I've been having a discussion with the people at my office about the
differences between Radiance and 3D Studio Viz & Lightscape. (They are
used to using Lightscape and don't know anything about Radiance) And I
was hoping some of you might have some comparison or contrasting
comments on this subject since I'm sure most of you have had some
experience with this sort of discussion.

Viz, I think this is in part more a comparison about radiosity Vs
raytracing. From what I know, the big problem with radiosity is that it
simplifies things and assumes lambertian reflections. This is obviously
a deviation from reality, but I assume it's still possible to achieve a
pretty accurate calculation of some sort depending on the situation. (I
know the person I'm having this discussion with is certainly convinced
you can). Therefore what I am more interested in, is knowing what type
of calculation you really CAN'T do with radiosity (Viz or Lightscape)
that you could do with Radiance. Like if I am trying to evaluate the
performance of a light shelf, can I use a radiosity based engine to
compare the differences from one light shelf of material A with the same
light shelf made out of material B? I mean if material A is matte white
paint and material B is chrome, how could you compare the two scenarios
if the software assumes lambertian reflections?

Mark de la Fuente

Mark de la Fuente wrote:

I've been having a discussion with the people at my office about the differences between Radiance and 3D Studio Viz & Lightscape. (They are used to using Lightscape and don't know anything about Radiance) And I was hoping some of you might have some comparison or contrasting comments on this subject since I'm sure most of you have had some experience with this sort of discussion.

Hi Mark,
there's a 2001 Lightscape/Radiance comparison at http://www.pab-opto.de/ , although it would be worth redoing the calcs with Radiance+photonmap, to show the differences clearly.

Even though we are specifically talking about Radiance Vs Lightscape & Viz, I think this is in part more a comparison about radiosity Vs raytracing. From what I know, the big problem with radiosity is that it simplifies things and assumes lambertian reflections. This is obviously a deviation from reality, but I assume it's still possible to achieve a pretty accurate calculation of some sort depending on the situation. ...

Most materials show a specular peak at low incident angle. Interesting materials for light redirection are non-lambertian, starting with simple structures like outside louvers.
-Peter

···

--
pab-opto, Freiburg, Germany, http://www.pab-opto.de
[see web page to check digital email signature]

I've been having a discussion with the people at my office about the differences between Radiance and 3D Studio Viz & Lightscape. (They are used to using Lightscape and don't know anything about Radiance) And I was hoping some of you might have some comparison or contrasting comments on this subject since I'm sure most of you have had some experience with this sort of discussion.

Peter A-B's paper, mentioned in his reply to you, is an excellent resource. Print out Section 5.4 and hand it out to your co-workers, and then point them toward the radiance-online.org website. =8-)

I used Lightscape heavily for a number of years and beta tested both the last version of Lightscape and Viz4 (the first time Lightscape's code was used in Viz). But I was always looking toward Radiance and wishing I had the time to learn its many nuances. The demise of Lightscape (Viz is horrid) and the aquisition of a complex project finally conspired to push me into the Radiance learning mode full-time. (P.S. I'm still in that mode, three years later!) At this point, I use Radiance for all our calculations that require more than a simple zonal cavity or point calc.

Even though we are specifically talking about Radiance Vs Lightscape & Viz, I think this is in part more a comparison about radiosity Vs raytracing. From what I know, the big problem with radiosity is that it simplifies things and assumes lambertian reflections.

Right. Again, Peter & Kurt's paper is an excellent resource on all of this.

This is obviously a deviation from reality, but I assume it's still possible to achieve a pretty accurate calculation of some sort depending on the situation. (I know the person I'm having this discussion with is certainly convinced you can). Therefore what I am more interested in, is knowing what type of calculation you really CAN'T do with radiosity (Viz or Lightscape) that you could do with Radiance.

I think some confusion stems from the fact that Lightscape & Viz have a ray-tracing capability that you can apply, as a post-process. When you apply that to a Lightscape/Viz model, the specular reflections are *rendered*, but not *calculated*. So you have a rendering that looks somewhat accurate, but at that point that's all you have. You cannot derive quantitative luminance/illuminance information from the image at that point. That data always comes from the radiosity-calculated model, and that's fundamentally flawed because of the lack of specular reflections (and in Lightscape's case, no diffuse transmissions either).

