[Mihai]

@Mihai: If you split the red surface in a water/glass surface and a air/water surface it should work. But if it dosn't work, than for me something in Maxwell is wrong, because than anything should be physical correct.

Either you are talking about a material that when applied to a surface knows to treat rays differently depending on what side of the normal the ray has entered, but you only have ONE surface of polygons to represent this glass|water interface, OR you actually have one separate glass surface, and one glass surface. I'm not sure which ones of these you are referring to?

[DELETED]

DELETED

[rivoli]

i'm a bit ..err..puzzled..thomas' version still right, no?

i've got nothing left to believe in.

[Thomas An.]

err..i expected a conlusion post to be 1: the conclusion and 12 pages of congrats..reading through 12 pages of hybrid-mihai-b-rep-cutoff-rivoli-normal-inverted-snell-ray-knowing-where-it-came-from i'm a bit ..err..puzzled..thomas' version still right, no?

Yeah, I am puzzled too
I thought it was proven beyond reasonable doubt. I don't understand why people still object, since all the information needed is there to draw the conclusions in both the theoretical and experimental aspect.

It is basically people coming in late in the discussion and then keep asking the same things that others have asked before, so we are going in circles. It is like restarting the discussion several times over.

-

[Mihai]

It is "the most right" until Maxwell gets another way of handling these dielectric|dielectric situations. The photo comparison is good enough no?
It's the nonbelievers that cause confusion

[Micha]

Sorry. It seems, I have over seen some facts in last posts.

[Micha]

It seems, that pofis like the keytoon team have some problems with water too.
In this animation frame it looks like air between the water and the hands. But besides, the animation is great and my baby love to see it again and again.

www.keytoon.com

[Thomas An.]

Maybe it would be a good idea to start a "summary and conclusions" thread so people don't have to read through nine different threads and 54 pages of often conflicting/confusing experiments and theory.

The first page has been updated to include the conclusion and experiement result along with the theory.

-

[tom]

thx thomas!
glasses 4ever!...hehehe

[Thomas An.]

Your scene is fine (its perfect). The caustics will take a long time to clear. They will become more promonent after sample level 18.

Let it cook all the way to level 23

[daros]

I think they are not enough samples. To reach an image like this one i think you must reach 14-18 samples.

[GM5]

Wow I just read through this mind-numbing thread. Thanks to everyone who contributed! Also, thank you Thomas An for the great test that you set up. Using the photo as a reference is the way to go. It is funny that sometimes you think something should look a certain way, but in reality it doesn't. I've looked at liquid filled glasses for a lot of years, but didn't really see them until I tried to recreate them in 3D. I'm amazed at how close the glass, liquid, and caustics in your render are to your reference photo. Of course M~R certainly helped!

-Greg

[i-d]

Well I just got crazy from all this information and made this.
I must say that this test scene I use on all raytracers never
looked better. Off course I used dreaded gap technique but
beleive me I had to.

[pluMmet]

I realize this is stupid but this is driving me nutz:

In this original image the glass has ripples in the bottom of the glass causing the top of the liquid to also show the ripples.

I am now rendring the scene hopefully with the ripples (fingers crossed.)
And thanks everyone in this thread. Great stuff! It helped.

[pluMmet]

Well this scene keeps crashing: http://mysite.verizon.net/scanbay02/M~RglassTest13.zip

Could somone check it out

what a great compression this .7z is