those test suprised mi a little bit, but on the other hand they didnt as i was thinking about them before the tests..
It is all about Ray per second..
And if you look at non reflecting objects (empty sphere and teapot on the stand), they are both noisy in roughly same way.. But it suprised me, that reflection is noisy in same rate.. That was that suprise..
It would be pointless to implement it - but nice placebo :)
AD ps: jeah, this would be awesome, but there is problem with my bugdet :D
Anyway my PC is maybe HW broken. Because this scene started on ~1,3M rps and ended at shown rps. Or it is just computational way? Or is this logycal conclusion, becase CPU have "more and more to worry about" - parsing more and more pixels to VFB? I have to do another longer test if the value stops, or it contuiniues to drop to some ridiculously low number.
HaND Olt