Guidelines for trans Material

Dear Group

I am modeling a bathroom that will use translucent light fixtures similar to this:

I wish to accurately simulate the light distribution from the fixtures however I do not know where to begin as far as determining physically valid parameters for the trans material. Does anyone here know where I may obtain data for such light fixtures and how to correctly translate this data into Radiance?

I have some additional questions concerning the trans material type. In a Radiance reference manual, it is stated:

"Trans is a translucent material, similar to plastic. The transmissivity is the fraction of penetrating
light that travels all the way through the material. The transmitted specular component is the fraction of transmitted light that is not diffusely scattered. Transmitted and diffusely reflected light is modified by the material color. Translucent objects are infinitely thin."

First, should the transmissivity + transmitted specular component = 1 in order for the material to be physically correct? Also, since the transmitted specular component is the fraction of transmitted light that is not diffusely scattered, does this mean that this is apart of the direct calculation or is this too apart of the indirect calculation.

Thanks

Marcus

Hi Marcus,

Welcome to the wonderful, confusing world of "trans".

I don't know of any good way to measure oddly-shaped translucent materials. Perhaps others on the list will have suggestions for that....

I have some additional questions concerning the trans material type. In a Radiance reference manual, it is stated:

"Trans is a translucent material, similar to plastic. The transmissivity is the fraction of penetrating
light that travels all the way through the material. The transmitted specular component is the fraction of transmitted light that is not diffusely scattered. Transmitted and diffusely reflected light is modified by the material color. Translucent objects are infinitely thin."

First, should the transmissivity + transmitted specular component = 1 in order for the material to be physically correct? Also, since the transmitted specular component is the fraction of transmitted light that is not diffusely scattered, does this mean that this is apart of the direct calculation or is this too apart of the indirect calculation.

The reason the trans parameters are so peculiar is so their valid range is easy to understand -- between 0 and 1 for everything except roughness. (For physical value ranges, you may also consult the file "ray/doc/notes/materials".) Pages 325 and 326 from Chas' chapter in "Rendering with Radiance" are also invaluable. There have been a number of threads on trans over the years, and this is definitely one for the Radiance FAQ, should it ever materialize -- no pun intended.

-Greg

"Unlike most other Radiance material primitives, the trans material is neither intuitive nor straightforward to apply."
- Charles Ehrlich

Marcus Jacobs wrote:

Dear Group

I am modeling a bathroom that will use translucent light fixtures similar to this:

I wish to accurately simulate the light distribution from the fixtures however I do not know where to begin as far as determining physically valid parameters for the trans material. Does anyone here know where I may obtain data for such light fixtures and how to correctly translate this data into Radiance?

Hi Marcus,

Unfortunately this is one of those areas where the best course of action is some measurement, followed with some trial and error. Trans can effectively model this material, however. I did a bunch of measurements of a translucent glazing several years ago, to obtain the transmittance at the normal as well as every 15-degrees off-axis. A trans material was then created and progressively modified until the curves for the trans and the measurements aligned pretty well. This will be fairly easy, assuming you can get a flat glass sample of those tulip-shaped shades. (Full disclosure: Greg helped me fit the data to a valid trans description, which is a large part of why it was fairly easy.) Measuring the actual shade itself would be a real bear, I'd expect.

I have some additional questions concerning the trans material type. In a Radiance reference manual, it is stated:

"Trans is a translucent material, similar to plastic. The transmissivity is the fraction of penetrating
light that travels all the way through the material. The transmitted specular component is the fraction of transmitted light that is not diffusely scattered. Transmitted and diffusely reflected light is modified by the material color. Translucent objects are infinitely thin."

First, should the transmissivity + transmitted specular component = 1 in order for the material to be physically correct?

The best references for trans are the flowchart at Schorsch's site, (http://www.schorsch.com/rayfront/manual/transdef.html) and pp 325-326 in RwR.

Good luck; I know Visarc Jack did some courthouse renderings a long time ago that featured these large alabaster pendant bowl-like shades, and they looked amazing; the trans really captured the soft lamp image quite well. Maybe Jack can share what he did? Even if he measured the actual shade, they had a fairly large radius to them. Your tulip shades will be a real challenge, I think.

