I’m rendering a scene where I set the Irradiance map min/max to -5/-2
When I render the status window says Prepass 1 of 1 … shouldn’t it be more that 1 prepass if the setting is -5/-2 ?
Also, can I use a saved irradiance map on the same scene/view when rendering at double resolution or is it better to lower the min/max settings and generate the IR directly on the larger render ?
Thanks
Are you using distributed rendering? When you do there’s always only one pass. If not then yes, I would expect 4 passes.
My understanding is if you use the saved IRmap on a larger render, it might start to show artifacts. I would imagine that small details would be the problem areas.
Yes I’m using DR with another 3 machines. So are you saying that each machine is doing 1 pass (total 4 passes) or the way it’s calculated is just different ? Is the final result of the IR still the same?
By the way, I have 8 threads in my processor but when rendering (with DR) my machine name shows up in 16 buckets instead of 8. Is this normal ?
Thanks
I don’t know if the machines actually perform 4 passes or do it all at once. But it will only show you it as one pass when using DR, so it doesn’t really matter. The final IR is basically the same, you just don’t get to see the passes.
The IRmap is just a bunch of brightness calculations taken at various sample points in the scene. the early passes have fewer calcs so they run faster. the later passes add sample locations, each pass adds more and takes longer. there should be no difference between running four passes and running one pass, as far as i know. (someone please correct me if i’m wrong about that, it would be good to know.)
that’s weird about the threads. never seen that before. is the DRspawner running on your machine also? Maybe it would generate extra buckets (if so, turn it off, it’s not helping you).
“Multipass” enabled at the IM options? In the past I compared multi pass against single pass and I found no advantage of multi pass.
A low res IM cache can be used for a high res final render. If you render 800x600 at -1 and later the final pass at 1600x1200 with the cached IM, than you get IM details like from -2, your IM pass is more under sampled in relation to the final pass.
Yes, that’s right - you have to set the IM pass higher in the low res render. So you set it for the quality you need for the final render.
ever test if it takes less time that way? So, turn off render final image, does the higher setting at low res take less time than at high res?
Micha, I’m not sure that I’m clear. OK, I understand now that DR does only one pass.
Now, if I’m rendering only in the main machine and I disable multipass in the IR options, the various passes are all done at once ? and you found no advantage in multipass … would then also be faster ?
Regarding generating a low res IM cache and using it in a higher res rendering, (so that I understand), right now for a 2400px (larger side) image I set the IR at -5/-2, but I could set it to say -4/-1 and generate it at 1200px and use it for the larger image and it will be equivalent to -5/-2 ?
Having said that, what is a good guide to follow for max/min settings vs resolution ? Is -5/-2 too low for 2400px ?
emil, where do you turn render final image off ? I have a bunch of views of the same scene so I’m setting the IR cache to incremental and I could generate the total cache at half resolution, save it, and load it to render the images at full res in one batch.
But where do you turn off ‘render final image’ option ? I cannot find it in any the VR option tabs …
http://www.lookinglass.us/vray4rhinoWiki/tiki-index.php?page=Global+Switches
It’s in global switches, halfway down on the left.
As for IRmap settings, it depends on the scene, but I find a max rate of -2 is usually too blotchy or leaks light too much for anything except a draft, at almost any resolution. But that’s for architectural interiors. Could be that product shots would be ok.
Oh, and the method you are suggesting should work fine, just don’t change anything except the camera view between renders.
I don’t think IRmap is resolution-dependent. Should be the same time for the same settings regardless of output. But Light Cache is resolution dependent if you have the settings set to screen, and of course the final pass is.
I always heard that IRmap is resolution dependent so for a bigger rendering you can get away with smaller max/min values.
I could be mistaken.
yes I think you’re right, I was confused.
I use a higher IM interpolation (30..40) or more IM subdivs and GI AO or detail enhancement to add the missing details.
Interesring. I’ll have to try it.