Faces become Blue under Water.

[victor.nsy]

I have had this experience that when an object is just a plane ( without any faces on the backside )
the fog color becomes difficult to control. My guess is that probably the rays requires a face on the
backside, as the amount of fog greatly depends on the thickness of an object apart from the intensity
setting.

But in my case the whole scene behind that object turned into a deep tone of the fog color. Whereas
this is quite new to me ... only the surface on the box is deep blue and the rest of the pool still looks
quite right!

Looks like a bug 0.O

best regards,

[werticus]

if you make the water properly, ie: a box solid that goes through the containing walls, does it work?

i've never found water to look right if its not made 'physically correct'

[^Lele^]

I *think* i fixed it.

I changed the shadow bias for the sun to 0.001, gave thickness to the water mesh (shell modifier), and fixed the fog material settings.
For the final render i also added caustics, as a swimming pool without them is not complete XD

As a side note, I also made the sun invisible, as it would lead to the usual ringed over bright spots with the filter used.


http://rapidshare.com/files/79885570...e_001.rar.html

[Aldaryn]

I was wondering, if it makes any difference to have a perfectly modeled water mesh as water for example in a pool, or just to have the top plane.

It is even hard to grasp what would be a perfect mesh here. There are two situations, where having faces other then the top ones makes any difference, and imo both will lead to rendering artifacts, or imperfections. First you can have a small gap between the pool surfaces, and the water, but in this case you'll get an unwanted refraction where irl you normally wouldn't.
Second, you can model the water surfaces coplanar to the pool walls, but I think this is one thing to be avoided.

In any other situation, rays would hit the pool walls first and will never intersect our additional water mesh faces, thus making them unnecessary.

So, considering refraction, I think it makes very little difference, and in my opinion, having only the top faces gives a more accurate solution. Same goes for fog too, Rays entering trough the top faces will start accumulating fog saturation as they travel, and eventually hit the pool walls and since the walls are opaque, it will also terminate fog saturation gain. If we have a volumetric water model, separated by a small gap from the pool, the fog gain termination will occur slightly sooner, when the ray exits out water mesh, but again this makes very little difference imo.

Of course, all this applies only if the water meets a completely opaque surface, if another refractive surface is involved, an accurate and intersecting mesh is required.

Correct me if I'm wrong, because I'm concerned about this.

Best regards,

A.

[^Lele^]

In VRay contact surfaces are correctly calculated.
A "good mesh" is as you see in the file i posted: a closed volume.
Slightly larger than the pool, and intersecting its walls.
In fact, think of a glass with some liquid in it.
You'd model that liquid as a plane?

[Aldaryn]

In fact, think of a glass with some liquid in it.
You'd model that liquid as a plane?

Fourth line from the end in my op.

So, if a ray hits a completely opaque face, it will continue to trace it further set by the bias value, to search for another surfaces because it presumes there might be a contact surface? This is whats really bugging me, even if it does, will it make any difference at all? And strictly talking about situations, where the liquid is in a completely opaque container with no openings sealed with dielectric materials.

Best regards,

A.

[^Lele^]

Oh sorry, I did miss it.
Yes, I see what you mean, in fact my geometry intersects the bottom too, making sure the only actually visible surface is the plane we started with.
What it really is for me is a safety measure.
It's a lot easier for me to visualise how far the fog tracing will go at most (despite density), and it's of course heritage of working with dielectric contact surfaces.
In any case, the effort is really negligible (pretty much any single-sheet object can be given thickness with one shell modifier, or for the precision jobs, an edit poly), and in cases where it produces no difference (like this one!) one still has a "correct" result, rather than a potentially wrong one (the client really wants glass bricks for that swimming pool...).

I traced the issue down to the fog material setup alone, please confirm this, in case.

[Aldaryn]

I totally get your point. Put it in real, get it out real. And this goes pretty well for everything in renderers nowdays. I was just very curious about your opinion on this.

Best regards,

A.

[Codi]

I had a look at the scene too and, as others have stated, the problem seems related to having a sun shadow bias too high (0.2m in the scene).
For some reason the faces that aren't affected by the water shadow despite being under the water surface, due the shadow bias, appear dark blue.
I found that if you enable 'force back face culling' in the general options the problem seems to go away, so it makes me think that there is a problem in the way shadows are calculated in this particular case, but I'm not really sure, maybe Vlado could explain why VRay behaves like this.

[Aldaryn]

Why was the shadow bias set to 0,2m in the first place? Obviously, a 20cm displace for the shadow *will* cause visible changes when rendering a pool in real world scale.

Best regards,

A.

[Codi]

Yup, simply setting the shadow bias to something more suitable for the scene will solve this problem as well as some other artifacts that are visible on the sample image (the strange shadow on the bottom right corner of the column).

BTW, the scene has an Architectural material on it, to be safe better stick to VRay materials.

[Da_elf]

i have never had problems using a plane for the surface of water. its nice and quick to setup and works

[vlado]

If you used fog for the water, and the shadow bias is too high, then V-Ray will figure out that the faces are inside a fog, but the shadow rays will start *above* the water and will never actually cross the water surface - so the fog will never be turned off and it will cause all shadow rays to end up blue.

The solution would be to reduce the shadow bias to something that makes more sense, e.g. several millimeters or something.

Best regards,
Vlado

[pskdude]

cool

Awesome yea must have missed that i work so many projects in different scales i lose track sometimes. Thanks so much for the help guys u where very helpful !!!!!!

=)