LuxCoreRender Forums

Any idea on that issue ?
BTW here is the file with the problematic smoke.

File

problem is band node, either use color math or leave the band sliders at default values and instead change absorption depth value.

I guess the fundamental problem is that you are reaching the limits of floating point precision by using a band from 0 to 0.02.

Hi, I've started to do some practice with smoke in Blender (the "old" one, I've tried the new mantaflow smoke but I don't like it, at the moment).
After some days of frustrating artifacts, I gave up with Cycles and tried with LuxCore: no more artifacts! yeee!

Now I would like to ask if there are any recommended settings, or tricks, to optimize render time with Heterogeneous Volume. I'm a beginner with LuxCore and I'm not so confident with all those settings and wiki descriptions. Thank you!

one more thing: I've noticed that LuxCore rendered viewport is much more slower than the Cycles one, when I change one setting (e.g. absorption depth of heterogeneous volume) the interface freezes and I need to wait approx. 30 sec to see the result. It's a normal behaviour or I can change some settings? Again, thanks!

(my PC: i7 8700K OC, GTX 1070Ti, 16GB RAM)

pixel_peeper wrote: ↑Mon Jan 27, 2020 10:42 am Now I would like to ask if there are any recommended settings, or tricks, to optimize render time with Heterogeneous Volume.

I would try to use as small path depths as possible, use only few lights, disable multibounce if you can afford it (it affects the look quite a bit).
Other than that, not sure, homogeneous volumes are complex things to compute.

Maybe this file helps to give you some ideas: viewtopic.php?f=2&t=737#p7633

pixel_peeper wrote: ↑Mon Jan 27, 2020 10:42 am one more thing: I've noticed that LuxCore rendered viewport is much more slower than the Cycles one, when I change one setting (e.g. absorption depth of heterogeneous volume) the interface freezes and I need to wait approx. 30 sec to see the result. It's a normal behaviour or I can change some settings? Again, thanks!

Currently the whole smoke data of the current frame is re-exported whenever you change some material setting.
This could be optimized in the future.

B.Y.O.B. wrote: ↑Mon Jan 27, 2020 12:18 pm I would try to use as small path depths as possible, use only few lights, disable multibounce if you can afford it (it affects the look quite a bit).
Other than that, not sure, homogeneous volumes are complex things to compute.

Maybe this file helps to give you some ideas: viewtopic.php?f=2&t=737#p7633

Ok, thanks! I need to accept that is slow, I just asked to be sure and I didn't expected some magic trick to dramatically reduce rendere time.

B.Y.O.B. wrote: ↑Mon Jan 27, 2020 12:18 pm Currently the whole smoke data of the current frame is re-exported whenever you change some material setting.
This could be optimized in the future.

Good!

I'm quite happy of these result, I prefer the first one. Both are denoised with OIDN.

Heterogeneous Volume Heterogeneous Volume with Multiscattering (render time is 31:46, I don't know why the statistics panel shows 0:58) Volume setup

Wow, looking great!

pixel_peeper wrote: ↑Mon Jan 27, 2020 3:54 pm Heterogeneous Volume with Multiscattering (render time is 31:46, I don't know why the statistics panel shows 0:58)
multiscatter.jpg

Multi-scattering is a beauty.

You can "play" with the step size to improve rendering times (i.e. disable auto step size): if you double the step size, you (may) cut in half the rendering time for the same amount of samples (more or less, if the bottleneck is the heterogeneous volume rendering). However it should increase the amount of noise too.

thanks B.Y.O.B. and Dade! I'll try!

at the moment the only thing that I'm unable to control properly is the color, if I change the scattering color I can't predict the color as I expect, if I use the absorption color the whole domain has this color, like an overlay. Myabe I should tune the absorption depth?