I did a lot of math while we were struggling with the dome tilt question ourselves a couple years back. The end result was that unless you committed to aMessage 1 of 15 , May 22, 2006View SourceI did a lot of math while we were struggling with the dome tilt question
ourselves a couple years back. The end result was that unless you
committed to a serious tilt (like 30 deg.), very little was gained in
the observers field of view. Once you do commit to the steeper tilt,
you are committing to 'filling' that space a lot when you aren't using
it and using more visuals that require you to move 'through' them.
Once you commit to a shallower tilt, little is gained by going from 5
degrees to 20 degrees(ish). More is gained by trying to get the
audience closer to the dome and filling the field of view.
We fought for, and got, a 50 foot dome with a 9.5 degree tilt. With a
little creative engineering our architects brought the dome right down
on our heads.
With the lights out, we can almost make the tilt disappear for our skies
shows, and the front of the dome (at 5.5 ft.) is low enough we can
'fake' a good tilt if we want a sense of forward motion.
We couldn't be happier with the results.
Science Center of Iowa
401 W. ML King.
Des Moines, Iowa 50309
... It all depends on what you are doing... if you are pushing the look of the content there ends up being issues you don t encounter with less sophisticatedMessage 1 of 15 , May 22, 2006View Source
> > Sorry but there is nothing "trivial" here, especially as I wait forIt all depends on what you are doing... if you are pushing the look
> > our latest renderings to finish... chugging along at 4-5 hours per
> > frame (only 2200x2200 too).
> While this does depend on the software you are using ... a "good"
> rendering engine should be able to render cubic maps much faster
> than fisheye, this is because it can cull lots of geometry (if it's a
> space based render) and perform similar optimisiations if it's an
> image based rendered (eg: raytracer). Such optmisiations also apply
> to memory requirements, for example I can generally render scientific
> visualisation style content using much less memory using cubic map
> views vs hemispherical views.
of the content there ends up being issues you don't encounter with
less sophisticated rendering requirements. At that higher
expectation level, even a "great" render farm with lots of fast
processors and lots of ram still ends up taking lots of time...
"lots" being the key word.
H o m e R u n P i c t u r e s
President & Creative Director
100 First Avenue - Suite 450
Pittsburgh, PA 15222