I decided to model a 'complex' indoor scene to test the abilities of OIDN when it comes to more complex edge situations. This scene contains 1.5 million triangles, but more importantly, over 20 million hair particles on the carpet, cussions, and grass outside the window. You'll notice that all (to my eye) detail in the hair particles, espectially on the carpet, are preserved, which is WAY ahead of other denoisers offered by Cycles or Renderman. The render ran for 21.5 minutes total, with a 20 minute halt condition for the render and 305 samples at the 20 minute mark.
The first is the non-denoised image, while the second is the final post-processed image. I encourage you to zoom into the shadows or carpet to check for noise! I don't understand how it's possible
Impressed with Interior Denoising
Impressed with Interior Denoising
Calling all Linux users: check out the new LuxRender 1.7 Flatpak! https://github.com/rrubberr/Flatpak-LuxRender
Light transport enthusiast - www.rrubberr.com
Light transport enthusiast - www.rrubberr.com
Re: Impressed with Interior Denoising
Very very nice result here . thank for sharing with us rrubberr.
More info ? hardware ? PGI on/off ? PGI debug view ?
More info ? hardware ? PGI on/off ? PGI debug view ?
Re: Impressed with Interior Denoising
I like it
Maybe the carpet could be a bit more trodden down?
Maybe the carpet could be a bit more trodden down?
Re: Impressed with Interior Denoising
Beautiful!rrubberr wrote: ↑Thu Mar 07, 2019 1:59 am I decided to model a 'complex' indoor scene to test the abilities of OIDN when it comes to more complex edge situations. This scene contains 1.5 million triangles, but more importantly, over 20 million hair particles on the carpet, cussions, and grass outside the window. You'll notice that all (to my eye) detail in the hair particles, espectially on the carpet, are preserved, which is WAY ahead of other denoisers offered by Cycles or Renderman. The render ran for 21.5 minutes total, with a 20 minute halt condition for the render and 305 samples at the 20 minute mark.
FinalNODN.png
FinalT1-01.jpeg
The first is the non-denoised image, while the second is the final post-processed image. I encourage you to zoom into the shadows or carpet to check for noise! I don't understand how it's possible