I have a scene that I am currently rendering. @ resolution 800 x 450 the render time is 00:53 seconds. At resolution 900 x 506 render time is 1:15. At 1280 x 1024, the render will sit at building light cache for 5+ minutes and move only a couple %. Using brute force or any other GI option, It basically wont render unless I leave it for a day.
Currently GI is on and I know this is where the issue is. I am using Primary (IR) Secondary (LC) with the low preset / basic settings.
I understand resolution changes the GI Calc time but within a couple hundred pixels, render time is going from 1 minute to almost 30 - 60 minutes. Is there something fundamental that I am missing? I am fairly new to VRay. Thanks for any advice.
Thanks for the reply Vlado. I been playing around with the quality settings in the GI tab and found some settings that seem to be working out pretty good. The dynamic memory limit was set to default of 512. How does that setting work and effect render time. In the scene there is a few sub d surfaces and will have displacement eventually. With the default settings though, I tried to do a full HD render and I let it sit for more than 24 hours and it was just past the GI stage. I think I still have some optimizations to learn about
Is is possible for me to get full forum access? I got my licence for my home today
The default for the dynamic memory limit is 400 MB and is too low for most purposes; depending on the amount of physical RAM, try a few gigabytes below that.
I agree with Vlado, first thing to try is to raise your dynamic memory limit.
Are your rendering with max VFB or Vray vfb ? Are you using the max bitmap pager ?
Just in case, few important things :
1-You need to know that some Vray parameters are strongly tied to the final resolution.
2-The Vray defaults settings are for a 640*480 resolution…
1-details :
- irradiance map Min rate and max rate. Each time your double the res, you can lower the min and max values to get identical result. For example -3/-1 (min and max) at 640*480 can be lowered to -4/-2 at 1280*960. Doubling the resolution is indeed an additionnal pass.
- DE (Irrmap Detail Enhance radius) needs also to be adjusted, dependant of resolution (if in screen mode)
- LC subdivs have to be doubled each time you double the res to get consistent results.
- Antialias can be lowered for high resolutions. each time the res double it’s equal to 1 subdiv (theorically.. in practice, it’s better to keep a bit more for security). But it’s a time gainer !
- you can also try to not use any image filter. Gives good results and render faster.
Last but important : Vray displacement options !
If your edge lenght is in pixel mode (View-dependant is checked), then 4.0 pixel can be a good value for 640*480, but can take infinite time in bigger resolutions !
Again, you need to raise this value while raising the resolution to avoid ultra complex computation of displace in background parts of your render.
Vlado, correct me if i’m wrong somwhere
Hope it helps, perhaps can you post some screengrabs to have a look ?
Thanks so much ! So a 1000 subdivs LC gives same quality at 640*480 and 4000*3000 ? damn i can’t believe that ! lol ! was absolutely sure that it was resolution dependant… So stupid am i.
I’m sure i were confused with PPT mode:
(from spot 3d, PPT examples: Increasing the resolution twice means that you will also have to increase the Subdivs twice to get the same quality)
right ?
Solidrocks will render way faster in next build for high res renderings as i was using a LC subdivs multiplier for all modes instead of only PPT
thanks Vlado !
For the PPT mode you are right, of course, but when using the light cache for secondary bounces only, this is not necessary. These examples show this: Chaos: Industry-leading design and visualization software - you can see that you get the same light cache, regardless of resolution.
