Coated glass from real measured spectral data

General discussion about Indigo Materials - material requests, material developement, feedback, etc..
User avatar
galinette
1st Place Winner
Posts: 923
Joined: Sat Jan 09, 2010 1:39 am
Location: Nantes, France
Contact:

Coated glass from real measured spectral data

Post by galinette » Sat Jan 09, 2010 11:18 am

Dear all,

I would like to simulate a coated glass material with following characteristics:
- The medium of the glass has a given index of refraction (for instance 1.52). Deviation of light by refraction should work exactly as the specular material does.
- But the reflectance/transmittance coefficients at the interface should not be governed by the medium refractive index (what specular does) but by a tabulated reflectance/transmittance angular spectral data (reflectance and transmittance values for each wavelength/angle_of_incidence pair)

I have the feeling that it is currently not possible from what I can see in the documentation. I do not want an approximative solution (such as blended phong and specular materials) which could give a similar look. I would like to do exactly the above function.

This simulates exactly the behavior of advanced coated glass, such as the ones you may see on modern skyscraper. It is made of standard glass on which an interferential system of several thin films is deposited.
The colors are highly angle dependant with a complex spectrum.

Thanks for any help!

Etienne
Eclat-Digital Research
http://www.eclat-digital.com

User avatar
fused
Developer
Posts: 3648
Joined: Fri Sep 22, 2006 7:19 am
Location: Berlin, Germany
3D Software: Cinema 4D

Re: Coated glass from real measured spectral data

Post by fused » Sat Jan 09, 2010 1:14 pm

Hey,

No its not possbile currently, but it sounds like a pretty cool and unique feature so...

+1

:)

User avatar
OnoSendai
Developer
Posts: 6241
Joined: Sat May 20, 2006 6:16 pm
Location: Wellington, NZ
Contact:

Re: Coated glass from real measured spectral data

Post by OnoSendai » Sat Jan 09, 2010 1:20 pm

Hi Galinette,
That's a very interesting request.
Such a request is definitely not possible with the current Indigo code.
The new 'absorption layer transmittance' specular material parameter is somewhat similar, but only acts as a multiplier on the transmitted light (e.g. on the transmission coefficient).
So it doesn't affect the reflection coefficient at all.

Are there any absorption effects in such coated glass? Presumably not, in which case the reflection coefficient is simply 1 - T where T is the transmission coefficient.

I think your request should be possible to add to Indigo. Do you have any example tabulated data that you could send to me?
Cheers,
nik

User avatar
galinette
1st Place Winner
Posts: 923
Joined: Sat Jan 09, 2010 1:39 am
Location: Nantes, France
Contact:

Re: Coated glass from real measured spectral data

Post by galinette » Sat Jan 09, 2010 1:47 pm

Hi,

Yes, there is definitely absorption, which is also spectroscopic... I have all the tabulated datas.

You can find a picture here illustrating this:
http://pt.saint-gobain-glass.com/newsle ... _home.html

You can see the color & reflectance angle dependance quite well. This gives the subtle shades you see on more and more towers in the world (all those that have energy efficient glass façades, and many are constructed like that nowadays!).

One of the simple way to do this, I think, would be implementing a "advanced_specular" material, where you can override the T & R coefficients (calculated with snell-descartes laws in the current implementation) by some 2D tabulated data (angle & wavelength).

Also, one interesting feature would be not only providing table data for the absorbance, but also for the refractive index. Cauchy law is somewhat restrictive for people like me working in optical engineering.

Thanks for your interest in this!

Etienne
Eclat-Digital Research
http://www.eclat-digital.com

User avatar
OnoSendai
Developer
Posts: 6241
Joined: Sat May 20, 2006 6:16 pm
Location: Wellington, NZ
Contact:

Re: Coated glass from real measured spectral data

Post by OnoSendai » Sat Jan 09, 2010 1:54 pm

Yup, tabulated refractive indices would be interesting too.
Can you send me some of the glass data to nick@indigorenderer.com ?
Cheers.

