Was wondering if others have noticed horizontal bands of lower irradiance values across the top 1/3 of the images produced from Radmap? These shifts look like strong shadow lines but don't relate to any occluding geometry in the scene. We seem to get these artifacts when using exterior views looking down onto the scene and using any number of weather files.
I can't say I've seen these before. It's also been a while since I've used radmap...
···
On Apr 28, 2009, at 1:50 PM, Chris Humann wrote:
Hello All,
Was wondering if others have noticed horizontal bands of lower irradiance values across the top 1/3 of the images produced from Radmap? These shifts look like strong shadow lines but don't relate to any occluding geometry in the scene. We seem to get these artifacts when using exterior views looking down onto the scene and using any number of weather files.
I have noticed that, actually I've noticed it with a custom script that
sums up CIE sunny skies, in a cumulative file, then renders irradiance
for a scene looking at direct only sun (ab 0). In my case I was looking
at an exterior aerial veiw of one building facade sited amongst other
buildings. The odd irradiance seems to occur on one of the neighboring
buildings that extends higher than all the others.
-Chris
···
________________________________
From: [email protected]
[mailto:[email protected]] On Behalf Of Chris
Humann
Sent: Tuesday, April 28, 2009 3:50 PM
To: Radiance general discussion
Subject: [Radiance-general] Re: Radmap irradiance shifts
Hello All,
Was wondering if others have noticed horizontal bands of lower
irradiance values across the top 1/3 of the images produced from Radmap?
These shifts look like strong shadow lines but don't relate to any
occluding geometry in the scene. We seem to get these artifacts when
using exterior views looking down onto the scene and using any number of
weather files.
____________________________________________________________
Electronic mail messages entering and leaving Arup business
systems are scanned for acceptability of content and viruses
I¹m going out on a limb here but I¹d suggest looking at dt and dc
parameters.
It could be that direct source testing stops earlier than you¹d like for
unobstructed surfaces. I expect that a scene with ~4000 suns should have
higher parameters for direct sampling anyway.
I have noticed that, actually I've noticed it with a custom script that sums
up CIE sunny skies, in a cumulative file, then renders irradiance for a scene
looking at direct only sun (ab 0). In my case I was looking at an exterior
aerial veiw of one building facade sited amongst other buildings. The odd
irradiance seems to occur on one of the neighboring buildings that extends
higher than all the others.
-Chris
From: [email protected]
[mailto:[email protected]] On Behalf Of Chris
Humann
Sent: Tuesday, April 28, 2009 3:50 PM
To: Radiance general discussion
Subject: [Radiance-general] Re: Radmap irradiance shifts
Hello All,
Was wondering if others have noticed horizontal bands of lower irradiance
values across the top 1/3 of the images produced from Radmap? These shifts
look like strong shadow lines but don't relate to any occluding geometry in
the scene. We seem to get these artifacts when using exterior views looking
down onto the scene and using any number of weather files.
____________________________________________________________
Electronic mail messages entering and leaving Arup business
systems are scanned for acceptability of content and viruses
This seems to make some sense. Anyone have a guess of whether it would
be more productive to set -dt 0 or -dc 1 in order to ensure no sources
are missed? In the man page those two settings seem to lead to avoidance
of estimations in the direct calculation.
-Chris
On Behalf Of Andrew McNeil
I'm going out on a limb here but I'd suggest looking at -dt and -dc
parameters.
It could be that direct source testing stops earlier than you'd like for
unobstructed surfaces. I expect that a scene with ~4000 suns should
have higher parameters for direct sampling anyway.
···
Sent: Wednesday, April 29, 2009 1:36 PM
____________________________________________________________
Electronic mail messages entering and leaving Arup business
systems are scanned for acceptability of content and viruses
This seems to make some sense. Anyone have a guess of whether it would
be more productive to set -dt 0 or -dc 1 in order to ensure no sources
are missed? In the man page those two settings seem to lead to avoidance
of estimations in the direct calculation.
Actually Francesco has always advised that -dt be set to 0 when using radmap skies because missing an entire sun could introduce a pretty large error. Maybe this is why I've never seen these artifacts before. Francesco?
Thank you all. Paying closer attention to the -dt and -dc parameters makes a lot of sense. Sure enough, setting -dt 0 and -dc 1 improves the results and fixes the artifact problem.
On Apr 29, 2009, at 10:36 AM, Andrew McNeil wrote:
I’m going out on a limb here but I’d suggest looking at –dt and –dc parameters.
It could be that direct source testing stops earlier than you’d like for unobstructed surfaces. I expect that a scene with ~4000 suns should have higher parameters for direct sampling anyway.
I have noticed that, actually I've noticed it with a custom script that sums up CIE sunny skies, in a cumulative file, then renders irradiance for a scene looking at direct only sun (ab 0). In my case I was looking at an exterior aerial veiw of one building facade sited amongst other buildings. The odd irradiance seems to occur on one of the neighboring buildings that extends higher than all the others.
-Chris
From: [email protected] [mailto:[email protected]] On Behalf Of Chris Humann
Sent: Tuesday, April 28, 2009 3:50 PM
To: Radiance general discussion
Subject: [Radiance-general] Re: Radmap irradiance shifts
Hello All,
Was wondering if others have noticed horizontal bands of lower irradiance values across the top 1/3 of the images produced from Radmap? These shifts look like strong shadow lines but don't relate to any occluding geometry in the scene. We seem to get these artifacts when using exterior views looking down onto the scene and using any number of weather files.
____________________________________________________________
Electronic mail messages entering and leaving Arup business
systems are scanned for acceptability of content and viruses