Metal Reflection Problem

Marcus_Jacobs · November 3, 2003, 12:15am

Dear Group

I have run into an issue with reflections for metal and metal2 materials. In my models, I am recieving some noisy reflections. In the following scene, I have modeled some kitchen appliances to have a brushed stainless steel finish with the following material:

void metal2 brushed_metal_01m1
4 1 1 1 .
0
6 1.0000 1.0000 1.0000 0.75000 0.040 0.075

Here are some renderings of the scene:

http://f1.pg.photos.yahoo.com/ph/marcdevon/vwp?.dir=/RadiancePics&.dnm=room2a_camera1.jpg&.view=t
http://f1.pg.photos.yahoo.com/ph/marcdevon/vwp?.dir=/RadiancePics&.dnm=room2a_camera2.jpg&.view=t
http://f1.pg.photos.yahoo.com/ph/marcdevon/vwp?.dir=/RadiancePics&.dnm=room2a_camera3.jpg&.view=t
http://f1.pg.photos.yahoo.com/ph/marcdevon/vwp?.dir=/RadiancePics&.dnm=room2a_camera4.jpg&.view=t
http://f1.pg.photos.yahoo.com/ph/marcdevon/vwp?.dir=/RadiancePics&.dnm=room2a_camera5.jpg&.view=t

Here are some of the pertinent render parameters:

-ps 1 -sj 1 -st 0 -dt 0

The pictures were rendered at 1000 x 1000 and filtered to 500 x 500.

I know that there should be some blurry reflections on the surfaces on the appliances due to the roughness of the material but I question whether or not the reflections should look like fractalized noise. They look far more fractalized in their native *.pic format that the *.jpg format that you see here. The following is a different scene but with the same appliances and associated materials as the previous scene. Instead of being illuminated by daylight, it is illuminated by some illumination squares. Here you can see the reflections are blurry yet smooth except for the reflection of the refridgerator handle on the refridgerator door. Here it seems that we have the same fractalized noise phenomenon (I can provide a close up of this to show more detail).:

http://f1.pg.photos.yahoo.com/ph/marcdevon/vwp?.dir=/RadiancePics&.dnm=room1_default1a.jpg&.view=t
http://f1.pg.photos.yahoo.com/ph/marcdevon/vwp?.dir=/RadiancePics&.dnm=room1_default2a.jpg&.view=t
http://f1.pg.photos.yahoo.com/ph/marcdevon/vwp?.dir=/RadiancePics&.dnm=room1_default3a.jpg&.view=t
http://f1.pg.photos.yahoo.com/ph/marcdevon/vwp?.dir=/RadiancePics&.dnm=room1_default4a.jpg&.view=t
http://f1.pg.photos.yahoo.com/ph/marcdevon/vwp?.dir=/RadiancePics&.dnm=room1_default5a.jpg&.view=t

As a side note, the reflections of the light sources on the floor do not seem right. I have noticed this before. Does anyone know what's causing this?

I have tried to figure out what's causing all of this but I can't get anywhere. I should note that all of the previous scenes were rendered with Desktop Radiance binaries. I have tested to see if I got similar results using Radiance 3.6 binaries using CYGWIN. I have confirmed this to be the case. The folling scene is rendered with Radiance 3.6 binaries for CYGWIN:

http://f1.pg.photos.yahoo.com/ph/marcdevon/vwp?.dir=/RadiancePics&.dnm=house6a_camera6_filt1.jpg&.view=t

Let me know if anyone has problems viewing these photos.

Thanks

Marcus

···

_________________________________________________________________
Want to check if your PC is virus-infected? Get a FREE computer virus scan online from McAfee. http://clinic.mcafee.com/clinic/ibuy/campaign.asp?cid=3963

Roland_Schregle · November 3, 2003, 12:32am

Marcus Jacobs wrote:

I have run into an issue with reflections for metal and metal2 materials. In my models, I am recieving some noisy reflections.

