Curved mirror surfaces

Hello,

Does anyone have experience with mirror surfaces and especially curved mirror surfaces? I am not having luck getting sunlight to bounce off of them the way they should. Are there any tricks? I am using Ecotect and exporting to Radiance.

Thank you!

Jill Dalglish

Hi Jill,

have you checked the material definition that is exported by Ecotect? if
I remember correctly Ecotect exports "mirrors" as "plastic" so you'll
need to manually edit the definition in the .rad file.
And check that the normals are oriented the right way around.
Finally curved mirror reflections necessarily have to be approximated by
a mesh in this approach, so it will never look absolutely right....

Hope this helps,

Giovanni

···

-----Original Message-----
From: Jill Dalglish [mailto:[email protected]]
Sent: 01 February 2012 16:11
To: radiance-general
Subject: [Radiance-general] Curved mirror surfaces

Hello,

Does anyone have experience with mirror surfaces and especially curved
mirror surfaces? I am not having luck getting sunlight to bounce off of
them the way they should. Are there any tricks? I am using Ecotect and
exporting to Radiance.

Thank you!

Jill Dalglish

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

some details on how to cut strange shapes, but in the end you need polygons….

http://web.mac.com/geotrupes/iWeb/Main%20site/RadBlog/19993E4C-9633-4930-9D7F-72C4115D2E62.html

···

On 1 Feb 2012, at 16:10, Jill Dalglish wrote:

Hello,

Does anyone have experience with mirror surfaces and especially curved mirror surfaces? I am not having luck getting sunlight to bounce off of them the way they should. Are there any tricks? I am using Ecotect and exporting to Radiance.

Thank you!

Jill Dalglish

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

Hi Jill!

What is the modifier used to model the mirror surface? If Giovanni's
assumption is correct and your surfaces get a plastic modifier (probably
with high specular reflection), you cannot expect accurate "sharp"
shadows/highlights on diffuse reflective surfaces (such as a ceiling) as
all the light transport is handled by the ambient calculation. You would
need a mirror modifier applied to your surface. Unfortunately, with a
meshed curved surface, you would end up with a practically infinite
number of such mirrors, thus with an infinite number of secondary light
sources. There is some define in the code setting the maximum number of
secondary light sources, and one could set it higher (this has been used
for reflector modeling), but don't expect this to be efficient with an
architectural model.

I quote this from time to time (Rendering with Radiance, p. 580, chapter
13 "Secondary light sources"):

[...] "Other cases involving curved, specular reflectors pose similar
difficulties for mkillum, and the only long-term solution seams to be
the creation of a forward raytracing module for computing these kinds of
illums." [...]

What are you trying to model? Specular, curved blinds? Some kind of a
collector?

Cheers, Lars.

···

Hello,

Does anyone have experience with mirror surfaces and especially curved
mirror surfaces? I am not having luck getting sunlight to bounce off of
them the way they should. Are there any tricks? I am using Ecotect and
exporting to Radiance.

Thank you!

Jill Dalglish

Hi Jill, hi Lars,

I'd like to chuck in my humble 2cents worth and point out that abovementioned limitation is the very reason the photon mapping addon was developed for RADIANCE. This is precisely what Jill needs. Although it won't eliminate the tesselation artifacts, it will simulate the resultant caustics more efficiently than virtual light sources.

RADIANCE binaries with photon mapping were available on Francesco Anselmo's website, but appear to have been removed from the downloads. Anyone have any recent binaries Jill could try out?

Regards,

--Roland

···

On Wed, 01 Feb 2012 21:20:59 +0100, Lars O. Grobe <[email protected]> wrote:

I quote this from time to time (Rendering with Radiance, p. 580, chapter
13 "Secondary light sources"):

[...] "Other cases involving curved, specular reflectors pose similar
difficulties for mkillum, and the only long-term solution seams to be
the creation of a forward raytracing module for computing these kinds of
illums." [...]

--
"END OF LINE" [MCP, 1982]
"It's just not right that so many things work when they shouldn't!" [Apologies to Steve Wozniak]