problem generation dmx matrix

Dear expert,

A brief introduction to the project:
I'm appling 3 phase method on a facede that is composed of a big cfs on the
bottom of which there are several openings (glass doorway) that are the
access to the internal area.

Therefore, I have decided to separet the facede in two groups (CFS &
GLASS_DOOR) since the BSDFs are different (I hope that it is the right way
to proceed).

But when I have generated the dmx matrix for both groups of windows, that
have the same orientation (-vd -0.98 -0.17 0), a warning message have
appeared only for the GLASS_DOOR group:
*view direction does not correspond to any surfaces.*

Controls:
1. checked geometry with objline and rvu (for the glow material) and all
seems work well.
2. checked oconv that contain material, scene without the window ,
skywhtie.
3. checked the command for the glass_door (that is the same of the
functioning one save for the geometry):
genklemsamp -c 1000 - vd 0.98 -0.17 0 materials.rad glass_door.rad |
rcontrib -n 4 -c 1000 -e MF:1 -f reinhart.cal -b rbin - bn Nrbins -m
sky_glow - faf temp/model_dmx.oct > data/glass_door.dmx

I don't know where find the solution of the the problem. Any advice?

Thanks in advance,
Giuseppe

Hi Giuseppe,

Thanks for sending your input files. There's nothing you are really doing wrong, but you've run up against a limitation in genklemsamp to find surfaces that are widely separated. Your windows are nice and close-together, so they don't pose a problem, but your doors are small relative to the space between them, such that the sampling process, which is based on the bounding cube, cannot locate them even with 1000 initial samples (your -c setting). The solution is to either run genklemsamp on each door individually, or to increase the sampling rate to -c 10000, which works in your case.

Best,
-Greg

···

From: Giuseppe De Michele <[email protected]>
Subject: [Radiance-general] problem generation dmx matrix
Date: July 13, 2014 4:39:02 AM PDT

Dear expert,

A brief introduction to the project:
I'm appling 3 phase method on a facede that is composed of a big cfs on the bottom of which there are several openings (glass doorway) that are the access to the internal area.

Therefore, I have decided to separet the facede in two groups (CFS & GLASS_DOOR) since the BSDFs are different (I hope that it is the right way to proceed).

But when I have generated the dmx matrix for both groups of windows, that have the same orientation (-vd -0.98 -0.17 0), a warning message have appeared only for the GLASS_DOOR group:
view direction does not correspond to any surfaces.

Controls:
1. checked geometry with objline and rvu (for the glow material) and all seems work well.
2. checked oconv that contain material, scene without the window , skywhtie.
3. checked the command for the glass_door (that is the same of the functioning one save for the geometry):
genklemsamp -c 1000 - vd 0.98 -0.17 0 materials.rad glass_door.rad | rcontrib -n 4 -c 1000 -e MF:1 -f reinhart.cal -b rbin - bn Nrbins -m sky_glow - faf temp/model_dmx.oct > data/glass_door.dmx

I don't know where find the solution of the the problem. Any advice?

Thanks in advance,
Giuseppe

Greg,

the dmx works well now (I have tried only the increasing of rays number for
now), so thanks again.

Cheers,
Giuseppe.

···

2014-07-15 18:59 GMT+02:00 Greg Ward <[email protected]>:

Hi Giuseppe,

Thanks for sending your input files. There's nothing you are really doing
wrong, but you've run up against a limitation in genklemsamp to find
surfaces that are widely separated. Your windows are nice and
close-together, so they don't pose a problem, but your doors are small
relative to the space between them, such that the sampling process, which
is based on the bounding cube, cannot locate them even with 1000 initial
samples (your -c setting). The solution is to either run genklemsamp on
each door individually, or to increase the sampling rate to -c 10000, which
works in your case.

Best,
-Greg

*From: *Giuseppe De Michele <[email protected]>

*Subject: *[Radiance-general] problem generation dmx matrix

*Date: *July 13, 2014 4:39:02 AM PDT

Dear expert,

A brief introduction to the project:
I'm appling 3 phase method on a facede that is composed of a big cfs on
the bottom of which there are several openings (glass doorway) that are the
access to the internal area.

Therefore, I have decided to separet the facede in two groups (CFS &
GLASS_DOOR) since the BSDFs are different (I hope that it is the right way
to proceed).

But when I have generated the dmx matrix for both groups of windows, that
have the same orientation (-vd -0.98 -0.17 0), a warning message have
appeared only for the GLASS_DOOR group:
*view direction does not correspond to any surfaces.*

Controls:
1. checked geometry with objline and rvu (for the glow material) and all
seems work well.
2. checked oconv that contain material, scene without the window ,
skywhtie.
3. checked the command for the glass_door (that is the same of the
functioning one save for the geometry):
genklemsamp -c 1000 - vd 0.98 -0.17 0 materials.rad glass_door.rad |
rcontrib -n 4 -c 1000 -e MF:1 -f reinhart.cal -b rbin - bn Nrbins -m
sky_glow - faf temp/model_dmx.oct > data/glass_door.dmx

I don't know where find the solution of the the problem. Any advice?

Thanks in advance,
Giuseppe

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

--
*Giuseppe De Michele*

By the way, I am working on a more general tool that will replace genklemsamp and run rcontrib for you, which should eliminate this issue. Hoping to have it done by the end of next month, maybe sooner.

-Greg

···

From: Giuseppe De Michele <[email protected]>
Subject: Re: [Radiance-general] problem generation dmx matrix
Date: July 18, 2014 12:14:04 AM PDT

Greg,

the dmx works well now (I have tried only the increasing of rays number for now), so thanks again.

Cheers,
Giuseppe.

2014-07-15 18:59 GMT+02:00 Greg Ward <[email protected]>:
Hi Giuseppe,

Thanks for sending your input files. There's nothing you are really doing wrong, but you've run up against a limitation in genklemsamp to find surfaces that are widely separated. Your windows are nice and close-together, so they don't pose a problem, but your doors are small relative to the space between them, such that the sampling process, which is based on the bounding cube, cannot locate them even with 1000 initial samples (your -c setting). The solution is to either run genklemsamp on each door individually, or to increase the sampling rate to -c 10000, which works in your case.

Best,
-Greg

From: Giuseppe De Michele <[email protected]>
Subject: [Radiance-general] problem generation dmx matrix
Date: July 13, 2014 4:39:02 AM PDT

Dear expert,

A brief introduction to the project:
I'm appling 3 phase method on a facede that is composed of a big cfs on the bottom of which there are several openings (glass doorway) that are the access to the internal area.

Therefore, I have decided to separet the facede in two groups (CFS & GLASS_DOOR) since the BSDFs are different (I hope that it is the right way to proceed).

But when I have generated the dmx matrix for both groups of windows, that have the same orientation (-vd -0.98 -0.17 0), a warning message have appeared only for the GLASS_DOOR group:
view direction does not correspond to any surfaces.

Controls:
1. checked geometry with objline and rvu (for the glow material) and all seems work well.
2. checked oconv that contain material, scene without the window , skywhtie.
3. checked the command for the glass_door (that is the same of the functioning one save for the geometry):
genklemsamp -c 1000 - vd 0.98 -0.17 0 materials.rad glass_door.rad | rcontrib -n 4 -c 1000 -e MF:1 -f reinhart.cal -b rbin - bn Nrbins -m sky_glow - faf temp/model_dmx.oct > data/glass_door.dmx

I don't know where find the solution of the the problem. Any advice?

Thanks in advance,
Giuseppe

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

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