In article <web.41dbd065b331cac5bf5c2ae10@news.povray.org>,
 "Loki" <nomail@nomail> wrote:

> Examining the use of Monte-Carlo methods in raytracing.  I don't exactly
> know what implementation Paris was getting at as I am by no means an expert
> on the inner workings of POV, though I am well acquainted with Monte-Carlo
> methods in physics models.  The fact that Paris knows enough about these
> matters to discuss them at all seems to indicate that he has an interest in
> this field beyond trolling.  Sure, he may be incorrect in his assessment of
> the application of a new algorithm, but why is that insulting to anyone?

I think it was more what he said about what POV does. POV-Ray radiosity 
uses a Monte-Carlo method where it stores pseudo-random samples in a 
quickly-searchable data structure, but from what I saw he seemed to 
claim only path tracing is the "one true Monte-Carlo method". Sort of 
similar to the argument that POV doesn't do "true radiosity" because it 
doesn't use the patch method. It simply isn't true, and appears to be an 
attempt to make POV look worse than it is. In other words, he seems to 
know enough that he should know better. Either he did and was trolling, 
he was pretending to know more than he really did and was trolling, or 
he was just really bad at saying what he really wanted, as well as 
jumping in way over his head as far as his knowledge of POV-Ray went.

For what it's worth, POV *could* get a speed benefit from path tracing, 
especially when you combine things like blurred reflection, radiosity, 
and focal blur. However, the current methods are no less "real". The 
essential difference is that path tracing follows a single possible path 
at a time, rather than branching through all possible paths up to a 
given depth. The result is that the deepest, least important rays are 
the first to get skipped, allowing more time to be spent on the higher 
level rays, and allowing the sampling of antialiasing, focal blur, 
blurred reflections, and radiosity (essentially maximally blurred 
reflection) to be done at the same time. The tradeoff is that things can 
get grainy if you don't use enough samples, even in scenes with just 
unblurred transparency and reflections and no GI or focal blur. 
Basically, heavy antialiasing is required to get a clear render.

-- 
Christopher James Huff <cja### [at] earthlinknet>
http://home.earthlink.net/~cjameshuff/
POV-Ray TAG: <chr### [at] tagpovrayorg>
http://tag.povray.org/

Post a reply to this message