Page 1 of 3

Luxcore Vs Octane Preparation test

Posted: Thu Jan 30, 2020 11:51 am
by Sharlybg
Hi guys ! before i can deep dive into the war between theses two engine i want to make sure everything is set to the correct way.
At the end this should lead to a better understanding of the key benefit from a spectral engine Vs RGB renderer. The is done with the same HDRI all deph 40.

Octane Render 2.35 mn
Octane_settings.jpg
octane__2.35mn.jpg


Luxcore PGI 2.43 mn
Luxcore PGI set.jpg
Luxcore__PGI__2.43mn.jpg



Luxcore Path 10.44 mn
Luxcore Path set.jpg
Luxcore__Path__10.44mn.jpg


Without Caches algo i really wonder how octane manage to strike a so clean image in just 2.35 mn (A kind of AI or clamping ?). Luxcore PGI is fast but loose lot of light details while being darker.

Luxcore pure path can be faster but it is really hard to remove hotpixels without loose too much light details ( render become darker and darker). Note also that both engine use transparent shadow for glass.

Re: Luxcore Vs Octane Preparation test

Posted: Thu Jan 30, 2020 12:09 pm
by Dade
Have you noticed that the HDR background has different orientation between Octane and LuxCore rendering :?:

Re: Luxcore Vs Octane Preparation test

Posted: Thu Jan 30, 2020 12:39 pm
by Sharlybg
Dade wrote: Thu Jan 30, 2020 12:09 pm Have you noticed that the HDR background has different orientation between Octane and LuxCore rendering :?:
Yes just slightly octane transformation node is not very smooth and precise in addition it doesn't work when you enter precise number . you have to manually play with the slider to see change in the viewport. But this doesn't change the Performace/quality story here. But anyway for consistency i've update luxcore render.

Re: Luxcore Vs Octane Preparation test

Posted: Thu Jan 30, 2020 12:41 pm
by zuljin3d
You used portals in octane?

Re: Luxcore Vs Octane Preparation test

Posted: Thu Jan 30, 2020 12:46 pm
by Sharlybg
zuljin3d wrote: Thu Jan 30, 2020 12:41 pm You used portals in octane?
No Portals in octane.

Re: Luxcore Vs Octane Preparation test

Posted: Thu Jan 30, 2020 1:10 pm
by zuljin3d
Octane looks good, it was interesting to see a comparison in difficult conditions, in a scene with materials and textures.

Re: Luxcore Vs Octane Preparation test

Posted: Thu Jan 30, 2020 1:28 pm
by Sharlybg
zuljin3d wrote: Thu Jan 30, 2020 1:10 pm Octane looks good, it was interesting to see a comparison in difficult conditions, in a scene with materials and textures.
I want them to be on par before i can add more complexity. i've to find where Octane gain time without cache. if there is a kind of AI something i want to be sure it doesn't make the test Unfair.

Re: Luxcore Vs Octane Preparation test

Posted: Thu Jan 30, 2020 3:47 pm
by Sharlybg
More test this time with direct light only. max bounce = 1 Max sample = 1 for both.

Octane
octane__direct__sample 1.jpg

Luxcore
Luxcore__direct__sample 1.jpg

Observation :

Look like Octane is using a kind of heavy direct light clamping that kill everything even though Clamping by default is set to 1000000 in octane.

Re: Luxcore Vs Octane Preparation test

Posted: Thu Jan 30, 2020 4:21 pm
by Sharlybg
Now Direct light with more samples per pixel (500) :


Octane Max deph 1 + 500 samples
Octane__direct__sample 500.jpg


Luxcore Max deph 1 + 500 samples
Luxcore__direct__sample 500.jpg

Observation

Even with more samples per pixel look how dark the octane render is. now wonder if we can clamp direct light alone in Luxcore too at some point.

Re: Luxcore Vs Octane Preparation test

Posted: Thu Jan 30, 2020 4:31 pm
by lacilaci
Sharlybg wrote: Thu Jan 30, 2020 4:21 pm Now Direct light with more samples per pixel (500) :


Octane Max deph 1 + 500 samples

Octane__direct__sample 500.jpg



Luxcore Max deph 1 + 500 samples

Luxcore__direct__sample 500.jpg


Observation

Even with more samples per pixel look how dark the octane render is. now wonder if we can clamp direct light alone in Luxcore too at some point.
Luxcore is always using whatever you input +1 bounce, that's how exporter is set up, Simon pls. Correct me if wrong.