Opaque object behind specular material causes lots of noise

Forums: Opaque object behind specular material causes lots of noise
Newtek Lightwave 3D (exporter developed by holocube, Integrated Plugin developed by juanjgon)

Moderator: juanjgon

Opaque object behind specular material causes lots of noise

Postby scooternva » Wed Apr 14, 2021 2:47 pm

scooternva Wed Apr 14, 2021 2:47 pm
I'm seeing a LOT of noise in a scene that uses the Path Tracing kernel where I have an opaque object nestled up very close behind a transparent specular object. The transparent object is supposed to be a windscreen on an aircraft. It's using an Octane Specular Material with the GGX BRDF model, an IOR of 2.417 (it's supposed to be diamond), a dispersion of 0.1, roughness 1% and film width of 1% with a film IOR of 1.7. The opaque object directly behind the windscreen is a blast shield that in the closed position completely blocks the window. The shield is surfaced with a simple plain white (RGB 248,248,250) Octane Glossy Material for now.

When the render first starts, the windscreen looks like dancing diamond dust was sprayed on its surface as OctaneRender resolves the image.

Noise example.png

At first I thought the noise was caused by rays bouncing back and forth between the windscreen and the blast shield, but the windscreen is not double-sided so I don't see how that could be the problem. And yet, as you can see in the image, another part of the windscreen where the blast shield is open only shows noise for a split-second before it mostly clears up.

Noise on the closed windscreen never really quite resolves itself, and the denoiser doesn't quite get rid of what's left; leaving faint multi-colored blobs where the noise used to be. And since the denoiser isn't denoising exactly the same way on every frame, I've got some ugly animated blob artifacts dancing around where the blast screens are closed. I get why the denoiser can't do a perfect job; the camera and the subject are both in motion relative to each other and both are in motion relative to the scene light sources, but this noise is ruining my scene.

On the specular windscreen material I've tried turning off the film width, setting dispersion to 0.0, and setting the windscreen IOR to 1.0, with no effect. I also tried adjusting the ray epsilon higher which did solve the problem, at the expense of my transparent objects no longer being transparent. The only thing that sort of works is setting the noise threshold extremely low (0.01%) and setting the maximum samples very high (8192), but holy hell the render times explode in duration. I've attached my camera and Path Tracing kernel settings below:

Camera Thin Lens Settings.png

Path Tracing Kernel Settings.png

Is there anything else I could try that won't be such a hit on image quality and/or rendering time? Thanks...
Octane 2020.2.3.1 | LightWave 2019.1.5 (3134) | Win10 Pro 19041.1052 | 2 x Gigabyte AORUS RTX 2080 Ti Xtreme w/NVLink | Gigabyte AORUS Xtreme X399 | AMD Ryzen 2950X | 64 GB RAM | SSDs: Samsung 970 PRO 512 GB + Seagate FireCuda 510 2 TB | blog.gammans.zone
User avatar
scooternva
Licensed Customer
Licensed Customer
 
Posts: 88
Joined: Thu Jul 04, 2019 6:13 pm
Location: Suburban Washington DC

Re: Opaque object behind specular material causes lots of noise

Postby scooternva » Wed Apr 14, 2021 3:13 pm

scooternva Wed Apr 14, 2021 3:13 pm
AHA! It's caustics that are causing the problem--specifically, the caustics of light coming through the windscreens and striking the blast shields. After I posted this I tried using the Direct kernel and the problem disappeared (along with the more interesting-looking refraction effects on the windscreen material). That's when I realized it was caustics; it's one of the few things that the Path Tracing engine has that the Direct kernel doesn't.

Setting caustic blur to 100% results in a noise-free image that looks like what the Direct kernel renders, at the expense of loss of image quality. The blurring actually darkens the windscreens with the blast shields closed, which is not what I want; the blast screens are almost white so they shouldn't look dark gray. I'm now experimenting with caustic blur percentage levels and max samples to find an acceptable balance between render time and image result. If there are any other suggestions I should try I'll gladly welcome them... thanks!
Octane 2020.2.3.1 | LightWave 2019.1.5 (3134) | Win10 Pro 19041.1052 | 2 x Gigabyte AORUS RTX 2080 Ti Xtreme w/NVLink | Gigabyte AORUS Xtreme X399 | AMD Ryzen 2950X | 64 GB RAM | SSDs: Samsung 970 PRO 512 GB + Seagate FireCuda 510 2 TB | blog.gammans.zone
User avatar
scooternva
Licensed Customer
Licensed Customer
 
Posts: 88
Joined: Thu Jul 04, 2019 6:13 pm
Location: Suburban Washington DC

Return to Lightwave 3D


Who is online

Users browsing this forum: No registered users and 6 guests

Thu Mar 28, 2024 9:41 am [ UTC ]