LuxCoreRender Forums

For comparison, this is a normal hybrid rendering:


Well, it isn't bad when compared with the starting point.

IT look to gorgeous to be real ! well done. Time to go to the pool.

Please where do you leave ?

Dade wrote: Tue Sep 03, 2019 9:57 am It is good but a bit too blurred/noised/slow to render SDS paths.

how slow?

Also, Dade in case you haven't noticed:

Dade wrote: Thu Aug 29, 2019 6:21 pm

epilectrolytics wrote: Thu Aug 29, 2019 4:22 pm In the paper they mention it runs on GPU, Iooks like it will also be a fast solution

It will run on CPU in our case because the light tracing is done there in our Hybrid Back/Forward path tracing but it isn't a big deal because the CPU will really handle only caustics and SDS paths and nothing else.

Dade wrote: Tue Sep 03, 2019 9:57 am It is good but a bit too blurred/noised/slow to render SDS paths.

Maybe it should better run on GPU as was mentioned in the paper, in a scene with a pool a significant part of the image will need fast SDS solving.
When there is also a noise problem, we need to ensure that OIDN can deal with that (albedo pass including specular surfaces).

how does it work with clamping? Can we use low clamping values for pathtracer and this lighttracing+ will still use unclamped values? Or this cannot be disconnected from clamping?

lacilaci wrote: Tue Sep 03, 2019 11:19 am

Dade wrote: Tue Sep 03, 2019 9:57 am It is good but a bit too blurred/noised/slow to render SDS paths.

how slow?

It is about a 2000 samples/pixel rendering done on the CPU with 50% of eye paths and 50% light paths (so there are about 1000 light trace samples).

My main concern at the moment it is about the initial blur, it doesn't seem to disappear over time but it should (and may be a bug).

lacilaci wrote: Tue Sep 03, 2019 11:22 am Also, Dade in case you haven't noticed:

I should have fixed this problem.

lacilaci wrote: Tue Sep 03, 2019 12:14 pm how does it work with clamping? Can we use low clamping values for pathtracer and this lighttracing+ will still use unclamped values? Or this cannot be disconnected from clamping?

It works like clamping with hybrid rendering or BiDir. So it should be disconnected from clamping (and there start to be very little, if no source at all, of fireflies, if you use the full "arsenal" of features).

It is good but a bit too blurred/noised/slow to render SDS paths.

Didn't notice this part. hope it can be solved with more development.