Sampling method ???

A forum where development builds are posted for testing by the community.
Forum rules
NOTE: The software in this forum is not %100 reliable, they are development builds and are meant for testing by experienced octane users. If you are a new octane user, we recommend to use the current stable release from the 'Commercial Product News & Releases' forum.
Post Reply
User avatar
bicket
Licensed Customer
Posts: 365
Joined: Thu Mar 01, 2012 10:46 am
Contact:

Hi Octane Team

I just wonder why the image sampler doesn't focuses on area that are noisy instead of working on the whole image ?
I wait 4 hours to complete a render that was 95% noise free after 30min.

Just a small aera was noisy but the renderer keep working on the the whole picture.
Is this a way to optimize render ?
Do the Otane Team have a plan to optimize the sampling ?
It would be a great performance improuvement.
PortFolio
4 x TITAN - Win 7 64 - Octane Rocks
User avatar
PAQUITO
Licensed Customer
Posts: 463
Joined: Wed Nov 21, 2012 4:50 pm
Location: Spain
Contact:

I wonder that exact thing about every unbiased render engine. Why not making an image based smart sampler? That is: make some samples - analyze the image to find noisier areas - redistribute de calculation density - start again.

or else: make some samples - Check the image for noise-free areas - Exclude that area from the calculation - start again.

That way when we have some areas very easy to clean (direct lit areas, for example), the sampler should just ignore it (base on an user-defined threshold or something), and focus in more needed areas. At the end of the render stage we will have the GPU working 100% in just small areas, such as small light sources and stuff like that. It would speed up the render greatly in my opinion. Is that possible?
AMD FX8350 8 cores. 16GBs RAM
GTX690
Win8 64.
Image
nagboy
Licensed Customer
Posts: 96
Joined: Thu Jan 03, 2013 5:32 am

Vray RT got this feature, it can't be difficult to implement such feature. Its a simple noise threshold filter. It would save us a lot of electricity and time.
manalokos
Licensed Customer
Posts: 441
Joined: Fri Nov 18, 2011 4:43 pm

I found that the best workaround if you have time is to separate the render into several different "passes", one for natural light and some other for different groups of artificial lights.
You can combine them afterwards with the advantage to be able to change relative intensities of each "pass", and greatly reducing the total render time.
User avatar
PAQUITO
Licensed Customer
Posts: 463
Joined: Wed Nov 21, 2012 4:50 pm
Location: Spain
Contact:

Could you explain how you do that? Sounds interesting.
AMD FX8350 8 cores. 16GBs RAM
GTX690
Win8 64.
Image
User avatar
bicket
Licensed Customer
Posts: 365
Joined: Thu Mar 01, 2012 10:46 am
Contact:

ok ! So, any news about bidirectional Pathtracing ?
PortFolio
4 x TITAN - Win 7 64 - Octane Rocks
MaTtY631990
Licensed Customer
Posts: 754
Joined: Thu Apr 08, 2010 8:38 pm

There was word quite a while back that the octane team will add this algorithm or similar dependent on research in the near future, and more specifically that we might see this implementation in the post version 1.9 releases. Although hopefully we might see this feature a bit closer.
User avatar
profbetis
Licensed Customer
Posts: 279
Joined: Tue Dec 21, 2010 11:00 am
Contact:

Lol you guys know that's exactly what the PMC kernel does right?
GTX 1080Ti 11GB (3x), Water-cooled
Intel i7-5820K 6-core @ 3.3GHz
Windows 10 Pro 64-bit, 32GB RAM
User avatar
bicket
Licensed Customer
Posts: 365
Joined: Thu Mar 01, 2012 10:46 am
Contact:

profbetis wrote:Lol you guys know that's exactly what the PMC kernel does right?
maybe can you explain use... with simple words :?
PortFolio
4 x TITAN - Win 7 64 - Octane Rocks
Post Reply

Return to “Development Build Releases”