As has already been pointed out, in my scene there really was no reason to 
change intervals to 4 from 1. That was my lazy way of increasing the sample 
size based on my understanding of how type 2 media works. My understanding 
of type 3 is not so good though.

Intervals are sections of the ray along which samples are taken from the 
camera (or is it the light? I forget) to the first object hit. Then samples 
are taken within each interval up to the samples value. When I made type 2 
media I just altered the existing code so that those samples were created at 
even intervals within each interval rather than at random, which helped 
smooth out some things. In essence it restored the way some of the old halo 
and atmosphere worked in 3.0 and found it made some scenes look much better.

The total number of samples taken is min samples*intervals in type 2. 
Looking at the docs I can see that more than one interval is not recommended 
for type 3 media and max samples is ignored because it uses an antialiasing 
method.

"Kenneth" <kdw### [at] earthlinknet> wrote in message 
news:web.49403952c8b65ac578dcad930@news.povray.org...
> William Tracy <wtr### [at] calpolyedu> wrote:
>
>> No extinction (too fiddly for me, honestly) and no intervals--I need to
>> re-red the docs on what intervals does.
>>
>
> I kind of gave up on trying to understand what intervals does. :-|
> Be aware that the docs (those that come with the program, anyway) have 
> wrong
> default media values. The correct defaults are:
>
> method 3
> intervals 1
> samples 10
>
> It seems that a fair number of users try fiddling with the intervals 
> value; but
> IMO, it's just easier (and more understandable) to leave intervals at 1 
> and
> crank up the samples until things looks good. I think that's a 
> recommendation
> in the docs anyway (when using method 3.) Others have noted that it's 
> faster,
> too, although I haven't experimented to find out.
>
> KW
>
>

Post a reply to this message