windows/glass turns black in SkIndigo

[lape]

Beeing still a noob I struggled the whole evening with a model where some of the glass panes turned black while rendered by indigo... it gave me headaches, because the model was free of mistakes, all normals pointing outwards..

and then i found this:
http://www.indigorenderer.com/forum/vie ... 37&t=10384

Perhaps this should be mentioned for beginners in some manual for SkIndigo. (Or did I miss something in the tutorials/manuals?)

[cotty]

Because Indigo is a physically renderer, the geometry have to be physically correct and a thickness of 0m isn't...

[CTZn]

Lape does understand the cause. He is saying that it should be documented better

[lape]

Because Indigo is a physically renderer, the geometry have to be physically correct and a thickness of 0m isn't...

I know that geometry should have a thickness, but that is not the problem. The problem with the black glass occurs, when geometry with different materials in SketchUp is not modelled apart from each other:

" to avoid problems, the glass cube must be modelled separately in a place clear from other geometries, then grouped, then moved to its desired location"

[Voytech]

Because Indigo is a physically renderer, the geometry have to be physically correct and a thickness of 0m isn't...

I know that geometry should have a thickness, but that is not the problem. The problem with the black glass occurs, when geometry with different materials in SketchUp is not modelled apart from each other:

" to avoid problems, the glass cube must be modelled separately in a place clear from other geometries, then grouped, then moved to its desired location"

Yea, I've spent many hours fixing my first building when I realized that. All kinds of wonky stuff happen when you have you glass in the same geometry as the rest of the model. When it does render, it sometimes does this weird diagonal stuff.

[lycium]

Hiya lape, there's a page in the docs describing this issue in detail, the causes and solution

Please check it out, and let me know if there's anything unclear: http://www.indigorenderer.com/documenta ... ling-glass

[lape]

Hiya lape, there's a page in the docs describing this issue in detail, the causes and solution

Please check it out, and let me know if there's anything unclear: http://www.indigorenderer.com/documenta ... ling-glass

ok, sorry, I've also seen this manual, but it was not clear for me that the glass needs to be modeled apart from other geometry and needs to be grouped... well, now I know it

[galinette]

Because Indigo is a physically renderer, the geometry have to be physically correct and a thickness of 0m isn't...

Well, Indigo assumes that a ray is 0m wide (base assumption in geometry optics)

"Physically correct" has no real meaning, physics is all about hypothesis, models and assumed simplifications. AFAIK, Indigo doesn't handle quantum effects.

diffuse_transmitter is an example of modelling a complex structure with non-zero thickness and SSS by a single sheet, as it can be seen as such at macroscopic scale.

Etienne

[CTZn]

"Physically correct" has no real meaning, physics is all about hypothesis, models and assumed simplifications. AFAIK, Indigo doesn't handle quantum effects.

I'm risking: physically correct in regard with optics, as a field of sciences attached to model the photographic reality of optical events ?

I suppose that going down to a "quantic renderer" would have a very little interest in simulating actual photography optics.

Can iridescence be qualified as a quantic effect (out of a model indeed) because it implies a sub-periodic phenomena ? Is this effect at the limit of... anything ?

[lycium]

Can iridescence be qualified as a quantic effect (out of a model indeed) because it implies a sub-periodic phenomena ? Is this effect at the limit of... anything ?

lolwut

Sorry mate, that makes absolutely no sense About iridescence though, that is indeed due to quantum behaviour, however we can (and do) simulate this at a higher level rather than from (quantum) first principles.

I suppose an analogy can be made, for rendering rainbows: you can either do it physically-based from first principles (per-wavelength optical scattering), or make some assumptions/approximations and use an ISL shader for the final spectrum.

[OnoSendai]

"Physically correct" has no real meaning, physics is all about hypothesis, models and assumed simplifications.

It certainly has no precise meaning.
I still think it's a somewhat helpful term though, or at least 'physically based' is.
I would class a rendering systems as physically based if it at least tries to use real world physics and units where possible.
For example, a lot of computer game opengl / direct3d rendering is not physically based because
a) There is no attempt to use real world units or physical quantities - the sun might be set to (1,1,1) RGB colour, because it 'looks good', as opposed to using real-world measured spectral radiance values.

b) The opengl lighting model has constant and linear falloff terms for point light radiance, along with the correct 1/r^2 falloff. It's not like the correct 1/r^2 term is more difficult to compute, but the incorrect terms are still there.

EDIT: another example of non-physically based rendering: Operating directly on 8-bit textures in non-linear colour space.

[CTZn]

Can iridescence be qualified as a quantic effect (out of a model indeed) because it implies a sub-periodic phenomena ? Is this effect at the limit of... anything ?

lolwut

Sorry mate, that makes absolutely no sense About iridescence though, that is indeed due to quantum behaviour, however we can (and do) simulate this at a higher level rather than from (quantum) first principles..

So, iridescence is at the limit of the standard model for optics I'm assuming (as opposed to what you call the first principles of quantum physics)... I admit that the question waise asking for a sense more than bringing one out, thanks !