Mirror?

[Woodie]

So how to make a real mirror? I tried few ways here are the results:


notice a lack of lit bright areas in the mirror. used material - specular IOR 100.




here I moved the light closer to the board. I finally start to see something but it's not it. mirror material first specular, second cr.nk (chrome).

any ideas?

[Kram1032]

where do you think shall those highlights come from?
you need to let something very bright reflect in the mirror

[Woodie]

So I gave it a bit more time. With bigger MNCR (10000). Result (after 20h):



Conclusion - the reflection of the highligts is the result of caustic? Why? Simplest renderers have no problem with mirror.

PS: I tested that with Maxwell with similar results.

Now check this:



Do you see a proper light reflection in the mirror. This was Vray. That is the desired result.

[OnoSendai]

Hi Woodie.
The reason it's so slow to converge, is the problem of 'insufficient techniques'.
It looks like your lights are mirror lights, hence they are effectively casting caustics downwards. And simply put, caustics seen in a mirror = slow.
VRay doesn't have that problem because it's usually running in biased mode AFAIK.

[Woodie]

Thanks Ono. That is correct. I basically modeled light fixture. Now the question is then how to use IES light with mesh? As I understand Indigo can't use regular lights?

[OnoSendai]

Unfortunately Indigo doesn't support IES lights currently, which would be perfect for this scene.

[Kram1032]

should be sort of an expanded tabulated data mode, so that it also can handle IES, shouldn't it?

[OnoSendai]

Yeah, the main difference is that it is a two dimensional function, not 1 dimensional.

[Kram1032]

I see