···

On May 18, 2004, at 7:07 PM, Mark de la Fuente wrote:

=================
Rob Guglielmetti
www.rumblestrip.org

Hi Mark,

see below.

Martin Moeck

···

-----Original Message-----
From: Mark de la Fuente [mailto:[email protected]]
Sent: Tue 5/18/2004 7:07 PM
To: [email protected]
Cc:

Even though we are specifically talking about Radiance Vs Lightscape & Viz, I think this is in part more a comparison about radiosity Vs raytracing. From what I know, the big problem with radiosity is that it simplifies things and assumes lambertian reflections. This is obviously a deviation from reality, but I assume it's still possible to achieve a pretty accurate calculation of some sort depending on the situation. (I know the person I'm having this discussion with is certainly convinced you can). Therefore what I am more interested in, is knowing what type of calculation you really CAN'T do with radiosity (Viz or Lightscape) that you could do with Radiance.

1) Venetian blinds: If you cannot subdivide surfaces like Venetian blinds into very small patches and override the automatic triangular subdivision (i.e., Lumen Designer), the calculation is meaningless. Exception: The new AGI32

2) Specular highlights are not considered in the hybrid raytracing method in AGI32, Lumen Designer and VIZ4. You might see a small reflection of a light source on a varnished oak floor. If you increase -ad to a large number, i.e., 5000 - 10,000, Radiance will consider the contribution of that specular highlight. The luminance of this highlight is typically very significant, even if the specular reflectance of the varnished floor is 1% or less. A sun beam reflected off that floor can create an additional 200 Lux on the ceiling in the specular direction, which the radiosity calculation never catches. This error is very significant.

3) You can combine Radiance plastic, metal and glass with the mirror material and turn any Radiance surface that has a slight or large specular component into a virtual light source. Again, the error is very significant in radiosity since those specuar reflections are ignored.

4) The important underlying assumption in using radiosity is that building materials are mostly diffuse, and might only have a specularity of 0.1% or so. Assume even a very low specularity of 0.1%, times the luminance of the sun or an HID lamp reflected in that "matte" material. This is an extremely bright virtual light source, no matter if the material is black or white. That contribution cannot be ignored. If you have an aluminum ceiling, aluminum lightshelf and some blinds, radiosity cannot be used at all, since the specularity of most aluminum materials is 50% to 902%.

Like if I am trying to evaluate the performance of a light shelf, can I use a radiosity based engine to compare the differences from one light shelf of material A with the same light shelf made out of material B? I mean if material A is matte white paint and ! material B is chrome, how could you compare the two scenarios if the software assumes lambertian reflections?

Not at all. You would have to do it by hand and it would be much more accurate. The diffuse and specular reflectance of both materials is known, and all you need is Excel to calculate the specular and diffuse light transfer. Then add 15% indirect and you got it.

"3) You can combine Radiance plastic, metal and glass with the mirror
material and turn any Radiance surface that has a slight or large specular
component into a virtual light source. Again, the error is very significant
in radiosity since those specuar reflections are ignored. "

unfortunately you cannot use mixfunc with virtual sources...
....I think that virtual sources nested into complex material descriptions

giulio

···

-----Original Message-----
From: Martin Moeck [mailto:[email protected]]
Sent: 19 May 2004 15:56

Hi Mark,

see below.

Martin Moeck

-----Original Message-----
From: Mark de la Fuente [mailto:[email protected]]
Sent: Tue 5/18/2004 7:07 PM
To: [email protected]
Cc:

Even though we are specifically talking about Radiance Vs Lightscape & Viz,
I think this is in part more a comparison about radiosity Vs raytracing.

From what I know, the big problem with radiosity is that it simplifies

things and assumes lambertian reflections. This is obviously a deviation
from reality, but I assume it's still possible to achieve a pretty accurate
calculation of some sort depending on the situation. (I know the person I'm
having this discussion with is certainly convinced you can). Therefore what
I am more interested in, is knowing what type of calculation you really
CAN'T do with radiosity (Viz or Lightscape) that you could do with Radiance.

