Page 3 of 3

Re: Dealing with this sort of reflection problem

Posted: Wed Sep 23, 2020 6:36 am
by bepeg4d
As Jay said, glass this liquid is an old issue for all unbiased render engines. C4D internal render engine, or RedShift, are biased and can cheat in this scenario.
To simplify the complexity of glass with liquid, we are used to slightly overlap the liquid inside the thickness of the glass.
This is giving a more correct rendering result of the refraction effect, but it is not properly correct.
Here is a diagram on how you should need to model a liquid inside glass (without nested dielectric), for any unbiased physically based render engine, not only OctaneRender:
F5644106-3A71-4F7B-B04C-6F100A9D6D67.jpeg
ciao Beppe

Re: Dealing with this sort of reflection problem

Posted: Wed Sep 23, 2020 1:57 pm
by mitchino
Can you explain how you would model that? Is each section made of a single 'skin' of polys? i.e no closed volumes?

Re: Dealing with this sort of reflection problem

Posted: Thu Sep 24, 2020 6:59 am
by bepeg4d
Basically you need to cut the internal glass where meets the liquid, and invert its normal pointing outward, to mimic the liquid that touches the glass, with IOR = 1.53/1.33 = 1.15.
Then add the top part of the liquid as single surface with normal pointing upward, to mimic the glass that touches the air, with IOR = 1.33.
In this way, you have the volume generated by three open surfaces, but with different IOR values, since Octane does not change the behavior of the ray, until it touches another surface with different IOR.
It is the same principle of the sphere around the Camera, with inverted normals, IOR =1, and Scattering Medium, for mimic the fog.
F8DB709D-2F3A-4297-A23F-DB100AA441C1.jpeg
GlassLiquidScheme.c4d.zip
(97 KiB) Downloaded 123 times
Obviously, this solution works great only for static liquids, for animated liquids better to use the overlapping solution between glass and liquid.

ciao Beppe