Low glossiness of Rhino materials produce noise ...

[Micha_cg]

... because it seems to be that subdivs are not set at typical values. It's not a bug, but a pitfall.
I think, good subdivs values could be (for standard Adaptive amount 0.85 and noise 0.01)
glossiness - subdivs
100 - 8
95 - 11
90 - 22
80 - 33
70 - 44
60 - 55
50 - 66

Maybe, it's possible to find a function to calculate a good subdiv value from the glossiness value for Rhino materials. For example:
110 - glossiness = subdiv
It's not perfect, but better than allways subdiv 8.

[Gijs]

Re: Low glossiness of Rhino materials produce noise ...

I don't know but using noise treshold and global subdiv. multiplier seems to work pretty well. I think you can set all materials to a high setting as well and just use the noise treshold to control quality. As you know, the actual samples are always dependant on the qmc sampler settings.
On the other hand, sometimes the produced noise is desirable to give the material a sandblasted look instead of a perfectly smooth gaussian blurry surface.

[Micha_cg]

Re: Low glossiness of Rhino materials produce noise ...

I tested an interior with with one blurry material - subdivs 22 and a very high count (66 or so). The speed was much slower with 66, same noise level. So, I think, it's good to use specific subdiv values.

[Gijs]

Re: Low glossiness of Rhino materials produce noise ...

One other thing, it is also dependant on the geometry: flat surfaces for example need higher subdivisions for reflections, than single or double curved shapes, because reflections on those objects are optically smaller.

I'll see if I can run some tests myself though, because this is indeed a tricky area of Vray.

[Micha_cg]

Re: Low glossiness of Rhino materials produce noise ...

One of the reasons why the full adaptive mode should more improved.

The biggest problem of the full adaptive mode is, that the IM produce low frequency noise sometimes with noiselevel 0.01... 0.02. Fine QMC noise is ok, but splotchy IM noise is ugly. The IM should be independent from the noise level or ... .