Hi,
I have simple question: How could I determine illuminance distribution on a workplane?
My scene is simple:
1.Rectangular room, with ideally diffuse surfaces. Reflectances are 0.8 0.8 0.8.
2.Room dimensions are 10x10x10 and the center is positioned in the (0,0,0).
3.In the middle of the room is placed a luminaire, with indirect light distribution. It is oriented toward ceiling, so workplane illuminance has only indirect component.
I found somewhere in the mail archives a command for illuminance calculation for a grid of points. For my testing room the command is:
cnt 6 6| rcalc -e '$1=$1*2-5; $2=$2*2-5; $3=-4; $4=0; $5=0; $6=1'| rtrace -h -oov -I -ab 2 room.oct| rcalc -e '$1=$1; $2=$2; $3=$3; $4=47.4*$4+120*$5+11.6*$6'>illuminance.txt
Which value is the most appropriate for -ab parameter? Should I also use some other ambient parameters?
I've tried with different -ab values and results are ascending, so I don't know which result is the most accurate.
Purpose of this calculations is to compare values from Radiance with values obtained form "Infinite Plane-Exitance Difference Method" approved from IESNA for calculation of indirect illuminance.
The answer to the first question is, whatever value you deem appropriate. The answer to the second one is, probably.
The raw truth of using Radiance is that there is no single correct answer for parameterizing your simulations. The proper settings for the myriad options will differ with every single simulation you do, and depend on... pretty much everything. The good news is that the rad utility can be put to good use as you get started with Radiance. rad makes good guesses about proper parameters for your simulations based on a few basic assumptions about your scene. You can even save those assumptions to a file and reuse them in subsequent rpict and rtrace runs. Refer to the manpage of rpict for more onfo on that. There are also some good discussions of this on this list's archives (I should know, since I learned about this from this very list).
···
On Apr 22, 2005, at 4:32 AM, Marija wrote:
My scene is simple:
1.Rectangular room, with ideally diffuse surfaces. Reflectances are 0.8 0.8 0.8.
2.Room dimensions are 10x10x10 and the center is positioned in the (0,0,0).
3.In the middle of the room is placed a luminaire, with indirect light distribution. It is oriented toward ceiling, so workplane illuminance has only indirect component.
Which value is the most appropriate for -ab parameter? Should I also use some other ambient parameters?
I've tried with different -ab values and results are ascending, so I don't know which result is the most accurate.
I am simulating to find illuminance values for different angles of Venetian blinds, Using Mkillum.. Do we need to consider Blinds as part of the Luminous plane generated by mkillum..? We are getting better results when using just a plane polygon as Luminous secondary source as compared to when we include Blinds in the fenestration system as luminous plane. ( comparing two different systems of blinds of similar angles and characteristics.)
I am simulating to find illuminance values for different angles of
Venetian blinds, Using Mkillum.. Do we need to consider Blinds as part
of the Luminous plane generated by mkillum..? We are getting better
results when using just a plane polygon as Luminous secondary source
as compared to when we include Blinds in the fenestration system as
luminous plane. ( comparing two different systems of blinds of similar
angles and characteristics.)
If I understand you right - are you asking wether to replace the each
blind's surface by a mkillum-generated illum??? This makes no sense -
you have aluminium blinds, and no light will pass through a blind. This
must lead to really funny results.
Mkillum has been written to precalculate the effect of compex
fenestration such as blinds. You calculate a light source on the inside
of the complex fenestration using mkillum, so that during your (rpict or
rtrace) rendering, the light source for ambient calculations is not
obstructed. The reasons for this are explained in detail in the book
"Rendering with Radiance".
In your case there are only two reasonable options as far as I
understood so far. Either render without mkillum, but very(!!!) high
quality settings for ad, as, aa, ab (which will take a long time), or
seal your complete fenestration system with mkillum-generated illum
surface(s) as explained in the book.
Mkillum has been written to precalculate the effect of compex
fenestration such as blinds. You calculate a light source on the inside
of the complex fenestration using mkillum, so that during your (rpict or
rtrace) rendering, the light source for ambient calculations is not
obstructed. The reasons for this are explained in detail in the book
"Rendering with Radiance".
Lars is correct; as Greg Ward says, you want the virtual light source positioned just inside of "where it gets nasty". You're basically taking the backwards ray tracing operation away from the view plane and placing it right up against the blinds to help guarantee that plenty of rays get through the "nastiness" and find those bright light sources out there. The result of that calculation is then saved as a .dat file and used on that virtual light source plane in the regular calculation, giving much better results in less time (opposed to relying on brute force with the ambient calculation). As Lars said, there is good information on this in the RwR text.
On Wed Mar 18 11:15:30 EDT 2009, "Lars O. Grobe" <lars.grobe@nus.edu.sg> wrote:
Hi...
I am simulating to find illuminance values for different angles of
Venetian blinds, Using Mkillum.. Do we need to consider Blinds as part
of the Luminous plane generated by mkillum..? We are getting better
results when using just a plane polygon as Luminous secondary source
as compared to when we include Blinds in the fenestration system as
luminous plane. ( comparing two different systems of blinds of similar
angles and characteristics.)
If I understand you right - are you asking wether to replace the each
blind's surface by a mkillum-generated illum??? This makes no sense -
you have aluminium blinds, and no light will pass through a blind. This
must lead to really funny results.
Mkillum has been written to precalculate the effect of compex
fenestration such as blinds. You calculate a light source on the inside
of the complex fenestration using mkillum, so that during your (rpict or
rtrace) rendering, the light source for ambient calculations is not
obstructed. The reasons for this are explained in detail in the book
"Rendering with Radiance".
In your case there are only two reasonable options as far as I
understood so far. Either render without mkillum, but very(!!!) high
quality settings for ad, as, aa, ab (which will take a long time), or
seal your complete fenestration system with mkillum-generated illum
surface(s) as explained in the book.