Here are some of the pertinent render parameters:

-ps 1 -sj 1 -st 0 -dt 0

Hi Marcus,

try upping -ps. Default for rpict is 4.
You might also wanna reduce -aa, and increase -ar and -ad. There's some splotches on the walls, possibly due to caustics from the appliances.

As a side note, the reflections of the light sources on the floor do not seem right. I have noticed this before. Does anyone know what's causing this?

Don't seem right as in no penumbrae? That's expected for a solar source modeled with the SOURCE primitive. AFAIK these sources are immune to -dj and -ds anyway. Maybe a Fully Licensed Guru[tm] here can tell you more.

Btw, that's a big-ass microwave! :^)

--Roland

···

--
Roland Schregle
PhD candidate, Fraunhofer Institute for Solar Energy Systems
RADIANCE Photon Map page: www.ise.fhg.de/radiance/photon-map

END OF LINE. (MCP)

Carsten_Bauer2 · November 3, 2003, 10:51am

Hi Marcus,

in pic room1_default5a you probably miss the reflection of the big light-source rectangle on the floor. This is the classical backward ray tracing problem, with photon-map things would look differently.

concerning the metal reflections: high jittering parameters can be tricky, a good way to find out what's happening is to start with very low settings for both material and rendering params (i.e. almost smooth, polished surfaces and the default -dj -dc -dt -sj -st vals) and raise them in little steps subsequently to grasp the point when the modelling differs from the expectations.

the latter should not be too high, either, many special surface properties of custom materials can not be modeled adequately with the standard Radiance models. This is understandable as things become really difficult when looking at the details. sometimes it helps to add an extra normal perturbing texture modifier to the specification, as one often has to deal with "roughness structures" which are too big to fit into the microscopic roughness models employed by plastic(2)/metal(2) materials

-cb

Mark_Stock · November 3, 2003, 4:05pm

Marcus,

I think the "fractal pattern" you refer to in the specular
highlights can be solved by turning on Radiance's pure Monte-
Carlo sampling algorithm (requires re-compiling, though).
I read about it in a text file on the old Radiance site:

···

----------
RADIANCE Compile Switches

Here is a list of compile switches, used to customize Radiance code for
specific machines and users:

-DMC
If set, switches from default low-discrepency sequence sampling to
true (pseudorandom) Monte Carlo. Use if the "brushed" appearance of
specular highlights and penumbras bothers you.
----------

To get this to compile with the HEAD distribution, you need to
make a few changes to the src/common/random.h file. I can give those
to you if you like.

also...

On Mon, 3 Nov 2003, Roland Schregle wrote:

Marcus Jacobs wrote:

> As a side note, the reflections of the light sources on the floor do not
> seem right. I have noticed this before. Does anyone know what's causing
> this?

Don't seem right as in no penumbrae? That's expected for a solar source
modeled with the SOURCE primitive. AFAIK these sources are immune to -dj
and -ds anyway. Maybe a Fully Licensed Guru[tm] here can tell you more.

I've found that -dj works with the "source" primitive (again, I
use the HEAD distro, but I found it to work in the 3R5 distro, too).
It uses the last float argument, the angular disc size, to determine
the jittering. Go ahead and raise that from 0.5 to 5 degrees for an
unrealistic sun with nice-looking soft shadows.

I typically use -dj 0.6 to 0.7 for most of my scenes.

BTW: the soft shadows will exhibit the same fractal-like pattern
that you see in the specular highlights. -DMC cures this, too,
but can require higher-resolution runs before filtering down.

Oh, and -ps 1 is almost a necessity with -DMC turned on.

Mark

Marcus_Jacobs · November 4, 2003, 2:07pm

Dear Mark

I would greatly appreciate if you could show me as to how to make these changes.

