Search found 42 matches

by Zelcious
Fri Sep 28, 2012 11:16 am
Forum: General Development
Topic: The light image in BPT
Replies: 11
Views: 6529

Re: The light image in BPT

Are you sure you sample Le(1) by projected pdf? That would be my guess of error otherwise. How is Le defined?
by Zelcious
Wed Sep 19, 2012 6:29 pm
Forum: General Development
Topic: MIS on the GPU
Replies: 41
Views: 21966

Re: MIS on the GPU

Make sure you don't use half floats. Quite easy to overflow them. Is it ok with for example 1.5?
by Zelcious
Thu Sep 13, 2012 4:15 pm
Forum: General Development
Topic: PMC-ERPT
Replies: 1
Views: 2330

PMC-ERPT

Is it the mutated path that goes back into the re-sampling step or the original. If it's the mutation then is pmc-erpt unbiased? Should violate detailed balance right, since the number of mutation is dependent of path intensity. Trivia: I did a little calculation last where it tried to come up with ...
by Zelcious
Sat Aug 25, 2012 10:15 pm
Forum: Links & Papers
Topic: Vertex Merging
Replies: 21
Views: 13964

Re: Vertex Merging

The same idea was also more fully developed in this tech-report:

http://research.nvidia.com/publication/ ... simulation


I can't find this paper on my HD any more and the link is broken on nvidias homepage. Does anyone have a copy?
by Zelcious
Sun Aug 19, 2012 9:15 pm
Forum: General Development
Topic: ERPT - fixed length mutation question
Replies: 6
Views: 3965

Re: ERPT - fixed length mutation question

Yeah I was also disappointed when I reread that part in veach thesis. Thanks for pointing me in the right direction Dietger! But truth is that you very rarely find something revolutionizing. Usually you have a very tiny improvement of a existing method that is compared against X other (often porely ...
by Zelcious
Mon Aug 13, 2012 5:43 am
Forum: Visuals
Topic: Topmod!
Replies: 3
Views: 5461

Re: Topmod!

Solid or hollow?
by Zelcious
Fri Aug 10, 2012 4:10 pm
Forum: General Development
Topic: Metropolis PT vs metropolis BDPT
Replies: 4
Views: 2888

Re: Metropolis PT vs metropolis BDPT

I can see why a keleman style MLT would benefit more by having it ontop of a BDPT since mutations early in the path propagate and explicit connections will probably help a lot but for mutations based on directly in world space there should be less benefit. That is my thinking at least.
by Zelcious
Fri Aug 10, 2012 3:43 pm
Forum: General Development
Topic: ERPT - fixed length mutation question
Replies: 6
Views: 3965

ERPT - fixed length mutation question

For ERPT, why is mutation chains required to be fixed length for unbiased rendering? Is it per sample or a global constraint?
by Zelcious
Tue Jul 24, 2012 9:52 am
Forum: General Development
Topic: Veach thesis - formula question
Replies: 86
Views: 38972

Re: Veach thesis - formula question

This is how MIS works: You have a path generated by one of several sampling strategies. Lets say you generated the path by connecting to a VPL and the other sampling strategy you use in your software is plain Path tracing. In this particular case the connection was really close causing extremly high...
by Zelcious
Mon Jul 23, 2012 8:06 pm
Forum: General Development
Topic: Veach thesis - formula question
Replies: 86
Views: 38972

Re: Veach thesis - formula question

But how MIS can fix the problem ? The problem is due to very small value of ||x-x'|| and G is not part of the path pdf ! G is not related with the way we sample... there is no relation ! What do I miss ? [/url] You are missing the conversation from solid angle to area measure. That contains ||x-x'||.

Go to advanced search