Direct Light cache: hit rate not good
Forum rules
Please upload a testscene that allows developers to reproduce the problem, and attach some images.
Please upload a testscene that allows developers to reproduce the problem, and attach some images.
Re: Direct Light cache: hit rate not good
always with 1 sampling path?
So, I open the scene I shared, enable sun&sky and and stop?
So, I open the scene I shared, enable sun&sky and and stop?
Re: Direct Light cache: hit rate not good
Yes but with the fix to the number of light sources
Re: Direct Light cache: hit rate not good
I'm currently rendering it with sun+sky and 5 diffuse/glossy/spec/total path depth.
edit: here it is.
edit: here it is.
Re: Direct Light cache: hit rate not good
Here we are
This is with sun enabled 2min22sec (total time) This is with double resolution (actually my daily output resolution) 33min total time ( total depth 8, diffuse&glossy&specular 4) This is the denoised image I'd like to save the lightcache solution, load it and render this scene with GPU....
In any case a great improvement for this kind of scenarios.
Great job
edit: I didn't understand how to denoise the image with different settings after the render is finished. The "run denoiser" button is always grey (disabled)
This is with sun enabled 2min22sec (total time) This is with double resolution (actually my daily output resolution) 33min total time ( total depth 8, diffuse&glossy&specular 4) This is the denoised image I'd like to save the lightcache solution, load it and render this scene with GPU....
In any case a great improvement for this kind of scenarios.
Great job
edit: I didn't understand how to denoise the image with different settings after the render is finished. The "run denoiser" button is always grey (disabled)
Re: Direct Light cache: hit rate not good
Denoising after the render is finished is not possible.
I'd like to say "currently" but I fear it is not really solvable because of Blender API limitations (see https://github.com/LuxCoreRender/BlendL ... -398390095)
Re: Direct Light cache: hit rate not good
So, if I'd like to test different settings, what should be the workflow?
Something like:
- launch render
- when I'm quite satisfied of the sampling, I press Pause of the render
- then in the image editor I change the denoiser settings
is it correct?
Something like:
- launch render
- when I'm quite satisfied of the sampling, I press Pause of the render
- then in the image editor I change the denoiser settings
is it correct?
Re: Direct Light cache: hit rate not good
It works quite well, good.
I'm going to add the option to have persistent light cache, saved on file: for camera-only animations it will be possible to use the same cache for all frames. OpenCL support will be the next step once things are settle down.