[Stephen]

Needs to cook longer, but I think it's getting there... maybe. The dielectric is still quite noisy, and it's making it look a bit brown. That is going away though. This was SL16.


quick test I did on my slow machine, sorry for not making it look pretty, I had other work to do. I think it looks promising though.

I'll put it on my server over night once I get your file. Who knows, maybe it will come out well. I have to admit I haven't spent much time on materials. Mine are usually single Layer Lamberts.

I'm finding this material editor a great deal of fun. I'm glad something finally made me look deeper under the hood.

[Boris Ulzibat]

It's giving me access denied. I don't have permission

Corrected the permissions, sorry for this mess!

[Stephen]

no worries. Thanks.

[Stephen]

Hey Boris,
I'm afraid it didn't turnout too well. I really had a tough time with sss. It doesn't seem to work as I would expect it too. Also the HDR didn't transfer in the pack and go so I replaced a temporary one.

Mihai has something close, maybe you can get his mxm.
Good luck.

[michaelplogue]

Well, based off the suggestions from mverta and Mihai, I've come up with something a bit closer. Just by itself, it still didn't look right, so I went ahead and added some pulp, and it looks like something I'd might actually consider drinking.

My approach was very hodge-podge and I'm sure there's stuff in the material that I don't need or isn't really doing anything...

I ended up with three layers, using the additive mode as Mike suggested. One ( the 'juice' layer) has full roughness and the SSS applied. Another is a simple liquid layer (the 'gloss' layer), and the third is a lambert layer (to help fog things up a bit more). I applied a very thick coating on each layer - not sure if it's actually doing anything, but the end result looks better.

The pulp was made using p-array in Max (copy of the liquid mesh, with faces inverted as emitter). The pulp object had it's faces inverted as well (facing inwards). I used a similar material as the juice, but without the gloss layer. I also cheated a bit and added a very low emitter to the pulp material to fake SSS and make them stand out a bit better - though it certainly isn't helping the render times...

This is at SL 10, so it's still pretty noisy. I'll let it render for a few more hours to see how it clears up.

[Boris Ulzibat]

Well, based off the suggestions from mverta and Mihai, I've come up with something a bit closer. Just by itself, it still didn't look right, so I went ahead and added some pulp, and it looks like something I'd might actually consider drinking.

My approach was very hodge-podge and I'm sure there's stuff in the material that I don't need or isn't really doing anything...

I ended up with three layers, using the additive mode as Mike suggested. One ( the 'juice' layer) has full roughness and the SSS applied. Another is a simple liquid layer (the 'gloss' layer), and the third is a lambert layer (to help fog things up a bit more). I applied a very thick coating on each layer - not sure if it's actually doing anything, but the end result looks better.

The pulp was made using p-array in Max (copy of the liquid mesh, with faces inverted as emitter). The pulp object had it's faces inverted as well (facing inwards). I used a similar material as the juice, but without the gloss layer. I also cheated a bit and added a very low emitter to the pulp material to fake SSS and make them stand out a bit better - though it certainly isn't helping the render times...

This is at SL 10, so it's still pretty noisy. I'll let it render for a few more hours to see how it clears up.

How did you do the glass/liquid part? are they intersecting?

And what are the SSS settings you used?

[Thomas An.]

Well, we can push and shove the existing material system to fake something close, but even if it looks ok (it will still not be Kosher, to a purist).

The main issue appears to be a weak SSS control at this time.
If you take a look at this test:
http://www.maxwellrender.com/forum/view ... hlight=sss

The asorption parameter acts on the entire object evenly; causing a uniform tint (which IMHO is incorrect) ... we should see the center thicker region being darker first, and the corners/edges of the block should remain fairly clear.

The scattering parameter seems to work very nicely though ... (but we need both parameters in top condition in order to get best results)

In the past I did some SSS tests as well, but kinda gave up on it:

[michaelplogue]

The main issue appears to be a weak SSS control at this time.

Agree with you 100%. It shouldn't be this hard to get any decent result.

How did you do the glass/liquid part? are they intersecting?

And what are the SSS settings you used?

The liquid is a separate object, not intersecting the glass.

I'll try to upload the material so you can take a look at it. The SSS settings were Absorption= .8, Scatter = .9 (it really probably shouldn't be this, but I couldn't get any decent results with the current SSS implementation - which as Thomas states - is very weak right now).

[michaelplogue]

Heres the mxm I used for the juice:

http://www.divshare.com/download/14283-794

[glypticmax]

If Michael and Thomas can't work this out with relative ease, there's a weak spot in the software.
If that's not the case, let's have some of the NL folks weight in.
I thought diamonds were hard for a render engine, but when Boris started the orange juice thread, I quickly figured out he had really come up with a very tough nut to crack.
I look forward to seeing how you guys solve this one.
sign me clueless but interested.

[michaelplogue]

Here's the same material using a different lighting setup: Two emitters on each side, and HDR environment.

I'm calling this one done: it's as close as I can come up with, and this would be something that I'd consider drinking (I've actually consumed worse looking stuff than this! ). Modeling sucks, but the material works for me... However, I really think the pulp really makes the difference here. Without it, it just doesn't look as real.. ...

[Hervé]

It could be Tropicana... the texture is good, but I'd say the feel is more of some oily thing (but that's modeling maybe..) the pulp makes indeed the difference.. maybe more pulp..

[Boris Ulzibat]

The liquid is a separate object, not intersecting the glass.

That's it! When the liquid is not intersecting with the glass, everyting looks fine, except that the liquid/glass part doesn't look real.
When you make them intersecting - all the lambertian/SSS effects are gone - you can see a tinted transparent dielectric.