mkillum -ab 0 question

radiance-general
1

Hi,

I'm running some daylight calculations and experimenting with mkillum.
I'm interested which light components does mkillum takes into account as
input.

I've found in RWR book on page 380 nice diagrams, but no info what is taken
into account when -ab 0 is used.

What I've understood by now:
mkillum -ab any_value:
l- option in geometry file - direct light from light sources (eg. sun) is
skipped from mkillum calculation and used later in rtrace/rpict directly.
Output geometry is made of 'illum' material
l+ option in geometry file - direct light from light sources is taken as
input and included in output light distribution. Output geometry is made of
'light' material

mkillum -ab 0:
*-which sky/sun/reflected components are taken as input for mkillum?*

mkillum -ab 1:
-direct light from sky is included
-indirect solar component (once reflected from ground or some other object)

mkillum -av 0 0 0:
-what does this give us (I know that -av should be 'ground ambient level'
from gensky?

···

####
To explain why do I need such strange mkillum parameters

My idea is to get light patch, coming from the sun and reflected from
specular deflector surface into the room, but to exclude all other
influences - direct light coming from sky or sun, or reflected light from
outside/inside surfaces.
I modeled deflector as mirror with alternative metal material with
appropriate reflectance/specularity.

*I want to get completely black image with reflected light patch only, so I
can sum images from various time moments during day/year and show light
patch moving for some period of time.*

To get image with specularly reflected light only:
1. I defined sky with sun only (no diffuse sky or ground sources)
2. Calculated mkillum -ab 0 -av 0 0 0 (l+ option) for my deflector surface
3. Used mkillum output in new octree as only light source (removed sun
definition)
4. Calculated image with rpict -ab 0 -av 0 0 0
Problem is that beside direct light patch inside the test room, I still
get some light on surfaces above/below my deflector surface.

Even when I repeat calculation for sun behind my surface (so no direct
light on surface), my output from mkillum has some small radiance, and I
get some luminance on surfaces below/above deflector.

Any ideas are welcome?

I hope I managed to explain what I want, and which method I try to use to
get results (unsuccessfully for now :frowning: )

Thanks in advance,
Marija

2

To compare the effect of a new design with an existing situation you could
just create two images (before and after) and subtract the "before" image
from the "after".

Appart from that I think you should use a mirror material for your
reflector. If it's not a flat surface use a reasonable geometric
approximation. It will look different but probably will still give you the
right idea about the effect. If you are only interested in the direct
reflection of the sun (single virtual light source per mirror polygon) you
should be able to render the scene even if you have a lot of mirror
surfaces.

To eliminate the sky (direct or indirect) remove the sky and ground
definition from the sky file (you are already doing that).

If you are not set on a "black and white" footprint of the reflection you
could also use a coloured mirror (say red) and use only white/grey
materials. Your reflection will show as a red patch while direct sunlight
will still be white.

Regards,
Thomas

···

On Wed, Dec 14, 2011 at 8:39 AM, Marija Velickovic <[email protected]>wrote:

*I want to get completely black image with reflected light patch only, so
I can sum images from various time moments during day/year and show light
patch moving for some period of time.*

To get image with specularly reflected light only:
1. I defined sky with sun only (no diffuse sky or ground sources)
2. Calculated mkillum -ab 0 -av 0 0 0 (l+ option) for my deflector surface
3. Used mkillum output in new octree as only light source (removed sun
definition)
4. Calculated image with rpict -ab 0 -av 0 0 0
Problem is that beside direct light patch inside the test room, I still
get some light on surfaces above/below my deflector surface.

3

Hi Thomas,

To compare the effect of a new design with an existing situation you could

just create two images (before and after) and subtract the "before" image
from the "after".

First thing that I've tried when started working on this problem is
subtraction of images 'before' and 'after'. It gave me nice light patch,
but also some other bright pixels on the edges of the area where direct sun
patch was - because of randomness of Radiance calculation (I suppose) .

Appart from that I think you should use a mirror material for your
reflector. If it's not a flat surface use a reasonable geometric
approximation. It will look different but probably will still give you the
right idea about the effect. If you are only interested in the direct
reflection of the sun (single virtual light source per mirror polygon) you
should be able to render the scene even if you have a lot of mirror
surfaces.

To eliminate the sky (direct or indirect) remove the sky and ground
definition from the sky file (you are already doing that).

Surface is flat so I've put mirror, and I removed sky/ground.
Problem is how to get rid of the direct sun and keep reflected one.

I tried rtcontrib but all contribution go to the sun. Thats why I'm using
mkillum to make my mirror deflector separate light source (with l+ option).
After that I get nice light patch from deflector, but also some light
around (which shouldn't be the case since I use -ab 0 -av 0 0 0).

If you are not set on a "black and white" footprint of the reflection you
could also use a coloured mirror (say red) and use only white/grey
materials. Your reflection will show as a red patch while direct sunlight
will still be white.

Yes good idea, I've tried it also and it works - so If I don't find some
way to do it nicely in black-white, I'll go to solution.

Thanks again,
Marija

4

Hi Marija,

I'm afraid that I'm to blame - the description that I gave in the book is wrong!

I've posted a (long overdue) erratum here:

http://www.iesd.dmu.ac.uk/~jm/doku.php?id=resources:erratum-mkillum

Apologies to those who have been led astray by this (embarrassing) mistake.

Best
John Mardaljevic

Reader in Daylight Modelling
Institute of Energy and Sustainable Development
De Montfort University, The Gateway, Leicester, LE1 9BH, UK
Tel: +44 (0) 116 257 7972

[email protected]
http://www.iesd.dmu.ac.uk/~jm
http://dmu.academia.edu/JohnMardaljevic

5

Hi John,

Thanks for posting this erratum, it explains why my images with mkillum -ab
0 and rpict -ab 0 seem illogical to me :slight_smile:

I'll print it and include in my book - to be sure I remember that for the
next time I have similar calcs.

Thanks again,
Marija

···

On Thu, Dec 15, 2011 at 12:46 PM, John Mardaljevic <[email protected]> wrote:

Hi Marija,

I'm afraid that I'm to blame - the description that I gave in the book is
wrong!

I've posted a (long overdue) erratum here:

http://www.iesd.dmu.ac.uk/~jm/doku.php?id=resources:erratum-mkillum

Apologies to those who have been led astray by this (embarrassing) mistake.

Best
John Mardaljevic