hello, my 1st post..
I’m new to Vray(6yr Brazil user), i’m trying to lock down some settings for exterior daylight scences, here’s what i’ve come up with, any suggetions on other setups??
thanks…
If you have the time & render power it’s always best to use Brute force as it will produce the most detailed result.
My preference is
Brute Force -First bounce
Light Cache - Secondary bounces
Hope this helps
thanks 3DMK… right now as you can see in my tests, my brute force test has the most noise, what key settings should i play around with to eliminate the noise when using brute force as the fist bounce..
if accuracy in lighting is not essential to you, you may want to explore ambient light + occlusion method. Very fast, very clean.
There are these settings:
Where the noise is controlled by the Noise threshold.
actually, i have entirely moved away from dmc sampler. Its all adaptive subdivision from here on. Faster time, less noise…its great.
The beauty of Brute Force is it’s simplicity.
Simply increase the Brute Force subdivisions to get a cleaner solution.
You could also try using the Image sampler or Global DMC settings to control it as well, but they work globally so they effect other noise than just GI noise.
eg [glossy reflections, dof, etc]
I think us 3D guys worry about grain too much sometimes…I am starting to appreciate the noise in images these days. You get noise in real life photos as well so I think it actually looks more realistic than the very smooth GI from irradiance & light Cache.
Since using BF for the last 6 jobs I haven’t a client complain about noise in images yet. So far so good
Hope this helps
well there is different kinds of grain out there. Problem with sampling is that you will get more grain in the darker areas, where as film grain is uniform.
Also for film, the 3d images must have no grain or the least possible since the grain is applied in comp but also, render grain is not RGB and lacks many film properties, it would be very hard to match grain from 3d render to specific film grain.
this is interesting, I gave up using adaptive subdivisions ages ago (about the same time I quit using irr maps for good). but this post of yours almost makes me want to give it another go.. is it really faster? even with glossy stuff, dof, mblur..?
well, dmc is cool, but it offers almost too much control and that leads to errors. For us the toughest thing was to get smooth motion blur, with least possible grain. Dmc always produced some form of noise, while adaptive subdivision made much cleaner result in faster time.
Brazil and MR use only adaptive subdivision and they manage, so I thought I’d try it out for a while. I can render 2k frames for finals between 1-3 hours per frame with final motion blur+aa settings.
thanks everyone for your suggestions..I just watched this video, really good explaination on VRay Image Sampling.
http://www.thegnomonworkshop.com/store/products/tutorials/cni03/
I guess I’ll have to give it another go then. actually I remember finding it quite slower than dmc especially when mblur was concerned, but of course it could have been me not knowing what I was doing.. another reason for which I stopped using it was ram usage, but it was back in the 32bit days. I guess memory should be less of a problem today.
post your settings, let see if we can improve the time…if its slow.
I’ll definitely do it as soon as I get to it. need to find the time to test it seriously in between a couple of ongoing projects.
btw, I’m sorry for the little hijack. i’ll stop it now.