[update] translucent tests - flesh & juice

[flipside]

Hi!

Here's a flesh test after I tried boring 3D vray flesh material:
(model is from 3d cafe or turbosquid, don't remember) Don't mind rendertimes as they are rendered while I was doing tons of other testrenders.





Below are some juice tests in different lighting situations. No material settings were tweaked (at least not very much)! I took some pictures similar to these setups and juice comes out pretty good if I compare the renders with them. GI is ON everywhere.


Skylight+vraylight(plane)+box with vraylightmaterial


Skylight only


Trying to recreate flashlight pointed at the glass+caustics (notice the artifacts at the edges)


Same but no caustics (affect shadows for fog is on)


Flashlight in low position+caustics


Flashlight from the top (don't remember if I had caustics on here)




Artifacts on the juices edge are usually because of lightmapping. I'm doing more test on how to get rid of the dark border around the juice edge (although it's also there in the real life, but not in such a weird way).

Caustics seem to work much better than before! The max density parameter works great.

wouter

[TutanMark]

Hi Flipside, nice tests!

I love the juice shader... maybe it seemes a little too dense, but is only a feeling.

Very interesting.
Marco

[tom schuelke]

HI flipside,

would you mind posting the settings or the scene ?

i allways wanted to try this , but never found the time.

Tom

[flipside]

There are a lots of different juices, which all can look very very different. This one is a medium priced orange juice . Real fresh juice would be more orange, cheaper juice would be more transparant hehe.

Here's the material settings. The reflect on backside and double sided option actually have no effect at all turned on or off. If you want another color, simply change the hue slider of the fog and adjust the saturation of it a little bit. The diffuse color can be anything, it has no effect because refraction is pure white.

Notice how the glass/liquid is modeled (inner glass surface is removed where the liquid is)! This is important because when you model the liquid slightly larger than the glass innner surface, the juice gets much darker. This method also helps removing GI artifacts around the edge of the juice area.



Adjust settings carefully because it's a very tricky material. You can go on for hours testing with it. Different glossy/fog multipliers/colors can have the same visual result.

But this material works quite well under very different light setups, and only thing to change for color adjustment is the fog! That was my initial goal when trying to create this material.

I can't give you the scene though...

The flashlight is a very bright spotlight with inverse square falloff.

If you don't have many glasses behind each other, lower the max refraction depth and use the exit color to make the material lighter/darker, this will save some rendertime and is the easiest way to adjust the brightness (because fiddling with the fog color and multiplier gets you lost, I mean it gets me lost )

[percydaman]

great work flipside! I love messing around with SSS materials. Too bad vrays SSS is so slow...

[Gijs]

I particulary like the low flashlight version

nice work

[flipside]

here's one with different colors:



Rendertimes go crazy on this one. This is with store wit IR map for the vraylight turned on. When I turn this off, rendertime is 3h50min... (at this resolution!).

[vince_1972]

Wouha,

your last picture is fantastic !

Vince

[siliconbauhaus]

I'm thirsty now

[Da_elf]

me too. damn you. mmm. the green one looks interesting.

[tom schuelke]

Maybe put some candles with aura in between ....

[DaForce]

Thats pretty damn cool!!!

[TutanMark]

Hi Flipside,

WOW... that's really Cool!!

Just rendertimes are really ... .. but i see that u used 20 glossy refraction subdivisions . Considering that is juice... maybe lowering that value could help.

Marco

[flipside]

Hey tutan,

I'm also using QMC AA, so going as low as 4-8 like you would do with adaptive subdivision AA is not a good idea. In fact, 8 subs for example render slower than 20 with QMC AA in most cases

The effect on rendertime with lower glossies is less with adaptive QMC AA as with adpative subdivision AA.

I also used 0.001 for qmc threshold, for maximum AA quality on the glass. I guess this is just a very heavy scene, there's nothing but reflections & refractions & glossies & translucency Not to forget GI, I have refraction GI caustics turned on which doensn't make it easier I think...

[Da_elf]

was going through my hard drive and came across this old pic i did when 3s was making its early steps in vray. just a hand i found somewhere was so long ago i cant remember. it wasnt human hence the spikes. and i got a set of bones to add in from somewhere too