Page 5 of 7

Re: Octane render experiments

PostPosted: Wed Oct 20, 2021 3:48 pm
by whersmy
Thanks elska! Funny I always just set SR to 10k, but tweaking might help.

I did notice some colour difference in the passes. I think I read somewhere that`s a bug, but not sure.
The main pass differs in colour from the denoised pass.

Re: Octane render experiments

PostPosted: Wed Oct 20, 2021 4:04 pm
by elsksa
whersmy wrote:Thanks elska! Funny I always just set SR to 10k, but tweaking might help.

Incremental values of 1 should suffice, if there are 2 lights and the second light requires a higher sampling priority (what Sampling Rate does in very short), a Sampling Rate value of 2 should suffice.

whersmy wrote:I did notice some colour difference in the passes. I think I read somewhere that`s a bug, but not sure.
The main pass differs in colour from the denoised pass.

I believe it has been reported in the forum already. I remember seeing it but I would recommend to double check via the forum-search bar about it. It might even have been fixed in a 2021.1 RC build, that is also to verify as I am not entirely certain.

Re: Octane render experiments

PostPosted: Wed Oct 20, 2021 9:00 pm
by whersmy
Will look into it :)

Trying out different lighting. Also gonna test out the random tiles script this week

Re: Octane render experiments

PostPosted: Fri Oct 22, 2021 8:16 am
by whersmy
Another one with the lamp turned on. Really like how subtle ACES and 16-bit EXR`s give control on lighting. :)

Re: Octane render experiments

PostPosted: Fri Oct 22, 2021 11:16 pm
by whersmy
Seems like the partition / random tiles generator script works nice. A very basic setup already really convinces. Here`s the node-setup. Will do a more random example in the upcoming days. Really digging the 3080 and stability of 2021.1 XB1! After three days 0 crashes yet when using textures over 10k x 3k pixels, Universal materials, LUA scripts, ACES and 8K HDRI`s. :)

Re: Octane render experiments

PostPosted: Tue Oct 26, 2021 2:35 pm
by whersmy
Getting the random tiles for the floor a bit more decent. Also reset the up-vector so the camera is correct again.

Re: Octane render experiments

PostPosted: Thu Oct 28, 2021 1:57 am
by whersmy
Another try-out for more soft light and post-pro in PS with ACES.

Re: Octane render experiments

PostPosted: Sat Oct 30, 2021 8:02 pm
by whersmy
Trying out direct sunlight. Coming from a 1060GTX to a 3080 is just the best! :)

Re: Octane render experiments

PostPosted: Sat Oct 30, 2021 10:42 pm
by whersmy
And one with more diffuse light. About 1.5 hours on a 3080. It seems the universal material (on this scene f.e. wood) with mix textures, invert textures and color correction nodes do take way longer to clean up, without any real visual difference. :idea: Gonna optimize the scene further.

Edit: attached a version saved as 8-bit from 16-bit. It looks like contrast and depth is lost when saving as jpeg directly from 16-bit in PS :o

Re: Octane render experiments

PostPosted: Tue Nov 02, 2021 8:40 am
by whersmy
A few more shots. Next up is optimizing the scene. Theres a baked OSL Random texture on the displacement on the cloth. Universal material with the layered options on the ceramic pot. Lit by 1 HDRI. Used ACES again with some curves adjustment.