User avatar
galinette
1st Place Winner
Posts: 923
Joined: Sat Jan 09, 2010 1:39 am
Location: Nantes, France
Contact:

Re: Coated glass from real measured spectral data

Post by galinette » Sun Jan 10, 2010 12:40 am

For sodalime glass (ie glass used in all automotives and buidlings, including tinted glasses), n is given in the table below. The k coefficient depends on tint. Here it is given for average clear iron glass (Fe around 900ppm), which is the most common glass, on which you may see green tint on the side or when it is very thick.

wl(nm) n k
360 1.543 23.86277101
365 1.5425 22.22413242
370 1.542 20.59999759
375 1.541 19.01408062
380 1.54 17.44572974
385 1.5395 15.14090736
390 1.539 12.86841071
395 1.538 10.64965751
400 1.537 8.461423553
405 1.5365 8.273130014
410 1.536 8.085092169
415 1.5355 7.898185251
420 1.535 7.711549483
425 1.5345 7.728222614
430 1.534 7.744907501
435 1.5335 7.762459883
440 1.533 7.779985927
445 1.5325 7.364522819
450 1.532 6.950199301
455 1.5315 6.537858321
460 1.531 6.126640415
465 1.5305 5.796081583
470 1.53 5.466258807
475 1.5295 5.137150208
480 1.529 4.808752747
485 1.529 4.687608127
490 1.529 4.566552455
495 1.5285 4.467378604
500 1.528 4.368283218
505 1.5275 4.306099367
510 1.527 4.243948866
515 1.527 4.160054105
520 1.527 4.076202009
525 1.5265 4.103121828
530 1.526 4.130015908
535 1.526 4.134305348
540 1.526 4.13857633
545 1.5255 4.401756985
550 1.525 4.665326977
555 1.525 4.908296889
560 1.525 5.151643963
565 1.5245 5.542124958
570 1.524 5.933414212
575 1.524 6.304624649
580 1.524 6.67669115
585 1.5235 6.991301153
590 1.523 7.306430858
595 1.523 7.599436995
600 1.523 7.89298337
605 1.5225 8.627979544
610 1.522 9.366098877
615 1.522 10.08641033
620 1.522 10.80987905
625 1.522 11.34626447
630 1.522 11.88441787
635 1.5215 12.44697993
640 1.521 13.01132253
645 1.521 13.67321793
650 1.521 14.33775734
655 1.521 15.00496185
660 1.521 15.67485281
665 1.5205 15.93491888
670 1.52 16.19532594
675 1.52 16.43519528
680 1.52 16.67541343
685 1.52 17.3870192
690 1.52 18.10172355
695 1.52 18.81945193
700 1.52 19.54033191
705 1.5195 20.62998768
710 1.519 21.72657921
715 1.519 22.80731563
720 1.519 23.89514329
725 1.519 24.17230875
730 1.519 24.44996061
735 1.519 24.72803753
740 1.519 25.00658293
745 1.5185 26.36964095
750 1.518 27.7436595
755 1.518 29.10678542
760 1.518 30.48123793
765 1.518 31.06331289
770 1.518 31.6474211
775 1.518 32.23456543
780 1.518 32.82382276
785 1.5175 33.35918943
790 1.517 33.89617626
795 1.517 34.41284197
800 1.517 34.93114804
805 1.517 35.40063405
810 1.517 35.87145559
815 1.517 36.34364274
820 1.517 36.81715817

I checked, the n values fit pretty well in a Cauchy law in that case. But for other materials it does not.

For coated glass, I will e-mail you some measured data (interface R/T spectra) but I need to check the intellectual property issues first.

Thanks for support!

Etienne
Eclat-Digital Research
http://www.eclat-digital.com

User avatar
OnoSendai
Developer
Posts: 6241
Joined: Sat May 20, 2006 6:16 pm
Location: Wellington, NZ
Contact:

Re: Coated glass from real measured spectral data

Post by OnoSendai » Sun Jan 10, 2010 9:42 pm

I'm eagerly awaiting the coated glass data!
For the sodalime glass, such a material should already be able to be created in Indigo quite accurately, because Indigo supports tabulated specification of absorption coefficient.

User avatar
galinette
1st Place Winner
Posts: 923
Joined: Sat Jan 09, 2010 1:39 am
Location: Nantes, France
Contact:

Re: Coated glass from real measured spectral data

Post by galinette » Sun Jan 10, 2010 10:06 pm

For the non-coated glass with several tints, I'm already doing the test renders. I'll post materials this week.

The standard glass is really nice looking with spectral data. Look at the subtle greenish shades.

Spectral data will really make the difference when looking at the trench of a large piece of flat glass (such as 1m wide / 6mm thick) when you see strong green tint. It's impossible to reproduce this accurately with RGB data.