1) Venetian blinds: If you cannot subdivide surfaces like Venetian blinds
into very small patches and override the automatic triangular subdivision
(i.e., Lumen Designer), the calculation is meaningless. Exception: The new
AGI32

2) Specular highlights are not considered in the hybrid raytracing method in
AGI32, Lumen Designer and VIZ4. You might see a small reflection of a light
source on a varnished oak floor. If you increase -ad to a large number,
i.e., 5000 - 10,000, Radiance will consider the contribution of that
specular highlight. The luminance of this highlight is typically very
significant, even if the specular reflectance of the varnished floor is 1%
or less. A sun beam reflected off that floor can create an additional 200
Lux on the ceiling in the specular direction, which the radiosity
calculation never catches. This error is very significant.

3) You can combine Radiance plastic, metal and glass with the mirror
material and turn any Radiance surface that has a slight or large specular
component into a virtual light source. Again, the error is very significant
in radiosity since those specuar reflections are ignored.

4) The important underlying assumption in using radiosity is that building
materials are mostly diffuse, and might only have a specularity of 0.1% or
so. Assume even a very low specularity of 0.1%, times the luminance of the
sun or an HID lamp reflected in that "matte" material. This is an extremely
bright virtual light source, no matter if the material is black or white.
That contribution cannot be ignored. If you have an aluminum ceiling,
aluminum lightshelf and some blinds, radiosity cannot be used at all, since
the specularity of most aluminum materials is 50% to 902%.

Like if I am trying to evaluate the performance of a light shelf, can I use
a radiosity based engine to compare the differences from one light shelf of
material A with the same light shelf made out of material B? I mean if
material A is matte white paint and ! material B is chrome, how could you
compare the two scenarios if the software assumes lambertian reflections?

Not at all. You would have to do it by hand and it would be much more
accurate. The diffuse and specular reflectance of both materials is known,
and all you need is Excel to calculate the specular and diffuse light
transfer. Then add 15% indirect and you got it.

___________________________________________________________________
Electronic mail messages entering and leaving Arup business
systems are scanned for acceptability of content and viruses.

from the "Lord":

"That is correct -- Radiance only finds virtual source objects using

mirror or prism material types directly. It doesn't even look if you

have a modifier in front of the material in the chain (e.g., you

reference a pattern before the material)."

cheers!

···

-----Original Message-----
From: Giulio Antonutto [mailto:[email protected]]
Sent: 19 May 2004 16:04

"3) You can combine Radiance plastic, metal and glass with the mirror
material and turn any Radiance surface that has a slight or large specular
component into a virtual light source. Again, the error is very significant
in radiosity since those specuar reflections are ignored. "

unfortunately you cannot use mixfunc with virtual sources...
....I think that virtual sources nested into complex material descriptions

giulio

-----Original Message-----
From: Martin Moeck [mailto:[email protected]]
Sent: 19 May 2004 15:56

Hi Mark,

see below.

Martin Moeck

-----Original Message-----
From: Mark de la Fuente [mailto:[email protected]]
Sent: Tue 5/18/2004 7:07 PM
To: [email protected]
Cc:

Even though we are specifically talking about Radiance Vs Lightscape & Viz,
I think this is in part more a comparison about radiosity Vs raytracing.

From what I know, the big problem with radiosity is that it simplifies

things and assumes lambertian reflections. This is obviously a deviation
from reality, but I assume it's still possible to achieve a pretty accurate
calculation of some sort depending on the situation. (I know the person I'm
having this discussion with is certainly convinced you can). Therefore what
I am more interested in, is knowing what type of calculation you really
CAN'T do with radiosity (Viz or Lightscape) that you could do with Radiance.

1) Venetian blinds: If you cannot subdivide surfaces like Venetian blinds
into very small patches and override the automatic triangular subdivision
(i.e., Lumen Designer), the calculation is meaningless. Exception: The new
AGI32

2) Specular highlights are not considered in the hybrid raytracing method in
AGI32, Lumen Designer and VIZ4. You might see a small reflection of a light
source on a varnished oak floor. If you increase -ad to a large number,
i.e., 5000 - 10,000, Radiance will consider the contribution of that
specular highlight. The luminance of this highlight is typically very
significant, even if the specular reflectance of the varnished floor is 1%
or less. A sun beam reflected off that floor can create an additional 200
Lux on the ceiling in the specular direction, which the radiosity
calculation never catches. This error is very significant.

