Antimatter and windows

Zack_Rogers2 · September 9, 2009, 11:08pm

Hello,

I am working on a geometry conversion project and need to "carve" out a
hole in my wall(s) for any window(s). In lieu of actually breaking the
wall into smaller components or creating it as a large polygon wrapping
around windows I am considering using antimatter.

The 4 links below represent a building with window openings (no glass)
using antimatter to carve the holes. In these renderings I had a
polygon that defines the overall wall and then a co-planar polygon that
defines the opening. Two of the images were generated with rpict and
two with rvu. In rvu, depending on the view angle, portions of the
opening disappear. In rpict, all openings seem to be correct in size
but it seems to have a hard time determing which material to use (wall
or antimatter) creating a speckled frit look. Given that these two
surfaces do overlap, the rpict renderings make some sense as it is
similar to what I have seen with other overlapping surfaces.

My questions regarding antimatter:

1) Is the use antimatter validated? and for what, rtrace, rview, rpict?
I seem to remember the are certain situations to avoid with antimatter.
Are there alot? Should I just avoid it altogether?
2) How do you apply antimatter? The documentation I could find
discusses using antimatter applied to a "volume", and seems to indicate
that any other object in that "volume" with the modifier specified in
the antimatter definition will be invisible. How do you apply it to a
volume? From what I understand, sphere and cylinder may be the only
objects with volume in radiance. I tried applying it to a genbox that
went through my wall but that is basically applying it to 6 polygons and
I got no hole in my wall. I am worried about relying on surfaces having
to be co-planar because for any angled (off axis) planes it will be hard
to ensure the two surfaces are planar given rounding differences.
3) A hole seemed to be carved in my wall whether or not the modifier in
my antimatter definition matched my wall. From the definition, i
gathered that the modifiers listed are the only ones that will become
invisible when inside an antimatter volume, but that seems to only
sometimes be the case. Am I misunderstanding the modifier list?
4) If the rest of this is valid, what now if I put a window polygon
directly on top of the antimatter polygon? The renderings I got seemed
to take this ok. The glass modifier was not listed in the antimatter
list and it seemed to still be there intact.

Overall, I am looking for a go or no-go on this approach to carving out
holes for windows. If there are too many problems with this approach, I
will fall back on the much more time consuming process of writing an
algorithm that will create a polygon wrapped around windows - anyone
have one of those written in python?

In radiance terms, here is an example of what I am trying to do for a
single wall with window. Notice, the antimatter definition lists
grey_0.00 however it seems to do a fine job punching a whole through my
grey_0.30 wall but not my trans_0.30 window.

void antimatter opening
1 grey_0.00
0
0

void plastic grey_0.30
0
0
5 0.3 0.3 0.3
0 0

void trans trans_0.30
0
0
7 0.947 0.947 0.947 0.05 0.02 0.333 0.00

grey_0.30 polygon wall
0
0
12 0 0 0
    20 0 0
    20 0 10
    0 0 10

opening polygon hole
0
0
12 3 0 3
    17 0 3
    17 0 9
    3 0 9

trans_0.30 polygon window
0
0
12 3 0 3
    17 0 3
    17 0 9
    3 0 9

Thanks for the help!
Zack

···

________________________________

ZACK ROGERS PE, LEED (r) AP

INTEGRATED DESIGN ASSOCIATES, INC
437 Main Street
Longmont, CO 80501
tel: 303.848.8299, fax: 303.848.8290
www.ideasi.com <http://www.ideasi.com/>

Lars_Grobe · September 10, 2009, 12:19am

Hi Zack,

the problem you experience appears any time when you have coplanar
surfaces with different modifiers. A ray hitting such a surface sees
either material, and as they are coplanar this results in a random noise
/ mixture of both.

To use antimatter, place a box in you wall, so that the front side of it
is in front of the wall (you would see it from your position) while the
back one is behind. Make sure that there is no other geometry in the
volume of this box.

