Caustic issues

Use this forum for general user support and related questions.
Forum rules
Please upload a testscene that allows developers to reproduce the problem, and attach some images.
Post Reply
jgrover110
Posts: 47
Joined: Tue Apr 17, 2018 1:07 pm

Caustic issues

Post by jgrover110 »

I've got 2 issues when trying to render caustics with the cache and light tracing.

First issue is the fireflies. These paths really should be disabled because the caustics are handled by both the cache and light tracing, but there presence make my renders basically unusable.

Second issue is that the brightness of the caustics doesn't seem to be correct. The material of the cube is 100% reflective, so the caustic and reflection of the caustic should be the same brightness, yet the reflection is noticeably brighter. This may be an issue with the metropolis estimation, but I'm not entirely sure about it

The scene attached below used a daily build from a day ago, and a daily build of Blender 2.90
render.png
render.blend
(1.13 MiB) Downloaded 186 times
User avatar
Theo_Gottwald
Posts: 109
Joined: Fri Apr 24, 2020 12:01 pm

Re: Caustic issues

Post by Theo_Gottwald »

In Blender 2.9 i have several issues, especially crashes when using Mantaflow, but also with Luxcore.
So i recommend to stick - at least when there are problems - with 2.83.
Get Blender-Quickbuttons and configure over 100 Blender-Buttons individual to your needs.
Visit my YouTube-Channel: Theo's Fun Videos and watch several Blender related Videos.
Join @Dreamstime and sell your Renderings to the world.
jgrover110
Posts: 47
Joined: Tue Apr 17, 2018 1:07 pm

Re: Caustic issues

Post by jgrover110 »

These issues persist when using 2.83
This isn't an issue with blender, it's an issue with LuxCore
Martini
Posts: 125
Joined: Fri Nov 23, 2018 11:36 am
Location: Australia

Re: Caustic issues

Post by Martini »

jgrover110 wrote: Mon Jul 13, 2020 10:51 am First issue is the fireflies. These paths really should be disabled because the caustics are handled by both the cache and light tracing, but there presence make my renders basically unusable.
I suspect it's the Light Tracing, but I can't really speak to that.
jgrover110 wrote: Mon Jul 13, 2020 10:51 am Second issue is that the brightness of the caustics doesn't seem to be correct. The material of the cube is 100% reflective, so the caustic and reflection of the caustic should be the same brightness, yet the reflection is noticeably brighter. This may be an issue with the metropolis estimation, but I'm not entirely sure about it
I get the same result as you. Blender 2.83, LuxCore 2.4 beta 1


If I use PhotonGI only (no Light Tracing), at 12K samples, the brightness is equal, although it does seem like it might be a little too bright overall?
Denoised. 12K samples PhotonGI (no Light Tracing)
Denoised. 12K samples PhotonGI (no Light Tracing)
Raw render. 12K samples PhotonGI (no Light Tracing)
Raw render. 12K samples PhotonGI (no Light Tracing)

If I disable the PhotonGI cache and use Light Tracing, the result is still quite noisy at 12K samples; the brightness is equal.
Denoised. 12K samples Path + Light Tracing (no Photons)
Denoised. 12K samples Path + Light Tracing (no Photons)
Raw render. 12K samples Path + Light Tracing (no Photons)
Raw render. 12K samples Path + Light Tracing (no Photons)

If I disable the Light Tracing as well (pure path tracing), the brightness is equal, although a little noisier at 12K samples.
Denoised. 12K samples Path (no Light Tracing, no Photons)
Denoised. 12K samples Path (no Light Tracing, no Photons)
Raw render. 12K samples Path (no Light Tracing, no Photons)
Raw render. 12K samples Path (no Light Tracing, no Photons)

With BiDir Metropolis at 2K samples I get almost the correct brightness, it's just a little brighter in the reflection. :?
Denoised. 2K samples BiDir Metropolis
Denoised. 2K samples BiDir Metropolis
Raw render. 2K samples BiDir Metropolis
Raw render. 2K samples BiDir Metropolis

With BiDir Sobol at 2K samples I get equal brightness, it's just... very noisy and will take a long time to converge.
Denoised. 2K samples BiDir Sobol
Denoised. 2K samples BiDir Sobol
Raw render. 2K samples BiDir Sobol
Raw render. 2K samples BiDir Sobol

Conclusion:

In 2 of the 3 cases where Metropolis was used (BiDir, Light Tracing + PhotonGI) the brightness was not equal. However, when Light Tracing was used on its own, the brightness was equal.

With Light Tracing + PhotonGI the result feels quite like their contributions are additive, so you get the brightness from PhotonGI, plus the brightness from Light Tracing.


Hope this helps :geek:
AMD Ryzen Threadripper PRO 5995WX 64-Cores | 2x Gigabyte RTX 4090 Gaming OC
ASUS Pro WS WRX80E-SAGE SE WIFI | 256GB Kingston Server Premier ECC Unbuffered DDR4
Post Reply