[Mihai]

Even with the current single layer sss it should be possible to create almost transparent surfaces such as thin curtains, but it seems the scattering parameter doesn't work properly, or is too strong.

I thought if you set it to 0.01, you would get almost no scattering effect and the surface would look mostly transparent. But there is very little difference between 1 and 0.01 in a 1mm thick cube. It always assumes a very dense medium. Perhaps this is what's missing for now?

[x_site]

:: could it be scale problem... or i am just talking rubbish ::

[Aldaryn]

Thin, almost transparent surfaces are impossible right now with sss.
You'll get far further with a dielectric with high UV roughness, I think.

I think the main problem is that you cant really separate transparency and transluciency. Every material in real life is transparent, if you manage to cut it really thin, even metals, and dense minerals. And every material is a bit scattering medium also, even glass.

If the scattering object is fairly thin, you should see refraction, and transmitted light, but if so, than this would mean, that you need an IOR value, transmitted color, and other shader properties, so, the best vay to solve this, is to merge a "Scattering shader" and the Dielectric shader. And this leasd to a standardization of all available shaders (expect emittter) into one super shader, which can accurately model any type of material ranging from glasses, through jelly like shader to diffuse ones.

[mtripoli]

I don't think so! Have a look at the first image here:
http://www.maxwellrender.com/forum/viewtopic.php?t=3944

This is a 1m square screen, 0.1mm thick! THe tree object is back-lit, shadow projected on the screen with sss applied (diffuse, ab-scattering 0.1 each).

Am I missing something? Plus, in a different thread, Frances shows curtains using sss...

http://www.maxwellrender.com/forum/viewtopic.php?t=3716


Mike Tripoli

[Aldaryn]

Really great tests mtripoli! Convincing!

However, I still think, no matter how thin your surface, it will never show transparent like effects. (For example, a blue ball behind a really thin sss surface would appear as a black shadow, not as a blue, blurred spot)

Please, someone prove me that Im wrong! It would be great...

[mtripoli]

Really great tests mtripoli! Convincing!

However, I still think, no matter how thin your surface, it will never show transparent like effects. (For example, a blue ball behind a really thin sss surface would appear as a black shadow, not as a blue, blurred spot)

Please, someone prove me that Im wrong! It would be great...

I hate to sound stupid but do you want a black shadow or a blue blurred spot? I don't know which is correct!

Mike Tripoli

[Aldaryn]

Both of them. Put a saturated object behind a paper, and backlit it, you will see the transmitted color of the object, and also the shadow of the object on the back side of the paper.
Also, I've just had a sandwich with cheese on it, and Ive cut the cheese really thin so I could clerly see my fingers (color, texture) through it, where it was hin eneugh.

[mtripoli]

OK, here's a test; blue ball behind a thin (0.01mm) "sheet". Ran for 4 minutes...


Is this correct?

Mike Tripoli

[Aldaryn]

Yes, somewhat correct, but to test that its incorrect, try to recreate a wax-paper like effect. The "problem" with the current sss is like you cant adjust how "blurred" the effect will render.

[tom]

I think this is something what you're looking for

[psanitra]

Nice Tom! that`s the proof!

[Frances]

I think this is something what you're looking for

Can you please provide your settings - otherwise your good results are just annoying.

[mtripoli]

I think this is something what you're looking for

Can you please provide your settings - otherwise your good results are just annoying.

I AGREE!

MT

[tom]

sorry....

Curtain
RGB=200,200,200
Absorbtion=0.0 (I don't want to hold the light propagation)
Scattering=1.0 (I want it go full through the media)

Ball
RGB=0,0,220
Absorbtion=0.0
Scattering=0.1 (I just want it to wash around with its diffuse color, actually blue)

In fact, I should not need to set scattering to the ball but i want to pronounce the casted color.
If you don't use it, you'll have less blue color, but you'll have..no way

[Mihai]

Thank you for testing this, but it's not what I was looking for. The blue ball is completely blurry. I know the color transmission works and that's nice, but the scattering in all cases is much too strong. It's like it always assumes a very dense sheet of paper. That's why I said this is missing for now, a parameter to set the density of the medium.

Logically, the more you raise the scattering, the more you want the rays to scatter, thus producing a blurrier image. That's why I don't understand when you say:

"Scattering=1.0 (I want it go full through the media)"