Page 2 of 2
Re: Far end of shadow noise
Posted: Thu May 30, 2019 4:21 pm
by lacilaci
I find intels denoiser far better than nvidias. It is definitely good enough for stills in most cases, not sure about anims. And what luxcore had before doesn't come even close.
This random skipped area problem is basicaly the only problem I ever had with it, and in case of final renders even that wouldn't matter.
Re: Far end of shadow noise
Posted: Thu May 30, 2019 4:34 pm
by CodeHD
lacilaci wrote: Thu May 30, 2019 4:21 pm
I find intels denoiser far better than nvidias. It is definitely good enough for stills in most cases, not sure about anims. And what luxcore had before doesn't come even close.
This random skipped area problem is basicaly the only problem I ever had with it, and in case of final renders even that wouldn't matter.
Exactly. The image I showed is the only one that ever failed for me, and it is far from a typical scene...
Re: Far end of shadow noise
Posted: Thu May 30, 2019 4:34 pm
by B.Y.O.B.
kintuX wrote: Thu May 30, 2019 4:17 pm
And a reminder, during NVidias first AI filtering/denoising demo it was noted that this tools are meant for better, faster, more efficient previewing!!! not final HQ renders.
That is not the case for Intel's OIDN, as stated by the main developer:
voxelium wrote:we didn’t optimize the filter for low SPPs and we never said that we did. I don’t know where does this come from.
Source:
https://blenderartists.org/t/intel-open ... ?u=b.y.o.b
More random skipped denoising
Posted: Wed Jun 05, 2019 10:05 am
by lacilaci
noise:
denoised:
scene has hundreds of thousands lights probably... but somehow works well with the exception of random non denoised areas...
Re: More random skipped denoising
Posted: Wed Jun 05, 2019 11:09 am
by FarbigeWelt
lacilaci wrote: Wed Jun 05, 2019 10:05 am
scene has hundreds of thousands lights probably...
Sure, this works very well. I think you can increase to millions. Who knows?
Re: Far end of shadow noise
Posted: Thu Jun 06, 2019 12:29 pm
by Dade
Dade wrote: Thu May 30, 2019 7:37 am
I can reproduce the problem but I can not see any error on our side.
I will upgrade the binaries to the latest Oidn version than, if the problem is still there, I will report behavior to Intel.
I have updated our Intel Oidn to v0.9.0 but the problem is still there. I have reported the problem to Oidn developers:
https://github.com/OpenImageDenoise/oidn/issues/33
Re: Far end of shadow noise
Posted: Thu Jun 06, 2019 12:35 pm
by lacilaci
Dade wrote: Thu Jun 06, 2019 12:29 pm
Dade wrote: Thu May 30, 2019 7:37 am
I can reproduce the problem but I can not see any error on our side.
I will upgrade the binaries to the latest Oidn version than, if the problem is still there, I will report behavior to Intel.
I have updated our Intel Oidn to v0.9.0 but the problem is still there. I have reported the problem to Oidn developers:
https://github.com/OpenImageDenoise/oidn/issues/33
thanks, here's an example of an area in reflection...
Unfortunately, I can't share any of those scenes
Re: Far end of shadow noise
Posted: Thu Jun 06, 2019 2:27 pm
by Dade
The problem has already been fixed in internal Oidn sources, the fix will be available with the next release
Re: Far end of shadow noise
Posted: Thu Jun 06, 2019 4:29 pm
by lacilaci
wow, that's great news!
Re: Far end of shadow noise
Posted: Fri Jun 07, 2019 4:12 pm
by kintuX
Really admirable co-op!!!