The result is a broken or yagged figure, even with refraction index set to 1.00 the distortion is present.

I am using Poser6, I don't know if I have to setup something else or is a Poser6's bug or in Poser 7,8,9 this still happens. Later I shall try to render this same scene with LuxRender.

In this image I have used refraction index = 1.00

Kawecki, what's the effect you're going for?  I guess I'm not understanding the reason for the transparent plane with refraction...other than, well, it's doing what it does.  It's refracted the reflection :/

refraction does bend light, no? What were you shooting for?

Laurie

Quote - I'm not understanding the reason for the transparent plane with refraction

Changing the refraction index moves the position of the reflected image. Indexes greater than one move the image to the left and smaller than one move to the right and with inex = 1 the image stays in its original place. It is working, but the image is a crap even with index = 1.

Next you can curve the plane and more speciall effects can be done!!!

Ok, so, You have the figure, and a transparent refraction plane.  So, the mirror (which appears concave?) gets a refracted image, and the camera sees the refracted image through the transparent plane, so it's refracted again.  

Right?

Do you have a photo of the kind of effect you're going for?

Quote - Do you have a photo of the kind of effect you're going for?

No, I am creating the effects and what can be done with them. What about a mirror where you are looking at and see your back reflected? Well, you need two mirrors.

My best guess is your problem is that the light is being bent twice.  Once between the  lens and the mirror from the character and once again from the mirror to the camera.

I have no solution to the problem, but I think everyone is missing kawecki's point. Refraction bends light, the light leaves a medium at a different angle from that at which it entered the medium. The effect should be constant. If you put a stick into a bathtub, it will appear to have a kink at the point where it enters the water, but only at that point, the part before it enters the water will appear straight, and the part in the water will appear straight (providing the water is still)! Real refraction does not create sudden discrete jumps, except where an object crosses from one medium to another.

Is the refractor a single or double sided surface? If double sided, try using flattened cube in its place, with a little space between the two opposing sides of the cube, or try using displacement to separate the two sides of a square. Poser seems to have problems with double sided surfaces with no separation between the two sides. This is not guarenteed to work, but it's probably worth a try.

Quote - but I think everyone is missing kawecki's point. Refraction bends light, the light leaves a medium at a different angle from that at which it entered the medium. The effect should be constant. If you put a stick into a bathtub, it will appear to have a kink at the point where it enters the water, but only at that point, the part before it enters the water will appear straight, and the part in the water will appear straight

A window is even much better example, a window is made of glass that has a refraction index and you don't see through the window objects cut into pieces as in my rendering.

Quote - Is the refractor a single or double sided surface? If double sided

I am using one sided square, it is like a window with thickness = 0, I don't know what Poser will do. In real life in this case the glass will be as it doesn't exist because light refracts, travel a zero distance and then bends back as leaves the surface producing zero displacement of the light ray. Poser obviously is not doing this.

The other possibility is that the light bend when hit the surface and then continue to travel with this new direction. This will produce a change of the angle of viewing the object and no mesh deformation. Again Poser is not doing it.

I don't know what will happen with a two sided square or a cube. I don't know if Poser will treat this as a volume filled with a refractive material or it will be nothing more than two planes with normal media between them or it will bend even more when hits the exiting surface. Later will do the tests.

The most weird thing and this is a bug of Poser, at least Poser6, is that a medium with refraction index 1.00 is 100% transparent, doesn't bend the light, doesn't reflect the light, and doesn't deform the object, it is just invisible. This is not happening and the posted images were rendered with refraction index 1.00

The plane with refraction index combined with a paraboloid mirror behaves as a magnifying glass. You can change the maginification changing the refraction index of the plane or the focal distance of the paraboloid mesh.

If the plane is set with a refraction index = 1.0 the plane turns invisible and the rendering with or without the plane gives the same result, what is correct.

I'm still not sure what the point of all this is. I do realize you may have found a bug, but why do it in the first place?

Laurie

Did you try change square prop. If you use poser primitive 2 sided square can give weird results.

Ok, now I see what you're trying to do.  

It might be worth it to try the cube, just to see what happens.  I'd try it, but I'm at school right now, and don't have Poser on my laptop (although I should.  This thing is more powerful than ANY of the rigs I have at home :/ )

I checked and the square primitive is a quite nasty thing, made up by two squares coincident in space and with opposing normals. Not surprised that bad things happen; probably Lux works because it does not detect interactions with the back plane (I know that when you use it to render fluids in bottles you have to very slightly scale the mesh representing the fluid in order to big Lux a reliable glass->fluid interface).

Bye!!!

Quote - Did you try change square prop. If you use poser primitive 2 sided square can give weird results.

I am using one side square, so it has only have one face. I have tried with the square subdivided into many small squares, but the result was the same. I have not tried a two sided square with no welded vertices or a cube with some volume enclosured.

With LuxRender is working, but the problem is that is very slow and this makes very difficult to work. Poser is fast and so I can change some parameters and render, change and render, change and render. Within some minutes I have done a lot of experiments, this cannot do in LuxRender.

My latest tests:

1- Using a two side square instead of one side square gives the same broken mesh rendering.

2- Using a thin cube instead of the square nothing is changed.

3- Increasing the depth of the cube so the mirror is left inside the result continue to be the same.

4- Changing any parameter in the rendering setting has no effect on the mesh crash.

5- Replacing the paraboloid mirror with a one side square mirror, the mesh is not more broken.

6- Refraction index bigger than one makes the reflected image to be more near (plane mirror)

7- Refraction index smaller than one makes the reflected image more far away

8- Refraction index of 0.2 duplicates the reflected image, weird !!!

9- Negative refraction index bigger than one have a magnifying effect.

10- Negative refraction index smaller or equal to one produce no reflected image

11- Using the paraboloid this time with refraction and a reflective square behind doesn't break the mesh and have a magnifying effect depending on the focal distance of the parabolid.

Conclussion.

The parabolid cannot be used, at least in Poser6, as reflective surface if is behing a refractive surface because it makes the mesh's reflected image  into pieces. I suppose that the same will happen with any surface that is not a plane.

Instead a plane surface used as a mirror can be used behind a refractive surface and doesn't break the reflected mesh.

LuxRender doesn't makes the mesh crash and behaves normally, pity that is very slow and negative refraction indexes doesn't work

a larger prop (in relation to figure size) with more polygons may reduce caclulational errors in some version(s).   I haven't got poser 6 :crying:

I have increased the number of polygons and the scales and nothing has changed a bit

The only thing that haven't changed is Poser6.

Is a zip file, remove the txt extension

Well, it works in Pro 2012, so I need something newer than Poser6.

As is working, experiment with bigger refraction index and smaller than one. Also try with negative values and tell me what happens.

You can change the focal distace of the paraboloid changing the zScale value. The reflected image will be bigger or smaller. Also moving the firgure closer or more far awy the mirror will change the size. The parabolid is not a flat mirror, so it has a lens effect.

It looks that negative refraction indexes are not working in Pro 2012, the same with LuxRender. Is a pity to not be able to play with metamaterials