Physical sky sun artefacts

[Micha_cg]

Hi,

I post it from time to time and ask me - is it on the radar of the development?
The problem is, that the bright reflection intensity cause strong noise artefacts at blury reflections. Solutions could be:

* lower reflection intensity
* limited interereflection depth for sun reflections
* disable visible sun reflections
* user can switch between reflection or highlight effect

Here an old post example:
http://asgvis.com/index.php?option=c...90&topic=631.0

When can be a solution expected? Today I run in the problem again and I'm missing a solution. My current solution - avoid the sky/sun system, if blury reflections are at the scene (interiors, yacht design). :-\

[dalomar_cg]

Re: Physical sky sun artefacts

This one is an issue that would be resolved through the ChaosGroup. As you know it is a sampling issue, and that is something that we don't have control over.

The best work around right now is through clamping and subpixel mapping. I've also used different color mapping modes (like exponential based modes) to manage this as well. The coin caustics image that I did on the newsgroup a few weeks ago was having the problem of bright individual pixels and turned out fine with these adjustments. I know its not the best answer, but their are work arounds.

Post with image attached
http://news2.mcneel.com/scripts/dnew...m=294282&utag=

[Micha_cg]

Re: Physical sky sun artefacts

This one is an issue that would be resolved through the ChaosGroup. As you know it is a sampling issue, and that is something that we don't have control over.

I feel, it's a "house made" Vray sampling issue. I don't understand, why Vlado stick at the extrem high intensity reflection solution. At Maxwell I never have seen this issue. I remember me, that I create 360° maps from the Maxwell environment for use at VfR (befor the sky was implemented). The intensity of the sun was in a good range, not so extrem high like at Vray. Maybe it's not physical correct, but interesting, nobody of the Maxwell user seems to miss it.