3) You can combine Radiance plastic, metal and glass with the mirror
material and turn any Radiance surface that has a slight or large specular
component into a virtual light source. Again, the error is very significant
in radiosity since those specuar reflections are ignored.

4) The important underlying assumption in using radiosity is that building
materials are mostly diffuse, and might only have a specularity of 0.1% or
so. Assume even a very low specularity of 0.1%, times the luminance of the
sun or an HID lamp reflected in that "matte" material. This is an extremely
bright virtual light source, no matter if the material is black or white.
That contribution cannot be ignored. If you have an aluminum ceiling,
aluminum lightshelf and some blinds, radiosity cannot be used at all, since
the specularity of most aluminum materials is 50% to 902%.

Like if I am trying to evaluate the performance of a light shelf, can I use
a radiosity based engine to compare the differences from one light shelf of
material A with the same light shelf made out of material B? I mean if
material A is matte white paint and ! material B is chrome, how could you
compare the two scenarios if the software assumes lambertian reflections?

Not at all. You would have to do it by hand and it would be much more
accurate. The diffuse and specular reflectance of both materials is known,
and all you need is Excel to calculate the specular and diffuse light
transfer. Then add 15% indirect and you got it.

___________________________________________________________________

Electronic mail messages entering and leaving Arup business

systems are scanned for acceptability of content and viruses.

___________________________________________________________________
Electronic mail messages entering and leaving Arup business
systems are scanned for acceptability of content and viruses.

Another P.S... Following the quoted conversation from Pillo, I went off and improved the material handling for sources (and virtual sources) so that it does look up the modifier tree past patterns and things that might come between the material primitive and the object. (If you understood that last sentence, you're probably in the top 2% of this class.) However, the code still isn't smart enough to consider mixfunc's that make material assignment as part of ray evaluation, which is basically incompatible with a priori assessments of which surfaces are considered light sources and which surfaces aren't. In other words, don't expect me to fix this anytime soon.

If you are trying to reflect objects in a surface, the "mirror" primitive allows you to specify any alternate material you choose when it is not participating in the virtual source calculation. This should be enough to do what you want in most cases.

The Lord of Radiance has spoked.
-Greg

From: Giulio Antonutto <[email protected]>
Date: May 19, 2004 8:10:39 AM PDT

from the "Lord":

"That is correct -- Radiance only finds virtual source objects using

mirror or prism material types directly. It doesn't even look if you

have a modifier in front of the material in the chain (e.g., you

reference a pattern before the material)."

cheers!

Goooongggg! Goooongggg!
(the sound of temple gong...)

···

-----Original Message-----
From: [email protected]
[mailto:[email protected]]On Behalf Of Greg
Ward
Sent: 19 May 2004 17:05

Another P.S... Following the quoted conversation from Pillo, I went
off and improved the material handling for sources (and virtual
sources) so that it does look up the modifier tree past patterns and
things that might come between the material primitive and the object.
(If you understood that last sentence, you're probably in the top 2% of
this class.) However, the code still isn't smart enough to consider
mixfunc's that make material assignment as part of ray evaluation,
which is basically incompatible with a priori assessments of which
surfaces are considered light sources and which surfaces aren't. In
other words, don't expect me to fix this anytime soon.

If you are trying to reflect objects in a surface, the "mirror"
primitive allows you to specify any alternate material you choose when
it is not participating in the virtual source calculation. This should
be enough to do what you want in most cases.

The Lord of Radiance has spoked.
-Greg

From: Giulio Antonutto <[email protected]>
Date: May 19, 2004 8:10:39 AM PDT

from the "Lord":

"That is correct -- Radiance only finds virtual source objects using

mirror or prism material types directly. It doesn't even look if you

have a modifier in front of the material in the chain (e.g., you

reference a pattern before the material)."

cheers!

