creating these and would it
be possible

to use a different modifier? I am looking for the most accurate
calculation.

Thanks

Russell Maunder

BBSc (hons)

Victoria University of Wellington

School of Architecture and Design

New Zealand

Phil Greenup

Centre for Medical, Health and Environmental Physics

Queensland University of Technology

GPO Box 2434

Brisbane Qld 4001 Australia

Ph (07) 3864 5362

Fax (07) 3864 9079

Email [email protected]