- Rob

Hey Guys,

Thanks for the insight. First, concerning my oddly shaped lighting fixtures, the ones that I have modeled actually look like this:

So I think they are actually more complicated that the example that I showed to you before. Out of curiosity, just as many modeling software estimate a complex surface as a mesh of planes or triangles, could the same principle be used for the lighting fixtures themselves. In other words, instead of seeing each fixture as a single complex surface, can it be considered it as just a group of flat planes? One item that I need to have addressed is whether I should model the diffuse shades a single surface or if it acceptable to have inner and outer surfaces (you can think of it as having a thickness to it). See here:

The information in Rayfront's user manual and pages 325-326 from RwR were VERY useful to me in gaining insight as to how Radiance treats light. There was a word that was mentioned that has me worried. Measure. Aside from not having any experience with performing a empirical test of light distribution, measuring light sounds very expensive. May I ask, would purchasing the needed equipment require a second mortgage to be taken out or giving up one's first born? Is it required to use transfunc or transdata in lieu of trans to obtain acceptable accuracy?

Thanks,

Marcus

···

From: "Gregory J. Ward"

Hi Marcus,

Welcome to the wonderful, confusing world of "trans".

I don't know of any good way to measure oddly-shaped translucent
materials. Perhaps others on the list will have suggestions for
that....

> I have some additional questions concerning the trans material
> type. In a Radiance reference manual, it is stated:
>
> "Trans is a translucent material, similar to plastic. The
> transmissivity is the fraction of penetrating
> light that travels all the way through the material. The
> transmitted specular component is the fraction of transmitted light
> that is not diffusely scattered. Transmitted and diffusely
> reflected light is modified by the material color. Translucent
> objects are infinitely thin."
>
> First, should the transmissivity + transmitted specular component =
> 1 in order for the material to be physically correct? Also, since
> the transmitted specular component is the fraction of transmitted
> light that is not diffusely scattered, does this mean that this is
> apart of the direct calculation or is this too apart of the
> indirect calculation.

The reason the trans parameters are so peculiar is so their valid
range is easy to understand -- between 0 and 1 for everything except
roughness. (For physical value ranges, you may also consult the file
"ray/doc/notes/materials".) Pages 325 and 326 from Chas' chapter in
"Rendering with Radiance" are also invaluable. There have been a
number of threads on trans over the years, and this is definitely one
for the Radiance FAQ, should it ever materialize -- no pun intended.

-Greg

From: Rob Guglielmetti <rpg@rumblestrip.org>

Hi Marcus,

Unfortunately this is one of those areas where the best course of action
is some measurement, followed with some trial and error. Trans can
effectively model this material, however. I did a bunch of measurements
of a translucent glazing several years ago, to obtain the transmittance
at the normal as well as every 15-degrees off-axis. A trans material was
then created and progressively modified until the curves for the trans
and the measurements aligned pretty well. This will be fairly easy,
assuming you can get a flat glass sample of those tulip-shaped shades.
(Full disclosure: Greg helped me fit the data to a valid trans
description, which is a large part of why it was fairly easy.)
Measuring the actual shade itself would be a real bear, I'd expect.

>
> I have some additional questions concerning the trans material type.
> In a Radiance reference manual, it is stated:
>
> "Trans is a translucent material, similar to plastic. The
> transmissivity is the fraction of penetrating
> light that travels all the way through the material. The transmitted
> specular component is the fraction of transmitted light that is not
> diffusely scattered. Transmitted and diffusely reflected light is
> modified by the material color. Translucent objects are infinitely thin."
>
> First, should the transmissivity + transmitted specular component = 1
> in order for the material to be physically correct?

The best references for trans are the flowchart at Schorsch's site,
(http://www.schorsch.com/rayfront/manual/transdef.html) and pp 325-326
in RwR.

Good luck; I know Visarc Jack did some courthouse renderings a long time
ago that featured these large alabaster pendant bowl-like shades, and
they looked amazing; the trans really captured the soft lamp image quite
well. Maybe Jack can share what he did? Even if he measured the actual
be a real challenge, I think.

