Glass

[Tangled-Universe]

Disabling vtor disables calculation of other rays (than the camera's rays I presume) and fits with the subdivision principle I described.
The lack of subdivisions causes a lack of rays to be calculated correctly and gives black patches/triangles.
Disabling vtor disables the calculation of rays at all and therefore results in an entire black reflection.
Disabling detail padding (no detail settings) still calculates the rays, but performs no subdivision and therefore gives the lowest level quality possible then.

However...still wondering how Marc does this...I must be wrong then, but I can't see why yet

[dandelO]

Martin, yes, I understand how those ray detail settings work.

You will need insane values to get it work, if possible at all.

What I tried to say in that first reply to your post was, that I padded the detail in camera to a really insane amount(the same as the radius of the BG node = 2e+008). And this gave me the same result as the 'no detail' setting.

I think it's impossible, too.

EDIT: Or do I? Marc, wtf are you doing here?
It seems that there is no inversion of the sky over ground though so, it isn't refracting correctly, at least in the picture with the 3 spheres. The newspaper one however, that is possible in any case because you are not seeing any sky rays in that transparency. You're looking directly down on it. Try getting to an equal level of this, so some sky is visible. If it holds up... then, BRAVO!

...

[Tangled-Universe]

Wish the newspaper was TG2, caustic effect
Anyhow, Martin, I didn't miss your point about setting an insane value for region padding. Like I predicted it doesn't work because it works on a kind of "planar level".

[MGebhart]

Ok....here we go. First I'm not clear if I'm heading in the direction you want or getting close to the results you are looking for. That said...

If you notice I cranked up the Decay Distance to a ridicules 5000. The Index of refraction is set to 10. YIKES! The Master reflection is .8. I also am playing around with the Volumn 1 density and color. In this render they are set to 1. (I think this has possibilities)

I thought about how Martin (dandelO) created translucency on my Ant model by cranking up the value to a whopping 250 so, I played with greater values as well to see what would happen.

Anyway, I hope we are getting close.

Remember: I used the water trans 1.tgd and the glass.tgc with my spheres.

NOTE: I think the Decay Distance will change dramatically depending on the object size. Just a guess.

[MGebhart]

Down view.

[cyphyr]

The 'no detail' is atrocious and proves exactly my point. At least, for me Smiley
It has to do with reflections from outside the frustum (the sky) which aren't subdivided enough. "no detail" for ray detail padding means that there is no subdivision at all (never use this by the way!) and therefore gives the atrocious result.

This may be exactly the same reason I'm getting "shadow popping" from sunset scenes when the sun is low and behind the camera. Its not possible for the camera frustum to to extend beyond 180 deg beyond the cameras FOV.

Richard

[dandelO]

I see. It seems you're just over-riding any interior artefacts with surface reflection. Kind of just like a shiny ball that's a little see-through. It may be a possible workaround with some effort in the meantime. Though, it doesn't appear to be inverting the 'world' inside the glass, you know?

[MGebhart]

I'm going to bed.......

This is the last one.

It is transparent, reflective and turns the world upside down.

Included the clip file.

NOTE: I will never get this day back.

[Seth]

damn, i get off the forums for a few hours and you wrote 2 more pages about it !!! ^^
need to be tonite to play with you guys

[dandelO]

Hee-hee! Marc, you're persistent, eh? I gave up and went to bed long before you.
That looks very good, I'll have a look when I'm in from work later...

[Tangled-Universe]

Hee-hee! Marc, you're persistent, eh? I gave up and went to bed long before you.
That looks very good, I'll have a look when I'm in from work later...

Looks very good. Same here, will check it once back home.

Sorry that you've "lost" a day

[Dune]

I've just been playing with this as well, with some knowledge in my distant memory about using reflective+water shaders, and some weird settings. Someone did this before, I vaguely recall.

I found that what happened when doing this another way, was that the surroundings are projected on the inside of the ball. So it looks like transparency, but isn't.

And while I'm at it, got some strange result with the refraction index set at 0; the whole scene turned black. Strange, huh?

[dandelO]

Dune, that's great.
One, well, two questions about the working version; Is the sphere object the same settings in the first shot as is shown in the second screenshot(double sided/no shadows)? And, Have you disabled reflections/refraction on the water shader, too? Since you're adding reflection and refraction via the reflective shader afterwards.

I can't quite comprehend how this is working so well, although it clearly() is working.
To my mind, there should be no initial transparency from the water shader because the object is rendered as double sided and, why are there no atmosphere artefacts rendered in this way? I'll try this soon, too, thanks for chipping-in. Things always seem get worked out in the end in here, one way or another. Great!

[Dune]

I'll get back about this, have to check. But I don't think there's difference between 2-sided and 1-sided. Shadows enabled, I think. Reflection in the water shader is 0, I think. I think a lot  I'll check, that's better than thinking...

[Dune]

Checked. Two-sided is okay, shadows are okay, reflections are off in the water shader. But you'll have to use your own balls, so to speak. The indigenous sphere doesn't work.

Some more examples. I even caught a cloud in a sphere, but didn't render out.