WinOsi Caustic Test Scene

[Kram1032]

I've a weird material problem: My reflective mat isn't reflective at all.
It seems to be reversed: as if it is absorbing light. It's uniform black!

Is there any mistake with the material code?
I checked the normals: They're pointing in the right direction.

My mat code:
<material>
<name>Reflective</name>
<specular>
<normal_reflectance>1.000</normal_reflectance>
<ior>1.000</ior>
<transparent>false</transparent>
<cauchy_b_coeff>0.0</cauchy_b_coeff>
<rgb_absorptivity>0 0 0</rgb_absorptivity>
</specular>
</material>
unmodified, directly exported with blender. Used Indigo V0.6

[fused]

ior should be higher than 1


try 100 or 1000

it should be perfect reflective

[Kram1032]

thanks, I'll test it immediately.

[Kram1032]

Ok, thank you very much!
It works and it even doesn't seem to have this bug with a vertical line cutting to the caustics at each wall...

What's the IOR at reflective surfaces like? the reflection gets sharper the higher?
Is there any maximum value?

[fused]

there is no maximum value. minimum is 1.

the reflection doesnt get sharper but stronger the higher the value is. unlike to phong materials where the exponent controls the sharpness the reflection is always sharp.

[Kram1032]

I've used 1000 for now.
Is there any see able difference to 10000 or 1000000?

What's this cauchy_b_coeff-value?

And another question:
Please can anyone list up all the primitves, indigo uses?
I know, one is for sphere <sphere> and one for an infinite plane <forgot tag>
Is there also one for a finite plane and for a (finite and infinite) cylinder?
then, there may be a cube and a cone... These should be all the basic forms...
Am I right?

[CTZn]

I agree with you, kram, actually I was thinking the same about primitives yesterday when I posted this ball cut by a box of air, I told myself that if such primitives would exist it would then be easier to create a procedural test scene for Indigo.

I second the request and I'm going to create a thread about that.

As for now I know only the two primitives you cited.

[Kram1032]

Thank you^^

[manitwo]

Updated frontpage with a new render!

[Kram1032]

You fixed it! fine.
Looks good.

[nikq]

i dont think it looks too different.
Beside that lines (i posted that in the bug-forum) and the reflective parts
(which aren't show the caustics) the indigo image looks really good.
Especially for that time!
To be sure - you should see two images

Redqueen:


Indigo: (3h50min)

Hi, all.

I'm researching GI algorithm and told about that test scene
with Shinji and Bee( authors of RQ and Parthenon ).

Point of that scene is , difficulty of sampling LS+DS+E pathes in bi-dir way.
(note L:light, S:Specular(mirror), D:Diffuse, E:Eye)

Why LS+DS+E is hard?
Because specular brdf is formed in dirac-delta function.
delta(x) : Infinity if x = 0 , zero if x not 0

So in such point, result of Indigo seems to be... not good.
Please check the middle mirror'ed balls.
Indigo's reflected caustics are too dark.

[Kram1032]

Maybe, it just has to go on rendering...

[Kreten]

What up with this RedQueen?
Do they have a forum or site or something?

[Kreten]

What up with this RedQueen?
Do they have a forum or site or something?

[nikq]

you can get original scene file here.
http://www.teamredqueen.com/gallery.html

and there another comparisions w/z maxwell
http://www.bee-www.com/