7
$\begingroup$

Please see the both renders below with limited and full global illumination, I like the way full global illumination brighten up the things and details, but I am having weird reflections on the glass surfaces? is this behaviour normal and correct realistic result or i am doing something wrong? thanks for the help

enter image description here

Blender file here,

I am using cycles, Subsurface modifier at 3 for the glass objects to render

$\endgroup$
  • $\begingroup$ Sorry, I wasn't able to reproduce such reflections in my test file. Could you upload your blendfile (or a part of it) to inspect it? I never saw a similar effect in real life, usually you see max 1 reflex per interface...but there is always a first time. Anyway it seems to me that you are not dealing with interfaces in a stictly physical based way (adaptivesamples.com/2013/10/19/fluid-in-a-glass) as there seems to be a gap between the glass and the mousse 3.bp.blogspot.com/-1RBiOPezZHs/T1Ld5mEr71I/AAAAAAAACjA/… $\endgroup$ – Carlo Sep 2 '15 at 15:56
  • $\begingroup$ @Carlo thanks for your reply, i have added blender file. the surface of chocolate filling is slightly away from the inner surface of glass, i suppose thats may be the reason, I will make it between the glass surface and share the result asap, meanwhile you can download the link above, thanks $\endgroup$ – ideorium Sep 3 '15 at 8:48
12
$\begingroup$

Yes


Where are that reflections coming from?

Not from the chocolate material. While in your modeling there is actually a small gap between the mousse and the glass cup (this mess up IOR calculation, but that's another story...) that would surely create additional reflections, those are not caused by that.

If you make the chocolate object completely diffusive (or delete it), you'll se that the effect is still there.

enter image description here

Let's erase the object, leaving only the cup and put a small red ring near the base of the cup.

enter image description here

We see that on each stripe (reflectons) is coloured by the red ring. This means that the light travel through the glass object with the path (more or less, I'm not an expert of optics) that is highlited in the image below.

enter image description here

Notice how cutting away the half of the ring that is behind the glass make the red stripes disappear. This is because the reflex is just what the faces (opposite to the camera) of the glass cup base are pointing to.

Let's put a small object:

enter image description here

On the right, inside the black circles there is a small green spot. By maginifing the image we'll see...

enter image description here

...Suzanne!

I'm quite sure it is upside down, as when you look yourself in a spoon... we can run another test by pushing the lower faces of the base behind the plane surface. So the ray will hit a diffuse shader and should reflect nothing:

enter image description here

enter image description here


Why there are so many?

Because for each camera ray the light path is slightely different due to the different incident starting angle. At the end of the path they are probably "diverging", so in a small portion you'll be able to see (distorted) the whole surrounding.

enter image description here


Is this a physically based effect?

Yes it is. You can see it in glasses with a thick base.

Here's a comparison between two different angles.

enter image description here


Why choosing "Limited Global illumination" instead of "Full Global Illumination" don't show them?

Because Limited Global Illumination has lower number of "Max" & "Glossy" samples. That factor is ruling after of many light bounces the simulation is stopped. If they are set to a value lower than the noumber of bounces that the light need to travel to get out of the glass, you'll not see any full reflex from that ray. There would maybe some light path that would be able to get out of the glass even with few bounces, but increasing the values, would give higher chances, so it is more likely to se lot of reflexs.


Conclusion:

The effect is phisically based. Full Global Illumination's setting as espected helps in making the scene to look more realistic, as it takes in account more bounces.


Suggestions:

There are some materials who use the "Add Shader" node that isn't energy conservative, so it can lead to non stichtly-physically based behaviours.

When a fluid is stiched to the walls of a glass container, it is better to deal with interfaces as the link in my comment shows to get a "true" IOR effect. Notice that with that method, you'll also will get rid of the reflections, as the light is reflected (as in real life) towards the mousse, instead of bouncing all around trought the glass volume:

enter image description here

$\endgroup$
  • $\begingroup$ OK, GJ me. This would make sense. Kudos for being super clear. $\endgroup$ – Mörkö Sep 3 '15 at 15:51
  • $\begingroup$ Wow what a great answer, thanks a lot @Carlo I m away from my computer now, but will study it properly asap.. I guess leaving gab between the chocolate filling and glass inner surface wasn't a good thing at all .. Thanks a lot for this detailed explanation $\endgroup$ – ideorium Sep 4 '15 at 7:31
  • $\begingroup$ Feel free to argue my point of view. Despite I wrote many lines, I just know a little about Blender's optics capability... I'm still running some tests to clarify myself some points (one above all is "why are there so many?"). @user277143 $\endgroup$ – Carlo Sep 4 '15 at 15:39
1
$\begingroup$

Afaik. This is normal. What is happening is that the chocolate filling reflects the glass' internal surfaces multiple times. One way to avoid this would be to use the light path node to limit the glossy component of the chocolate to only non-glossy rays, effectively elliminating the problem. ANother way around the problem would be to remove the faces of the cup that are in direct contact with th chocolate.

$\endgroup$
  • $\begingroup$ thanks @user277143, Light patch node is actually confusing me at the moment, i don't know how to include it in the node system :( , can you explain a bit more please? thanks $\endgroup$ – ideorium Sep 3 '15 at 8:49
0
$\begingroup$

Your problem lies with the use of full white color. Use close to white so that those reflections will diminish with each iteration.

$\endgroup$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.