Now what happens is that a ray hitting the antimatter material with the
surface facing it will become blind, go on traveling the same direction
(or modify it according to what you give in the second line of the
antimatter modifier's definition), and start seeing geomertry only once
he passed thru the back of an antimatter surface again. In other words,
it is like if the ray jumps between the antimatter surfaces ignoring
everything in between.

Hope this is correct, but that was the result of looking at that part of
the code once

Cheers, Lars.

Zack_Rogers2 · September 10, 2009, 4:29am

Thanks for the insight Lars,

I had originally tried the box approach and moved onto the co-planar
surface approach as I felt like I was getting further.

Back to the box approach, I've been testing this out now on just a very
simple wall model and am seeing strange behavior. From all viewing
directions, the shadow from my hole in the wall looks correct. However,
when viewing the wall from one side it seems like I am seeing the inside
surfaces of my antimatter box as grey_0.60 but not the front surface.
At some view angles from the front side, the side of my antimatter box
disappear and I do get a glimpse through the wall. From the other side
(shadow side), everything looks as I would expect.

http://docs.google.com/Doc?docid=0AcMF73k-1deIZGdkNmhja3BfOTdnejRjbTloaw
&hl=en
http://docs.google.com/Doc?docid=0AcMF73k-1deIZGdkNmhja3BfOTVjamR4YnZkMw
&hl=en
http://docs.google.com/Doc?docid=0AcMF73k-1deIZGdkNmhja3BfOTNmOHF6N2docA
&hl=en

The box was created with genbox and I double checked that all surface
normals are pointing outward. I also tested other modifiers in my
antimatter definition and "void" and anything other than the wall
material did not do anything (disregard my earlier statement/question
about materials).

So several of my questions have been answered but there is still the
obvious artifacts in the rendering and I am still wondering about the
validity of this approach using rtrace. Will it see random errors?

The file I am running for this test is:

void plastic grey_0.25
0
0
5 0.25 0.25 0.25
0 0

void plastic grey_0.60
0
0
5 0.6 0.6 0.6
0 0

void antimatter opening
1 grey_0.60
0
0

grey_0.60 polygon Space.10_Wall.38
0
0
12
    0.00 0.00 10.00
    0.00 0.00 0.00
    30.00 0.00 0.00
    30.00 0.00 10.00

!genbox opening hole 10 1 6 | xform -t 10 -.5 2

grey_0.25 polygon ground
0
0
12 -100 -100 -0.5
    100 -100 -0.5
    100 100 -0.5
    -100 100 -0.5

!gensky 3 20 10.5 -a 42.8326 -o 106.329 -m 105.0 +s | xform -rz 0.0

skyfunc glow skyglow
0
0
4 0.9 0.9 1.0 0

skyglow source sky
0
0
4 0.0 0.0 1.0 180

skyfunc glow groundglow
0
0
4 1.0 0.8 0.5 0

groundglow source ground
0
0
4 0.0 0.0 -1.0 180

···

________________________________

ZACK ROGERS, P.E., LEED AP, IESNA

INTEGRATED DESIGN ASSOCIATES, INC
437 Main Street
Longmont, CO 80501
tel: 303.848.8299, fax: 303.848.8290
www.ideasi.com

-----Original Message-----
From: [email protected]
[mailto:[email protected]] On Behalf Of Lars
O. Grobe
Sent: Wednesday, September 09, 2009 6:20 PM
To: Radiance general discussion
Subject: Re: [Radiance-general] Antimatter and windows

Hi Zack,

the problem you experience appears any time when you have coplanar
surfaces with different modifiers. A ray hitting such a surface sees
either material, and as they are coplanar this results in a random noise
/ mixture of both.

To use antimatter, place a box in you wall, so that the front side of it
is in front of the wall (you would see it from your position) while the
back one is behind. Make sure that there is no other geometry in the
volume of this box.