- Rob

So I think they are actually more complicated that the example that I showed to you before. Out of curiosity, just as many modeling software estimate a complex surface as a mesh of planes or triangles, could the same principle be used for the lighting fixtures themselves. In other words, instead of seeing each fixture as a single complex surface, can it be considered it as just a group of flat planes? One item that I need to have addressed is whether I should model the diffuse shades a single surface or if it acceptable to have inner and outer surfaces (you can think of it as having a thickness to it). See here:

http://i35.photobucket.com/albums/d178/marcdevon/luminaireview.jpg

Hi Marcus,

Without actual photometric data (such as an ies file) to apply to your group of flat planes, you have to model the actual comple geometry. But I'm afraid I may have misunderstood your question. As for the single/dual surface question, I guess it depends on how close your views will bring you to the shade. The trans material is infinitely thin, and it'd be best to model the thng as a single surface. For closeup work, you may need to introduce thickness but that's gonna get a lot more complicated from a valid material definition standpoint.

The information in Rayfront's user manual and pages 325-326 from RwR were VERY useful to me in gaining insight as to how Radiance treats light. There was a word that was mentioned that has me worried. Measure. Aside from not having any experience with performing a empirical test of light distribution, measuring light sounds very expensive. May I ask, would purchasing the needed equipment require a second mortgage to be taken out or giving up one's first born? Is it required to use transfunc or transdata in lieu of trans to obtain acceptable accuracy?

Mmmm, measurements, yeah. The gear can be expensive. Let's see, at my old job, where I did these measurements, I used an illuminance meter, a luminance meter and a theatrical spotlight (ETC Source Four HID, aka "leko", or ERS). The light fixture was a free sample, but you could use other sources too. The illuminance meter (Minolta Somethingorother-nowdiscontinued) was about \$800 and the luminance meter (Minolta LS-100) was more expensive, I believe almost three thousand USD. Actually, here are a couple of links I just found for the meters I used (the illuminance meter I link to here is a newer model than I used, but it's basically the same thing).

It's expensive not to mention time-consuming, but it's also fun and is an excellent example of the power of Radiance, in that Radiance is happy to accept material data as complex as you are willing to gather.

Maybe Peter can chime in with his thoughts on this stuff, and maybe an update on his goniophotometer project?

- Rob

···

On Jun 3, 2006, at 8:30 AM, Marcus Jacobs wrote:

Just to add to what Rob said. I know that it is possible to arrange for equipment rentals directly from KonicaMinolta. This especially makes sense for the luminance meter which is indeed ~\$3,000 at retail. Also on the spot light that Rob mentioned, good camera equipment stores that are set up to provide rental equipment to professionals will have a wide selection of lighting. I know that for one exercise I rented a small spotlight where you could adjust the beam spread (sorry I do not remember the name/type) at a reasonable price.

···

On Jun 3, 2006, at 8:30 AM, Marcus Jacobs wrote:

So I think they are actually more complicated that the example that I showed to you before. Out of curiosity, just as many modeling software estimate a complex surface as a mesh of planes or triangles, could the same principle be used for the lighting fixtures themselves. In other words, instead of seeing each fixture as a single complex surface, can it be considered it as just a group of flat planes? One item that I need to have addressed is whether I should model the diffuse shades a single surface or if it acceptable to have inner and outer surfaces (you can think of it as having a thickness to it). See here:

http://i35.photobucket.com/albums/d178/marcdevon/luminaireview.jpg

Hi Marcus,

Without actual photometric data (such as an ies file) to apply to your group of flat planes, you have to model the actual comple geometry. But I'm afraid I may have misunderstood your question. As for the single/dual surface question, I guess it depends on how close your views will bring you to the shade. The trans material is infinitely thin, and it'd be best to model the thng as a single surface. For closeup work, you may need to introduce thickness but that's gonna get a lot more complicated from a valid material definition standpoint.

The information in Rayfront's user manual and pages 325-326 from RwR were VERY useful to me in gaining insight as to how Radiance treats light. There was a word that was mentioned that has me worried. Measure. Aside from not having any experience with performing a empirical test of light distribution, measuring light sounds very expensive. May I ask, would purchasing the needed equipment require a second mortgage to be taken out or giving up one's first born? Is it required to use transfunc or transdata in lieu of trans to obtain acceptable accuracy?

Mmmm, measurements, yeah. The gear can be expensive. Let's see, at my old job, where I did these measurements, I used an illuminance meter, a luminance meter and a theatrical spotlight (ETC Source Four HID, aka "leko", or ERS). The light fixture was a free sample, but you could use other sources too. The illuminance meter (Minolta Somethingorother-nowdiscontinued) was about \$800 and the luminance meter (Minolta LS-100) was more expensive, I believe almost three thousand USD. Actually, here are a couple of links I just found for the meters I used (the illuminance meter I link to here is a newer model than I used, but it's basically the same thing).

