some_yahoo <no### [at] waycom> wrote:
> Sampling method 1 would yield 25 samples, whereas method 2 would be a 
> maximum of 1089 samples.  And no amount of optimizing my media would 
> take that any lower.

  The idea with method 2 is that, while method 1 always sends a fixed amount
of samples (when the antialiasing threshold is reached), method 2 sends more
samples adaptively as needed, thus resulting in a better antialiasing
quality. If you wanted to reach the same antialiasing quality with method 1,
it would have to send a lot more rays than method 2 (because method 1 is not
adaptive).

  Which method is best suited depends on the scene. Sometimes simply using
method 2 with otherwise default antialiasing settings results in a much
higher-quality antialiasing than with method 1, without considerable increase
in rendering time (and, as said, to achieve the same visual quality with
method 1 a lot more samples would be needed, increasing the rendering time
significantly).

-- 
                                                          - Warp

Post a reply to this message