About noise in general

[lllab]

i use the full res hdri without any modification, no blur, no gamma lowered, it reflects and gives the light. in vray is no need to use a low res.

just loaded hdri and set input to sRGB (to be linear gamma 0.4545) so it suits the LWF. light intensity was 40.
material depth is default 5/cutoff 0.01, nothing special set, only an IOR 1.6, reduced intensity to 0.5 and set glossy to 0.8.
its rendered on one computer with 2 intels 5680 (12 cores), normal disk, 16gb ram. i used the c4d version of vray, which uses the 2.1 core so should be identical in result

p.s in glas i normally deactivate "generate gi and if not needed deactivate shadow) i test with glass in next round
p.s.s made fast test with glass material (IOR 1.6), "generate GI" deactivated, shadow on. seems no big difference on first sight, i post when ready.

stefan

[pengo]

Ok generate gi off and visible to gi? switched off I suppose
Sorry those so muche questions but more I know about the parameter use more causes of errors or artifacts I can exclude .
Surely and I am wrong but if one cpu as one of the two u have has 6 core phisycal 12 in total with 6 virtual x 2 aren' t 24 buckets calculating?
pls don't think I want to be polemic in any way I only ask to understand as better I can
Thanks

[BBB3]

Thanks for everyone who is trying to help find a solution. I'll do more tests here and try to post results and findings, although I'm not sure I can find the time to put together a scene specifically for testing. Anyway, I think it's better to try and find a solution in real-world conditions - i.e. using a genuine scene.

[lllab]

sorry visible to GI is off, generate/receive is both on (we have slightly different naming in c4d)

here the glass version, 24min at 1600px:

http://images-stefan.s3.amazonaws.co...in_glasson.jpg

[pengo]

Don't affect colors (adaptation only) checked or unchecked? (used or not)
thanks

[lllab]

as saved for 8 bit unchecked (32bit saved with adaption only on looks same)

[pengo]

lllab surely even faster u could get your render if u would use with QMC+LC the environment slot of vray for the hdr/exr map and the same or reduced in intensity in refraction and reflection slots. I'd like to see what time u would get respect domelight usage that for me is more needed with irrad map usage.
Thanks again for tests

[BBB3]

Yes but you are never going to get the sharp shadows from the HDR that you would get when using the Domelight... So this is certainly not equivalent, even if the render times end up being drastically lower.

[pengo]

You are surely right! But what times

[lllab]

yes dome light is much better,you never get nice sharp shadows with environment only. its my favorite light source.

i see not much render time increase with it by the way, only marginal, but much better quality.
dome light has very good importance sampling as far i see, so is very efficient.
important is to have high enough subdivision in the dome light (at least 64 i would say), so you dont generate grain that AA needs to clean up later. also beside speed gain ,the light is more beautiful with higher subdivision.

[simmsimaging]

yes dome light is much better,you never get nice sharp shadows with environment only. its my favorite light source.

i see not much render time increase with it by the way, only marginal, but much better quality.
dome light has very good importance sampling as far i see, so is very efficient.
important is to have high enough subdivision in the dome light (at least 64 i would say), so you dont generate grain that AA needs to clean up later. also beside speed gain ,the light is more beautiful with higher subdivision.

That is good to know. I always leave that at default (more or less following the old Universal method).

Thanks for that tip
b

[PIXELBOX_SRO]

brute force, glossy reflection and one hdri dome light, all a subdiv of 256,

hmm.... is there a global button for this or do i have to go to tweak every subdiv manually?

[BBB3]

Yes, that's the problem. I have a library full of models that all have glossy reflections subdivs set at 8...

[voltron7]

hmm.... is there a global button for this or do i have to go to tweak every subdiv manually?

...well, Olly's VMC will take care of the glossy subdivisions...

[lllab]

you could use the global subdivision multiplier, default subd. is 8, so set the multiplier to some like 32 shoudl be similar,
(if you compensate the Hem. subdivision if you use IR), or a script to set all wanted to 256 on one click.