interior test [Distributed Path integrator explained]

Please use this forum for general user support and related questions.

Moderator: coordinators

Forum rules
Please include your operating system type/version, LuxRender version and Exporter version used when submitting a support post.

Make sure you have read the Release forum thread for Release and RC (Release Candidates) builds as these threads contain information on known problems and workarounds: Test Builds Forum

Re: interior test [Distributed Path integrator explained]

Postby jeanphi » Wed Sep 15, 2010 1:31 pm

jensverwiebe wrote:27 s/px ? Thats nothing !!!
You can expect good renders > 500s/px.

With distributedpath, you can get an exponentially growing number of rays for a single sample so it is possible to get clean results with much less samples but the samples will take much longer to compute.

Jeanphi
jeanphi
Developer
 
Posts: 6573
Joined: Mon Jan 14, 2008 7:21 am

Re: interior test [Distributed Path integrator explained]

Postby brankovukelic » Wed Sep 15, 2010 1:51 pm

Other materials are ok. I've already had trouble with them so I fixed all of them... except the emitter, which is not visible. Does the appearance of the emitter affect the scene? It's a matte material.

Also, does swapping the emitter for a area light do anything useful? (Sorry I ask a lot, but it takes such long times to see if these things have any effect.)

EDIT: Just noticed the default emission color for area light (this is luxblend25 under Blender 2.54, btw) is 0.8 0.8 0.8, so I've changed my emitter mesh accordingly. Keeping my fingers crossed. eff is up to ~1500%, as well, which sounds good, right? :)
brankovukelic
 
Posts: 23
Joined: Sat Jun 05, 2010 10:52 am

Re: interior test [Distributed Path integrator explained]

Postby brankovukelic » Wed Sep 15, 2010 4:48 pm

Nope. Tweaking the emitter mesh material did nothing. I still have more or less the same amount of fireflies.
brankovukelic
 
Posts: 23
Joined: Sat Jun 05, 2010 10:52 am

Re: interior test [Distributed Path integrator explained]

Postby GioCa » Fri Nov 05, 2010 4:58 am

Hi all, first of all thanks for this interesting thread. After reading most of it I'd ask you a little help to nail down a very basic guide to choose the right settings:

If the scene is exterior or not complex interior go for...
Sampler: Low discrepancy;
Pixel sampler: Hilbert;
Pixel samples: start from 8 and go up as needed.
(not complex means not many windows / openings and generally simple light paths)

If the scene is complex interior go for Metropolis.

So I guess the most effective rule for a newbie could be Nizu's sentence: "Try LD+HILBERT for maximum speed, but if pic is still very noisy at 32-64 spp consider using MLT".

Obviously as experience grows many other considerations shall be done, the first to be learning Surface Integrator options.
GioCa
 
Posts: 2
Joined: Wed Aug 11, 2010 8:29 am

Re: interior test [Distributed Path integrator explained]

Postby binarycortex » Sat Dec 04, 2010 4:08 am

It has been my experience that bidir (eye path + light path) is the only thing that can produce caustics in a reasonable amount of time, even the regular path tracer (eye path only) has trouble finding them. The reason is because caustics are formed by very improbable paths to the light (when using eye path only). Bidir overcomes this because it traces from the light source (light path) as well as from the camera (eye path). LD+Hilbert is good for animations when you want predictable results every time. It's also good for preview renders.
Competition Coordinator.
Current Competition: Math is Beautiful / Abstract Wallpaper

Member of the first official jeanphi-fan club
User avatar
binarycortex
Developer
 
Posts: 1501
Joined: Fri Feb 22, 2008 10:44 pm

Re: interior test [Distributed Path integrator explained]

Postby Eneen » Sun Dec 05, 2010 5:22 pm

Hi!

First things first: hello All, because I'm new here (new to Lux and new to Blender 2.5)

I didn't want to create new topic, so I'm writing here. I made small test, one with MLT+BiDir and one MLT+DP (if I can use such combination), but both images are different, despite of scene beeing same. DP doesn't want to render blend mat (transluent+none) on transparent curtains (single face) and makes sun light much brighter. Also window glass is really transparent (no noise?), it's arch glass. Those are 64S/p pics. Which combination is more "correct"?
Attachments
t1.jpg
MLT+DP
t2.jpg
MLT+Bidir
Eneen
 
Posts: 11
Joined: Fri Nov 26, 2010 8:19 am

Re: interior test [Distributed Path integrator explained]

Postby jeanphi » Sun Dec 05, 2010 5:44 pm

Hi,

DP is extremely difficult to use, you first need to understand every parameter so that you can tune them according to your scene. I guess in this case some of the depths are too low. However in such a setting, bidir+mlt will probably be better. Also check that you have put a portal in front of the window.

Jeanphi
jeanphi
Developer
 
Posts: 6573
Joined: Mon Jan 14, 2008 7:21 am

Re: interior test [Distributed Path integrator explained]

Postby Eneen » Mon Dec 06, 2010 7:15 am

I used metropolis becase of gloss floor (it is good move right?).
There are 3 portals, Right and center window and one on the left without window, just hole, to keep lighting nice.
Those are done on default settings, but I've increased all depths to 10 and same thing, no transparency on curtains and brighter wall... I expected some difference but why this wall is so much brighter?
Eneen
 
Posts: 11
Joined: Fri Nov 26, 2010 8:19 am

Re: interior test [Distributed Path integrator explained]

Postby jeanphi » Mon Dec 06, 2010 7:20 am

Hi,

Distributed path started as a hack to meet deadlines when rendering sequences of a motion picture. However the math behind it is currently wrong. It's all fixable, but it will require a heavy rewrite.

Jeanphi
jeanphi
Developer
 
Posts: 6573
Joined: Mon Jan 14, 2008 7:21 am

Re: interior test [Distributed Path integrator explained]

Postby Eneen » Mon Dec 06, 2010 7:49 am

I see, so I will use MLT+BiDir to get physically accurate renderings.
And what about exphoton map? Will it give accurate results too? I mean not physically accurate, but more or less similar to MLT+Bidir.
Eneen
 
Posts: 11
Joined: Fri Nov 26, 2010 8:19 am

PreviousNext

Return to LuxRender User Support

Who is online

Users browsing this forum: No registered users and 1 guest