"Cousin Ricky" <ric### [at] yahoocom> wrote:
>   Total Time:   43 hours 51 minutes 29 seconds (157889 seconds)
> CPU time used: kernel 86.41 seconds, user 151368.00 seconds, total 151454.41
> seconds
> Render averaged 5.19 PPS over 786432 pixels

A render using two-pass radiosity.  I also discovered that +a0.0 +am2 +r3 +j
yielded results as good as focal blur for the checkered plane, and almost as
good as focal blur for the micronormal on the sphere's substrate (the "smoky"
part).  (I set jitter=off in my povray.ini some years ago after experiencing
dissatisifaction with the feature, so now if I forget to set j+, anti-aliased
checkered planes look horrible.  I soon changed my mind about jitter, but for
some reason I've resisted setting it back.)

______________________________________________________________

  Total Time:    8 hours 38 minutes 35 seconds (31115 seconds)
CPU time used: kernel 4.48 seconds, user 29950.97 seconds, total 29955.45
seconds
Render averaged 26.25 PPS over 786432 pixels
______________________________________________________________

Add 17 minutes for the first pass.

That's a lot better!  Odd thing is that on my first try, I forgot to turn on the
micronormal, and that run took 25% less time than this one.  I don't know why
there would be such a discrepancy.  The histograms didn't show the micronormal
area taking an inordinate amount of time, and with +a0.0 there shouldn't be
that much difference anyway.

Thanks for all the advice, especially for the 2-pass radiosity.  I'll be making
more use of that in the future.  Thanks especially to Jaime Vives Piqueres for
his response in p.a-u (and for rad-save-load.pov).

Alas, no progress has been made on the sphere itself.  You just can't find good
contractors nowadays...

Post a reply to this message

Attachments:
Download 'unfinished-a.jpg' (284 KB)

Preview of image 'unfinished-a.jpg'