Thanks

Marcus

···

Date: Mon, 3 Nov 2003 11:05:03 -0500 (EST)
From: Mark Stock <[email protected]>
To: [email protected]
Subject: Re: [Radiance-general] Metal Reflection Problem
Reply-To: [email protected]

Marcus,

I think the "fractal pattern" you refer to in the specular
highlights can be solved by turning on Radiance's pure Monte-
Carlo sampling algorithm (requires re-compiling, though).
I read about it in a text file on the old Radiance site:

----------
RADIANCE Compile Switches

Here is a list of compile switches, used to customize Radiance code for
specific machines and users:

-DMC
If set, switches from default low-discrepency sequence sampling to
true (pseudorandom) Monte Carlo. Use if the "brushed" appearance of
specular highlights and penumbras bothers you.
----------

To get this to compile with the HEAD distribution, you need to
make a few changes to the src/common/random.h file. I can give those
to you if you like.

also...

On Mon, 3 Nov 2003, Roland Schregle wrote:

> Marcus Jacobs wrote:
>
> > As a side note, the reflections of the light sources on the floor do not
> > seem right. I have noticed this before. Does anyone know what's causing
> > this?
>
> Don't seem right as in no penumbrae? That's expected for a solar source
> modeled with the SOURCE primitive. AFAIK these sources are immune to -dj
> and -ds anyway. Maybe a Fully Licensed Guru[tm] here can tell you more.

I've found that -dj works with the "source" primitive (again, I
use the HEAD distro, but I found it to work in the 3R5 distro, too).
It uses the last float argument, the angular disc size, to determine
the jittering. Go ahead and raise that from 0.5 to 5 degrees for an
unrealistic sun with nice-looking soft shadows.

I typically use -dj 0.6 to 0.7 for most of my scenes.

BTW: the soft shadows will exhibit the same fractal-like pattern
that you see in the specular highlights. -DMC cures this, too,
but can require higher-resolution runs before filtering down.

Oh, and -ps 1 is almost a necessity with -DMC turned on.

Mark

--__--__--

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

End of Radiance-general Digest

_________________________________________________________________

Carsten_Bauer2 · November 4, 2003, 6:30pm

Hi Marcus,

the necessary changes are some code-exclusions in two Radiance -source files of the main distribution.
Probably this has already been changed in the HEAD version (dunno...so far I haven't entered the Headshop).

Anyway, I once posted the changes in an answer-mail some months ago:

http://www.radiance-online.org/pipermail/radiance-general/2003-May/000702.html

good luck

-cb

Mark_Stock · November 4, 2003, 6:47pm

Marcus,

Hmmm, the changes listed in Carsten's e-mail aren't quite the
same ones that I came up with. They're simpler than mine, and
I'm sure they work, too.

Of course, this may all be moot, as I was able to build today's
HEAD with -DMC invoked without any code changes (except pointing
to /usr/lib/libtiff.a)

Now, if you're building the photon map enhancement, which requires
the 3R5 release, you'll need the changes to get -DMC working.

The changes that I came up with require a few modifications to the
src/common/random.h file:

1) in the "#ifdef MC" section, add:
extern unsigned short *urperm;

2) after the next "#else", add:
extern int initurand(int size);

3) remove, or comment out the following lines:

···

----------
#ifdef NOPROTO

extern int initurand();
extern int ilhash();
extern int urind();
extern void multisamp();

#else
/* defined in urand.c */
extern int initurand(int size);
----------

and the final:
----------
#endif
----------

I guess it's kind of long-winded, but it builds on Linux
(Linux 2.4.20-20_29.rh9.at, gcc (GCC) 3.2.2 20030222)

Mark

On Tue, 4 Nov 2003, Marcus Jacobs wrote:

I would greatly appreciate if you could show me as to how to make these
changes.

>Date: Mon, 3 Nov 2003 11:05:03 -0500 (EST)
>From: Mark Stock <[email protected]>
>Subject: Re: [Radiance-general] Metal Reflection Problem
>
>I think the "fractal pattern" you refer to in the specular
>highlights can be solved by turning on Radiance's pure Monte-
>Carlo sampling algorithm (requires re-compiling, though).
>I read about it in a text file on the old Radiance site:
>
>----------
>RADIANCE Compile Switches
>
>Here is a list of compile switches, used to customize Radiance code for
>specific machines and users:
>
>-DMC
> If set, switches from default low-discrepency sequence sampling to
>true (pseudorandom) Monte Carlo. Use if the "brushed" appearance of
>specular highlights and penumbras bothers you.
>----------
>
>To get this to compile with the HEAD distribution, you need to
>make a few changes to the src/common/random.h file. I can give those
>to you if you like.