_______________________________________________
[email protected]

___________________________________________________________________
Electronic mail messages entering and leaving Arup business
systems are scanned for acceptability of content and viruses.

void plastic df
0 0 5
.2 .2 .2 .2 .01

void mirror floor
1 df
0
3 .2 .2 .2

does it

Martin

···

-----Original Message-----
From: Giulio Antonutto [mailto:[email protected]]
Sent: Wed 5/19/2004 11:03 AM
Cc:

"3) You can combine Radiance plastic, metal and glass with the mirror material and turn any Radiance surface that has a slight or large specular component into a virtual light source. Again, the error is very significant in radiosity since those specuar reflections are ignored. "

unfortunately you cannot use mixfunc with virtual sources...
....I think that virtual sources nested into complex material descriptions are simply ignored... but I could be wrong (any confirm about this?).

giulio

-----Original Message-----
From: Martin Moeck [mailto:[email protected]]
Sent: 19 May 2004 15:56

Hi Mark,

see below.

Martin Moeck

-----Original Message-----
From: Mark de la Fuente [mailto:[email protected]]
Sent: Tue 5/18/2004 7:07 PM
To: [email protected]
Cc:

Even though we are specifically talking about Radiance Vs Lightscape & Viz, I think this is in part more a comparison about radiosity Vs raytracing. From what I know, the big problem with radiosity is that it simplifies things and assumes lambertian reflections. This is obviously a deviation from reality, but I assume it's still possible to achieve a pretty accurate calculation of some sort depending on the situation. (I know the person I'm having this discussion with is certainly convinced you can). Therefore what I am more interested in, is knowing what type of calculation you really CAN'T do with radiosity (Viz or Lightscape) that you could do with Radiance.

1) Venetian blinds: If you cannot subdivide surfaces like Venetian blinds into very small patches and override the automatic triangular subdivision (i.e., Lumen Designer), the calculation is meaningless. Exception: The new AGI32

2) Specular highlights are not considered in the hybrid raytracing method in AGI32, Lumen Designer and VIZ4. You might see a small reflection of a light source on a varnished oak floor. If you increase -ad to a large number, i.e., 5000 - 10,000, Radiance will consider the contribution of that specular highlight. The luminance of this highlight is typically very significant, even if the specular reflectance of the varnished floor is 1% or less. A sun beam reflected off that floor can create an additional 200 Lux on the ceiling in the specular direction, which the radiosity calculation never catches. This error is very significant.

3) You can combine Radiance plastic, metal and glass with the mirror material and turn any Radiance surface that has a slight or large specular component into a virtual light source. Again, the error is very significant in radiosity since those specuar reflections are ignored.

4) The important underlying assumption in using radiosity is that building materials are mostly diffuse, and might only have a specularity of 0.1% or so. Assume even a very low specularity of 0.1%, times the luminance of the sun or an HID lamp reflected in that "matte" material. This is an extremely bright virtual light source, no matter if the material is black or white. That contribution cannot be ignored. If you have an aluminum ceiling, aluminum lightshelf and some blinds, radiosity cannot be used at all, since the specularity of most aluminum materials is 50% to 902%.

Like if I am trying to evaluate the performance of a light shelf, can I use a radiosity based engine to compare the differences from one light shelf of material A with the same light shelf made out of material B? I mean if material A is matte white paint and ! material B is chrome, how could you compare the two scenarios if the software assumes lambertian reflections?

Not at all. You would have to do it by hand and it would be much more accurate. The diffuse and specular reflectance of both materials is known, and all you need is Excel to calculate the specular and diffuse light transfer. Then add 15% indirect and you got it.

___________________________________________________________________

Electronic mail messages entering and leaving Arup business

systems are scanned for acceptability of content and viruses.

sorry, I thought you where combining things with mixfunc...
good trick!
However what I wanted was to create holes on a surface that was acting like
a mirror but unfortunately for me you cannot use mixfunc with virtual
sources, is it?
cheers,
giulio

···

-----Original Message-----
From: Martin Moeck [mailto:[email protected]]On
Behalf Of Martin Moeck
Sent: 19 May 2004 18:27

void plastic df
0 0 5
.2 .2 .2 .2 .01

void mirror floor
1 df
0
3 .2 .2 .2

does it

Martin

-----Original Message-----
From: Giulio Antonutto [mailto:[email protected]]
Sent: Wed 5/19/2004 11:03 AM
Cc:

"3) You can combine Radiance plastic, metal and glass with the mirror
material and turn any Radiance surface that has a slight or large specular
component into a virtual light source. Again, the error is very significant
in radiosity since those specuar reflections are ignored. "

unfortunately you cannot use mixfunc with virtual sources...
....I think that virtual sources nested into complex material descriptions

giulio

-----Original Message-----
From: Martin Moeck [mailto:[email protected]]
Sent: 19 May 2004 15:56

Hi Mark,

see below.