Etienne
Attachments
test_glass_PLX.jpg
Test render with glass material designed from spectral data measured on real glass (sodalime - standard Fe content)
test_glass_PLX.jpg (100.26 KiB) Viewed 16889 times
Eclat-Digital Research
http://www.eclat-digital.com

User avatar
Zom-B
1st Place 100
Posts: 4700
Joined: Tue Jul 04, 2006 4:18 pm
Location: ´'`\_(ò_Ó)_/´'`
Contact:

Re: Coated glass from real measured spectral data

Post by Zom-B » Mon Jan 11, 2010 4:23 am

I jut tried to use the tabulated data 1:1 for the absorption Spectrum of a Specular Material. My result is quite strong tinted. Do I have to modify the data (devide by 100 or something) to get get a propper result?!
sodalime.jpg
Default Test Scene Whitebalance of E
sodalime.jpg (118.21 KiB) Viewed 16842 times


**edit**

ok, I changed the white balance from preview scene default E to F7, so the result is nicer.
WB Setting E realy tints the image green (as you can see in the BG ramp)...
Hell knows why it is the default in the Preview scene...
sodalime2.jpg
Whitebalance set to f7
sodalime2.jpg (111.27 KiB) Viewed 16822 times
polygonmanufaktur.de

User avatar
galinette
1st Place Winner
Posts: 923
Joined: Sat Jan 09, 2010 1:39 am
Location: Nantes, France
Contact:

Re: Coated glass from real measured spectral data

Post by galinette » Mon Jan 11, 2010 11:09 am

Nice looking!

The data is to be used "as provided". I'ts already in m^(-1) units and I'm 100% sure of it.
I'm not surprised a 10cm sphere looks tinted anyway. This glass is used for windows, thickness is generally 6mm (or 2*6mm for insulating glazing units), and I know the tint is already sightly perceptible. So for a 10cm sphere, it has to look tinted.

I think your model is OK. Just check the refractive index coefficient : good values are 1.51 and 0.00424672.

More glass data to come soon...

Etienne
Eclat-Digital Research
http://www.eclat-digital.com

User avatar
OnoSendai
Developer
Posts: 6241
Joined: Sat May 20, 2006 6:16 pm
Location: Wellington, NZ
Contact:

Re: Coated glass from real measured spectral data

Post by OnoSendai » Mon Jan 11, 2010 11:17 am

Looks good. Can you upload to the MatDB Zom-B?

User avatar
Zom-B
1st Place 100
Posts: 4700
Joined: Tue Jul 04, 2006 4:18 pm
Location: ´'`\_(ò_Ó)_/´'`
Contact:

Re: Coated glass from real measured spectral data

Post by Zom-B » Mon Jan 11, 2010 12:12 pm

The 0.00424672 is meant to be the cauchy B parameter, right?!

Rendering a preview image and uploading when done!
polygonmanufaktur.de

User avatar
Zom-B
1st Place 100
Posts: 4700
Joined: Tue Jul 04, 2006 4:18 pm
Location: ´'`\_(ò_Ó)_/´'`
Contact:

Re: Coated glass from real measured spectral data

Post by Zom-B » Mon Jan 11, 2010 2:04 pm

OK, I uploaded the material to the Database: http://www.indigorenderer.com/materials/materials/250

tanks a lot for the Data galinette, I'll use it for sure :)

looking forward to more Tabulated data, and having tabulated IOR would be a nice feature too master Ono :wink:
polygonmanufaktur.de

User avatar
OnoSendai
Developer
Posts: 6241
Joined: Sat May 20, 2006 6:16 pm
Location: Wellington, NZ
Contact:

Re: Coated glass from real measured spectral data

Post by OnoSendai » Mon Jan 11, 2010 2:08 pm

Thanks Zom-B!

User avatar
Jambert
Indigo 100
Posts: 545
Joined: Wed Jul 30, 2008 10:08 pm
Location: France, Lyon

Re: Coated glass from real measured spectral data

Post by Jambert » Mon Jan 11, 2010 7:03 pm

Zom-B wrote:I jut tried to use the tabulated data 1:1 for the absorption Spectrum of a Specular Material. My result is quite strong tinted. Do I have to modify the data (devide by 100 or something) to get get a propper result?!
sodalime.jpg


**edit**

ok, I changed the white balance from preview scene default E to F7, so the result is nicer.
WB Setting E realy tints the image green (as you can see in the BG ramp)...
Hell knows why it is the default in the Preview scene...
sodalime2.jpg
:shock: impressive :shock:
realy good work thanks all :)

Post Reply
29 posts

Who is online

Users browsing this forum: No registered users and 9 guests