Now what happens is that a ray hitting the antimatter material with the
surface facing it will become blind, go on traveling the same direction
(or modify it according to what you give in the second line of the
antimatter modifier's definition), and start seeing geomertry only once
he passed thru the back of an antimatter surface again. In other words,
it is like if the ray jumps between the antimatter surfaces ignoring
everything in between.

Hope this is correct, but that was the result of looking at that part of
the code once

Cheers, Lars.

Lars_O_Grobe1 · September 10, 2009, 4:40am

Hi Zack!

void antimatter opening
1 grey_0.60
0
0

This will make the backface of your antimatter appear grey.

void antimatter opening
1 void
0
0

Will have invisible backside faces, I think this is what you want.

Cheers, Lars.

Zack_Rogers2 · September 10, 2009, 5:12am

Hi Lars, I tried "void" as the modifier and did not get any hole in my
wall from either side. Also, if the first definition makes the backside
of the surfaces grey, then it would have been seen that way from both
sides of the wall. The shadow side looks correct. It seems like
inconsistent behaviour.

Zack

···

________________________________

ZACK ROGERS, P.E., LEED AP, IESNA

INTEGRATED DESIGN ASSOCIATES, INC
437 Main Street
Longmont, CO 80501
tel: 303.848.8299, fax: 303.848.8290
www.ideasi.com

-----Original Message-----
From: [email protected]
[mailto:[email protected]] On Behalf Of Lars
O. Grobe
Sent: Wednesday, September 09, 2009 10:40 PM
To: Radiance general discussion
Subject: RE: [Radiance-general] Antimatter and windows

Hi Zack!

void antimatter opening
1 grey_0.60
0
0

This will make the backface of your antimatter appear grey.

void antimatter opening
1 void
0
0

Will have invisible backside faces, I think this is what you want.

Cheers, Lars.

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

Zack_Rogers2 · September 10, 2009, 5:39am

I think I figured it out. If I use a box for the wall instead of just a
plane, the antimatter box seems to work correctly from all sides. Here
are pictures of my working wall, with glass now placed in the opening:

http://docs.google.com/Doc?docid=0AcMF73k-1deIZGdkNmhja3BfMTAxYzIydGd2aD
c&hl=en
http://docs.google.com/Doc?docid=0AcMF73k-1deIZGdkNmhja3BfOTlnNDRmcThnYw
&hl=en

And here is the file that generated it:

void plastic grey_0.25
0
0
5 0.25 0.25 0.25
0 0

void plastic grey_0.60
0
0
5 0.6 0.6 0.6
0 0

void antimatter opening
1 grey_0.60
0
0

void glass glass_050
0
0
3 0.5 0.5 0.5

!genbox grey_0.60 wall 30 .8 10 | xform -t 0 -.4 0

#grey_0.60 polygon Space.10_Wall.38
#0
#0
#12
# 0.00 0.00 10.00
# 0.00 0.00 0.00
# 30.00 0.00 0.00
# 30.00 0.00 10.00

!genbox opening hole 10 1 6 | xform -t 10 -.5 2

glass_050 polygon window
0
0
12 10 0 2
  10 0 8
  20 0 8
  20 0 2

grey_0.25 polygon ground
0
0
12 -100 -100 -0.5
  100 -100 -0.5
  100 100 -0.5
  -100 100 -0.5

!gensky 3 20 10.5 -a 42.8326 -o 106.329 -m 105.0 +s | xform -rz 0.0

skyfunc glow skyglow
0
0
4 0.9 0.9 1.0 0

skyglow source sky
0
0
4 0.0 0.0 1.0 180

skyfunc glow groundglow
0
0
4 1.0 0.8 0.5 0

groundglow source ground
0
0
4 0.0 0.0 -1.0 180

···

________________________________

ZACK ROGERS, P.E., LEED AP, IESNA

INTEGRATED DESIGN ASSOCIATES, INC
437 Main Street
Longmont, CO 80501
tel: 303.848.8299, fax: 303.848.8290
www.ideasi.com