OTOY Forums

Hi Pgatsky, I have had a chance to look at your file. Thank you for posting it.

**Edit: As noted in a subsequent by Beppe, this requires Octane 2020.2 XB1 to properly resolve**

The issue is you do not want all of those reflections of the embossing. I do not know how many shots and from what angles you are looking to do, but my fix is to cheat and break things. Make the clear section of the bottle solid, not hollow. That way, you do NOT see the reflections of the embossing on the inside of the glass. You would see the reflection on the interior of the back of the bottle, but that would not produce the reflection that you are objecting to.

You WOULD still see the reflection of the embossing on the cognac mesh, however. In that case, you can set your index to 1, which would minimize that reflection. You would effectively just be tinting the mesh. Not physically correct, but the result I think you might be going for.

I have two other recommendations for you regardless:

1. Make sure your grayscale images are set to float. Currently, you have the displacement set to Normal, which assumes RGB space. You are wasting VRAM that way. It might not matter much in this scene, but it is a good practice.

2. Your glass material needs a roughness value greater than 0. In these cases, I typically use 0.001 or similar. Mess around with that a bit to get the look you are after. The reason that you do this is that no human-made material is perfectly smooth. In fact, the only thing in your scene which should have a roughness value of 0.0 is the cognac, as only liquids can achieve true smoothness.

Sorry I could not offer a simpler suggestion. Hope this helps!

Thank you again jayroth for your very informative reply, I really appreciate the effort you have shown to explain.

Hi,
since the issue is in refraction, due to nested dielectrics, to solve, you need to export to Standalone 2020.2-XB1, and render from there: ciao Beppe

Thank you Beppe I will look into that.

This subject always breaks my brain a little bit (nested dielectrics), but might the issue be the fact that the difference between indices of refraction of the back surface of the glass and the whisky is not the same as the difference between the glass and the air? Essentially, glass and whisky are much closer to each other than glass and air, and the resulting internal reflection off the glass/whisky interface should be significantly less than you get from glass/air. Geometry wise, you may want two surfaces, glass/air and glass/whisky, instead of glass, glass, whisky. I may be thinking too old-school, though, and I only have to deal with this issue once in a blue moon.

Thats deep frankmci but I get your thinking...I think?? Interesting insight there

Frank has a good point. The air gap could be very helpful here. Or, per Beppe, take it into the latest standalone. Obviously I need to amend my earlier post on this.

I think you've all had good points and been very helpful.

I think I get what Frank is saying but doesn't that go against how to build glass bottles and thickness? Shouldn't in theory the section between the outer part of the glass and the inner section been seen as a solid component rather than air? If I had taken out the inner wall of the glass then the object would just become one solid section of glass. The index of refrac for the glass is 1.52 and the liquid is 1.33 but does that mean that whats in between might be a lot less than the difference?

Why would a programme such as C4D get something like that so mixed up? When the standalone as Beppe has show will work fine? Is this something Maxon should be aware of? Or do you reckon they've got no motive to help OTOY users over something like this when they now have Redshift?

This all has to do with the physics of light, and the way to best simulate those physics in a software rendering program. Nested dielectrics is currently the state of the art in simulating the interaction that you have created with your scene, that of the interface of two different transparent mediums (glass and alcohol, in your case). Ray tracers have always been challenged by this interaction, and Frank mentioned the air gap idea. To further explain that, early ray tracers such as LightWave would not render the interior of a glass volume correctly unless you had a corresponding interior shape with an IOR value of 1.0 (that of air). Modern engines do not need the "air gap" typically, though as we have seen here, it is one way to address the situation when nested dielectrics are not available as a solution. Fortunately, as of Octane standalone 2020.2-XB1, nested dielectrics are supported and this issue becomes moot (once all of the plugins are updated to that version of Octane). For the time being, Standalone is the best way to go.

This is NOT an issue with Cinema 4D per se, as it is only a rendering issue. If rendering with the native Cinema engines it may be, but not in our case as Octane users.

Thanks again jayroth I'll be investigating the standalone more now.