Martin Moeck

-----Original Message-----
From: Mark de la Fuente [mailto:[email protected]]
Sent: Tue 5/18/2004 7:07 PM
To: [email protected]
Cc:

Even though we are specifically talking about Radiance Vs Lightscape & Viz,
I think this is in part more a comparison about radiosity Vs raytracing.

From what I know, the big problem with radiosity is that it simplifies

things and assumes lambertian reflections. This is obviously a deviation
from reality, but I assume it's still possible to achieve a pretty accurate
calculation of some sort depending on the situation. (I know the person I'm
having this discussion with is certainly convinced you can). Therefore what
I am more interested in, is knowing what type of calculation you really
CAN'T do with radiosity (Viz or Lightscape) that you could do with Radiance.

1) Venetian blinds: If you cannot subdivide surfaces like Venetian blinds
into very small patches and override the automatic triangular subdivision
(i.e., Lumen Designer), the calculation is meaningless. Exception: The new
AGI32

2) Specular highlights are not considered in the hybrid raytracing method in
AGI32, Lumen Designer and VIZ4. You might see a small reflection of a light
source on a varnished oak floor. If you increase -ad to a large number,
i.e., 5000 - 10,000, Radiance will consider the contribution of that
specular highlight. The luminance of this highlight is typically very
significant, even if the specular reflectance of the varnished floor is 1%
or less. A sun beam reflected off that floor can create an additional 200
Lux on the ceiling in the specular direction, which the radiosity
calculation never catches. This error is very significant.

3) You can combine Radiance plastic, metal and glass with the mirror
material and turn any Radiance surface that has a slight or large specular
component into a virtual light source. Again, the error is very significant
in radiosity since those specuar reflections are ignored.

4) The important underlying assumption in using radiosity is that building
materials are mostly diffuse, and might only have a specularity of 0.1% or
so. Assume even a very low specularity of 0.1%, times the luminance of the
sun or an HID lamp reflected in that "matte" material. This is an extremely
bright virtual light source, no matter if the material is black or white.
That contribution cannot be ignored. If you have an aluminum ceiling,
aluminum lightshelf and some blinds, radiosity cannot be used at all, since
the specularity of most aluminum materials is 50% to 902%.

Like if I am trying to evaluate the performance of a light shelf, can I use
a radiosity based engine to compare the differences from one light shelf of
material A with the same light shelf made out of material B? I mean if
material A is matte white paint and ! material B is chrome, how could you
compare the two scenarios if the software assumes lambertian reflections?

Not at all. You would have to do it by hand and it would be much more
accurate. The diffuse and specular reflectance of both materials is known,
and all you need is Excel to calculate the specular and diffuse light
transfer. Then add 15% indirect and you got it.

___________________________________________________________________

Electronic mail messages entering and leaving Arup business

systems are scanned for acceptability of content and viruses.

___________________________________________________________________
Electronic mail messages entering and leaving Arup business
systems are scanned for acceptability of content and viruses.

Hi Pillo,

It wouldn't be exactly right, but you could make a pattern that created dark dots on your mirror in place of holes, then use as an alternate material a mixfunc with proper holes. You'd get a holey reflection as well as the correct view of the object when you render it.

-Greg

From: Giulio Antonutto <[email protected]>
Date: May 19, 2004 10:35:18 AM PDT
To: "'Radiance general discussion'" <[email protected]>

sorry, I thought you where combining things with mixfunc...
good trick!
However what I wanted was to create holes on a surface that was acting like a mirror but unfortunately for me you cannot use mixfunc with virtual sources, is it?
cheers,
giulio

···