It's expensive not to mention time-consuming, but it's also fun and is an excellent example of the power of Radiance, in that Radiance is happy to accept material data as complex as you are willing to gather.

Maybe Peter can chime in with his thoughts on this stuff, and maybe an update on his goniophotometer project?

- Rob

_______________________________________________

--
# Jack de Valpine
# president
#
# visarc incorporated
# http://www.visarc.com
#
# channeling technology for superior design and construction

Thanks

Although I would like to avoid the time and expense of performing this analytical work, I am thinking that it may be necessary. First, is there any existing data that is readily available for translucent materials that are used as diffusing elements for light fixtures? Also, are there any specifications or procedures as to how to perform these tests?

Thanks

Marcus

···

From: Jack de Valpine

Just to add to what Rob said. I know that it is possible to arrange for
equipment rentals directly from KonicaMinolta. This especially makes
sense for the luminance meter which is indeed ~\$3,000 at retail. Also on
the spot light that Rob mentioned, good camera equipment stores that are
set up to provide rental equipment to professionals will have a wide
selection of lighting. I know that for one exercise I rented a small
spotlight where you could adjust the beam spread (sorry I do not
remember the name/type) at a reasonable price.

>
> On Jun 3, 2006, at 8:30 AM, Marcus Jacobs wrote:
>> So I think they are actually more complicated that the example that I
>> showed to you before. Out of curiosity, just as many modeling
>> software estimate a complex surface as a mesh of planes or triangles,
>> could the same principle be used for the lighting fixtures
>> themselves. In other words, instead of seeing each fixture as a
>> single complex surface, can it be considered it as just a group of
>> flat planes? One item that I need to have addressed is whether I
>> should model the diffuse shades a single surface or if it acceptable
>> to have inner and outer surfaces (you can think of it as having a
>> thickness to it). See here:
>>
>> http://i35.photobucket.com/albums/d178/marcdevon/luminaireview.jpg
>
> Hi Marcus,
>
> Without actual photometric data (such as an ies file) to apply to your
> group of flat planes, you have to model the actual comple geometry.
> But I'm afraid I may have misunderstood your question. As for the
> single/dual surface question, I guess it depends on how close your
> views will bring you to the shade. The trans material is infinitely
> thin, and it'd be best to model the thng as a single surface. For
> closeup work, you may need to introduce thickness but that's gonna get
> a lot more complicated from a valid material definition standpoint.
>
>>
>> The information in Rayfront's user manual and pages 325-326 from RwR
>> were VERY useful to me in gaining insight as to how Radiance treats
>> light. There was a word that was mentioned that has me worried.
>> Measure. Aside from not having any experience with performing a
>> empirical test of light distribution, measuring light sounds very
>> expensive. May I ask, would purchasing the needed equipment require a
>> second mortgage to be taken out or giving up one's first born? Is it
>> required to use transfunc or transdata in lieu of trans to obtain
>> acceptable accuracy?
>
> Mmmm, measurements, yeah. The gear can be expensive. Let's see, at
> my old job, where I did these measurements, I used an illuminance
> meter, a luminance meter and a theatrical spotlight (ETC Source Four
> HID, aka "leko", or ERS). The light fixture was a free sample, but
> you could use other sources too. The illuminance meter (Minolta
> Somethingorother-nowdiscontinued) was about \$800 and the luminance
> meter (Minolta LS-100) was more expensive, I believe almost three
> thousand USD. Actually, here are a couple of links I just found for
> the meters I used (the illuminance meter I link to here is a newer
> model than I used, but it's basically the same thing).
>
> http://www.tequipment.net/MinoltaTL1.asp
> http://www.tequipment.net/MinoltaLS100.asp
>
> It's expensive not to mention time-consuming, but it's also fun and is
> an excellent example of the power of Radiance, in that Radiance is
> happy to accept material data as complex as you are willing to gather.
>
> Maybe Peter can chime in with his thoughts on this stuff, and maybe an
> update on his goniophotometer project?
>
> - Rob
>
> _______________________________________________