Feature request: render as single-sided

[Proupin]

Would it be possible to eventually have this feature? what I think it would be useful in some cases.

Simply put, it would consist of a material property that when set, would cause all rays to pass through those polygons whose normal are not facing the ray.

[radiance]

Hi,

this is something that's not usually implemented in a physically based engine, it would give incorrect results and has no basis in reality.
why would you need it ?

Radiance

[Proupin]

I think it would serve as a workaround in some situations dealing with refractive-to-refractive interactions... it's hard to explain and I guess my english is not that good, or I'm too lazy to try convincing anyone. But if you want I can try.

Is non-manifold geometry physically based, open surfaces for example? I believe not, but it's still useful; as is a parameter to control the max reflections (bounces), already implemented, and that is not physically based either... But this feature is golden nevertheless... I mean Maxwell for example doesn't have a parameter to control the max bounces, too strict in my opinion. Is bump mapping physically-based, or normal mapping for that matter? what does physically-based mean anyway?

On a related topic, the current fresnel model, is it complete or it needs further development? I mean internal reflections etc...

[gristle]

It would be useful for having an enclosed studio, polys flipped inwards, so the camera can see inside the studio even though it is outside.
Can sort of do this at present, but need a really wide angle lens. Yeah, not like reality, but if you want a complete real solution, i'll rent a studio!

[radiance]

i'll think about it but i don't think it's a solution as it will cause lighting issues that are completely incorrect in your scene...

Radiance

[gristle]

Cool, your call.

Because such features are physically inaccurate, does this mean others such as exclude shadow, self shadow, unseen by camera etc will be excluded?

Cheers

[radiance]

Cool, your call.

Because such features are physically inaccurate, does this mean others such as exclude shadow, self shadow, unseen by camera etc will be excluded?

Cheers

it depends.
exclude shadows/self shadow is no issue, unseen by camera you can already do with opacity slider.

the shadow stuff is ok and sheduled for beta3, the rest i need to have a think about after 2.1 is out.

Radiance

[gristle]

Cheers

I guess a per material ray recursion is out of the question? Useful if the floor is reflectiive and you have luminous panels that you do not want to be seen reflected in the floor.

One feature I have been meaning to ask about is generating object masks, surface ID/Object ID/Depth maps for comping.
I guess I can achieve this at the moment by using the surface editor, but it is not ideal.

[Proupin]

Radiance, thanks for the answer, about the fresnel model I guess it will keep improving as time goes by...

yes the studio example flipped polys inwards is something I hadn't thought about when I thought of the single sided request, but hey good point... if it will give incorrect results or not in regards of real stuff, well why if we just thought it existed a material in real life that could really behave like that? I mean there are lasers, birefringence, gravity bending light, even invisibility cloaks being developed, real life stuff... not to mention in real life the range is much higher than we would ever have in a screen or print, so our "physically-based" is capped right from the beggining...