WinOsi Caustic Test Scene

[manitwo]

Hi guys

When i saw this thread on the radium-forum i wondered how indigo would handle this scene ...


Redqueen reference pic: rendered in 30min (biased i hope)




Radium (PT): rendered in 3d by "tg_jp"




Indigo 06 (Bidir+mlt): rendered in 8h 10min




Maxwell 1.1 Demo: rendered in 2h




Indigo 07testversion (PT): rendered in 10min (i expected that - looks kinda artistic )




As IanT said on the radium-forum this is nearly the worst scene you could try with an unbiased renderer. (but anyway interesting )
Theres a really small emitter behind a glas lens - so everthing is lit by caustics.

Something is wrong with my indigo-caustics (maxwell caustics looks good) but overall its not so bad i think.
The Indigo image is much brighter in the corner than the maxwell one after that 15min.
But it looks like if a indigo-caustic meet the blue wall the caustic gets kinda purple (is this okay?)

@Ono: what do you think?

btw, winosi+rq testscene: http://www.teamredqueen.com/gallery.html

[kepler]

Why Indigo doesn't show the light's rays on the floor (those circles)?

[manitwo]

look a bit closer - they are there - not as bright and strong as in the maxwell
one but they are there. They would get sharper if i let it render for a longer time.

[kepler]

look a bit closer - they are there - not as bright and strong as in the maxwell
one but they are there. They would get sharper if i let it render for a longer time.

Ok, it's a matter of time, right?

[manitwo]

Ok, it's a matter of time, right?

yes i think so
I'll render it for a longer time when i figure out what's wrong with the shape of the caustics.

[CTZn]

But it looks like if a indigo-caustic meet the blue wall the caustic gets kinda purple (is this okay?)

Looks like you used a very low value for <large_mutation_prob>, so if this is the case any color is ok at the beginning, indeed it needs more time.

OOps that was said for the last image (0.7t3) !

[IanT]

This exactly proves my theory that BiPT will do fine with the direct caustics but will still struggle with the indirect ones. Radium's only PT (at the moment) hence the HUGE difference in results

Just so others can have a laugh ... that Redqueen image was claimed to be "unbiased" on the site it was originally show on

Ian.

[manitwo]

Radium seems to have a really good PT-code.

[IanT]

If any Fryrender beta user is reading this ... would be an excellent test for that too

(There is some doubt that Fryrender even does indirect caustics yet...)

Ian.

[zsouthboy]

I am currently rendering this scene in Indigo 7t2.

And yes, ... its a little "artistic" at this point ~5 mins.

It just doesn't "feel" right at all.

Indigo works so awesome otherwise; it just seems weird to find a scene that it doesn't in.


Question: who the hell are these "redqueen" guys?

EDIT: oh, and what settings should the emitters have? There wasn't a mesh defined in the 3ds file for a "light", so I simply copied the glass lenses, changed the material to emit, and slid them up Z a few notches.

[manitwo]

there are emitters in the file. as i said they are very very small and
above the lenses. These meshes should emit light in two directions.
To make that I copied these emitter meshes and inverted the normals.

[Zom-B]

To get some caustics in Indigo 0.6 just disable BiDir

[arneoog]

Am I the only one who can't see the Indigo Renders!!??
Just getting "File not found" ...
All the other images loads fine..

[arneoog]

Sh*t... my first dubble post..

[blueplanetdesign]

I can't see them either...