BlendLuxCore Development
Re: BlendLuxCore Development
Yes, that's possible. It's already there in LuxCore, we just have to use it in the addon.
Re: BlendLuxCore Development
Basic AOVs are now in master.
Todo: https://github.com/LuxCoreRender/BlendLuxCore/issues/18
Todo: https://github.com/LuxCoreRender/BlendLuxCore/issues/18
Re: BlendLuxCore Development
I'm currently implementing a faster AOV import routine in C++.
Managed to get the import time of the RGB AOV from around 0.7 seconds down to 0.005 seconds
Managed to get the import time of the RGB AOV from around 0.7 seconds down to 0.005 seconds
Re: BlendLuxCore Development
Results of the framebuffer acceleration (importing one RGB pass with 1920x1080 pixels):
Before:
Now:
About 0.3 seconds of each measurement are caused by engine.end_result(result). I don't think we can optimize this because it's a Blender function.
The accelerated C++ function accounts for about 0.07 seconds of the time in the new measurements.
Before:
Code: Select all
FrameBuffer draw took 1.841s
FrameBuffer draw took 1.744s
FrameBuffer draw took 1.378s
FrameBuffer draw took 1.604s
Code: Select all
FrameBuffer draw took 0.493s
FrameBuffer draw took 0.358s
FrameBuffer draw took 0.320s
FrameBuffer draw took 0.328s
The accelerated C++ function accounts for about 0.07 seconds of the time in the new measurements.
Re: BlendLuxCore Development
Yes, because it is enabled in Blender by default (in Cycles it's the same).
It probably should not be, but if I don't enable it by default, there is a black "Depth" pass in the render result. I can't remove it unfortunately.
However if you think it costs too much RAM or so I can disable it by default and we just have to live with the black "Depth" pass.
It probably should not be, but if I don't enable it by default, there is a black "Depth" pass in the render result. I can't remove it unfortunately.
However if you think it costs too much RAM or so I can disable it by default and we just have to live with the black "Depth" pass.
Re: BlendLuxCore Development
It costs both a bit of ram and performance, not a big deal but it looks like a waste.B.Y.O.B. wrote: ↑Thu Feb 15, 2018 12:53 pm Yes, because it is enabled in Blender by default (in Cycles it's the same).
It probably should not be, but if I don't enable it by default, there is a black "Depth" pass in the render result. I can't remove it unfortunately.
However if you think it costs too much RAM or so I can disable it by default and we just have to live with the black "Depth" pass.
Re: BlendLuxCore Development
I found a way to disable Blender's Z-pass by default, so the problem is now solved and the DEPTH AOV is disabled by default
-
- Supporting Users
- Posts: 141
- Joined: Tue Jan 09, 2018 6:48 pm
Re: BlendLuxCore Development
@Simon
You run into trouble using:
Flag is an python 3.6 feature but next blender will stay with py 3.5 afaik.
Verify by testing in master ( py 3.6 ) vs. 2.79_rc latest src ( 3.5 )
You run into trouble using:
Code: Select all
from enum import Flag, auto
Verify by testing in master ( py 3.6 ) vs. 2.79_rc latest src ( 3.5 )
Re: BlendLuxCore Development
Damn it. I actually checked if it was supported, but I used the wrong Blender version by mistake.
Thanks, I'll revert it.
Thanks, I'll revert it.