Marcus_Jacobs · November 2, 2003, 4:34pm

Dear Group

I have run into an issue with reflections for metal and metal2 materials. In my models, I am recieving some noisy reflections. In the following scene, I have modeled some kitchen appliances to have a brushed stainless steel finish with the following material:

void metal2 brushed_metal_01m1
4 1 1 1 .
0
6 1.0000 1.0000 1.0000 0.75000 0.040 0.075

Here are some renderings of the scene:

http://f1.pg.photos.yahoo.com/ph/marcdevon/vwp?.dir=/RadiancePics&.dnm=room2a_camera1.jpg&.view=t
http://f1.pg.photos.yahoo.com/ph/marcdevon/vwp?.dir=/RadiancePics&.dnm=room2a_camera2.jpg&.view=t
http://f1.pg.photos.yahoo.com/ph/marcdevon/vwp?.dir=/RadiancePics&.dnm=room2a_camera3.jpg&.view=t
http://f1.pg.photos.yahoo.com/ph/marcdevon/vwp?.dir=/RadiancePics&.dnm=room2a_camera4.jpg&.view=t
http://f1.pg.photos.yahoo.com/ph/marcdevon/vwp?.dir=/RadiancePics&.dnm=room2a_camera5.jpg&.view=t

Here are some of the pertinent render parameters:

-ps 1 -sj 1 -st 0 -dt 0

The pictures were rendered at 1000 x 1000 and filtered to 500 x 500.

I know that there should be some blurry reflections on the surfaces on the appliances due to the roughness of the material but I question whether or not the reflections should look like fractalized noise. They look far more fractalized in their native *.pic format that the *.jpg format that you see here. The following is a different scene but with the same appliances and associated materials as the previous scene. Instead of being illuminated by daylight, it is illuminated by some illumination squares. Here you can see the reflections are blurry yet smooth except for the reflection of the refridgerator handle on the refridgerator door. Here it seems that we have the same fractalized noise phenomenon (I can provide a close up of this to show more detail).:

http://f1.pg.photos.yahoo.com/ph/marcdevon/vwp?.dir=/RadiancePics&.dnm=room1_default1a.jpg&.view=t
http://f1.pg.photos.yahoo.com/ph/marcdevon/vwp?.dir=/RadiancePics&.dnm=room1_default2a.jpg&.view=t
http://f1.pg.photos.yahoo.com/ph/marcdevon/vwp?.dir=/RadiancePics&.dnm=room1_default3a.jpg&.view=t
http://f1.pg.photos.yahoo.com/ph/marcdevon/vwp?.dir=/RadiancePics&.dnm=room1_default4a.jpg&.view=t
http://f1.pg.photos.yahoo.com/ph/marcdevon/vwp?.dir=/RadiancePics&.dnm=room1_default5a.jpg&.view=t

As a side note, the reflections of the light sources on the floor do not seem right. I have noticed this before. Does anyone know what's causing this?

I have tried to figure out what's causing all of this but I can't get anywhere. I should note that all of the previous scenes were rendered with Desktop Radiance binaries. I have tested to see if I got similar results using Radiance 3.6 binaries using CYGWIN. I have confirmed this to be the case. The folling scene is rendered with Radiance 3.6 binaries for CYGWIN:

http://f1.pg.photos.yahoo.com/ph/marcdevon/vwp?.dir=/RadiancePics&.dnm=house6a_camera6_filt1.jpg&.view=t

Let me know if anyone has problems viewing these photos.

Thanks

Marcus

···

_________________________________________________________________
Concerned that messages may bounce because your Hotmail account has exceeded its 2MB storage limit? Get Hotmail Extra Storage! http://join.msn.com/?PAGE=features/es