LuxCoreRender Forums

Dade wrote: Fri Mar 01, 2019 1:44 pm I have fixed this damn problem.

Congrats !
You are a god!!!

lacilaci wrote: Fri Mar 01, 2019 4:30 pm How do I know I have actually the newest build...
I'm asking cause azure shows linux version build already and I downloaded it and I see no difference in rendering.
Is it possible that even though azure shows upload succeeded that I still have to wait a while?

Builds are uploaded to GitHub at the very end as far as I remember.

Dade wrote: Fri Mar 01, 2019 5:26 pm

lacilaci wrote: Fri Mar 01, 2019 4:30 pm How do I know I have actually the newest build...
I'm asking cause azure shows linux version build already and I downloaded it and I see no difference in rendering.
Is it possible that even though azure shows upload succeeded that I still have to wait a while?

Builds are uploaded to GitHub at the very end as far as I remember.

Alright, will try the fix tomorrow then

lacilaci wrote: Fri Mar 01, 2019 5:35 pm

Dade wrote: Fri Mar 01, 2019 5:26 pm Builds are uploaded to GitHub at the very end as far as I remember.

Alright, will try the fix tomorrow then

Yup, first builds for all platforms are completed, then all of them are uploaded to GitHub.

Well.. So I couldn't wait and downloaded just now... And it seems it is just as it was, nothing changed (tried linux opencl version) /tried 2 different hdri's and while sun and sky give 1:1 result against Pathtracing HDRI still has a massive difference.

cache unclamped render: no cache unclamped: Tried luxcoreui, tonemapping is different but difference is still there:

I have exposed the caustic cache merging options in BlendLuxCore.

By the way, this is a rendering with merge radius scale = 1:


Is this expected behaviour, or a bug?

B.Y.O.B. wrote: Fri Mar 01, 2019 7:18 pm By the way, this is a rendering with merge radius scale = 1:

Yes, with 1.0 you are not allowing any overlap (i.e. merging everything), the default value is 0.25 to allow a 75% overlap.

So again, to make sure I re-downloaded both addon and latest linux daily and ran few example tests with...

no-cache: cache: no-cache - different hdri: cache - different hdri: no-cache benchmark scene + hdri: cache benchmark scene + hdri: no-cache benchmark scene (Sun&Sky): cache benchmark scene (Sun&Sky): So, the issue is still present here and only sun and sky looks proper - even more light that PT only, but that's probably due to better propagation of sun in scene.

The difference is sometimes bigger sometimes smaller, but if it can work properly with luxcore lightsources I think hdri should too.

I have renamed the "Usage Threshold Scale" to "Brute Force Radius Scale" and added some tooltips:
https://github.com/LuxCoreRender/BlendL ... ebb26be9de
Dade, maybe you can take a look at this commit and check if my explanations in the descriptions are correct.

B.Y.O.B. wrote: Sat Mar 02, 2019 11:30 am I have renamed the "Usage Threshold Scale" to "Brute Force Radius Scale" and added some tooltips:
https://github.com/LuxCoreRender/BlendL ... ebb26be9de
Dade, maybe you can take a look at this commit and check if my explanations in the descriptions are correct.

i think it's more like a "mix" between cache and pure brute force than radius