hi,

i've been mucking about with 3.1 media trying to understand the difference
in results of increasing intervals vs. increasing samples, and on the whole
i've found increasing intervals to produce a better effect (ie. smoother).

reading the docs, i think i understand why, but "think" is very much the
operative word here.

intervals are the actual number of transformations that may be performed on
a beam passing through a media.

samples is the number of samples taken to ensure that the interval effect is
accurate.

therefore, as an analogy, the question is, if you want to create a
small-scale representation of a crowd at a sports-match, do you randomly
pick 50 of the crowd, but don't worry as to whether they are particularily
representative (fifty intervals, one sample per interval) or pick one member
of the crowd, but ensure that they are the most average member of the crowd
(one interval, but many samples to determine the outcome of that interval).

is this right? if so, it explains why increasing intervals seems better than
increasing intervals. fifty non-average members of a crowd looks more like a
crowd than one member of a crowd, however representative that one member may
be.

Post a reply to this message

I agree with your hypothesis, intervals being the key to more rays thus
more refineable media samples. The reverse isn't possible. And also
don't forget that the more lights you have using 'media_interaction' the
more intervals needed to account for the interacting rays, otherwise
they would "dropout" when attempting to be sampled.
I too do not know the true workings but from observation alone this
seems true.


Tom Melly wrote:
> 
> hi,
> 
> i've been mucking about with 3.1 media trying to understand the difference
> in results of increasing intervals vs. increasing samples, and on the whole
> i've found increasing intervals to produce a better effect (ie. smoother).
> 
> reading the docs, i think i understand why, but "think" is very much the
> operative word here.
> 
> intervals are the actual number of transformations that may be performed on
> a beam passing through a media.
> 
> samples is the number of samples taken to ensure that the interval effect is
> accurate.
> 
> therefore, as an analogy, the question is, if you want to create a
> small-scale representation of a crowd at a sports-match, do you randomly
> pick 50 of the crowd, but don't worry as to whether they are particularily
> representative (fifty intervals, one sample per interval) or pick one member
> of the crowd, but ensure that they are the most average member of the crowd
> (one interval, but many samples to determine the outcome of that interval).
> 
> is this right? if so, it explains why increasing intervals seems better than
> increasing ([intervals]?

 replace with "samples").

 fifty non-average members of a crowd looks more like a
> crowd than one member of a crowd, however representative that one member may
> be.

-- 
 omniVERSE: beyond the universe
  http://members.aol.com/inversez/homepage.htm
 mailto:inv### [at] aolcom?Subject=PoV-News

Post a reply to this message