Rose Center, SGI, and Standards
- Mickey correctly identified the need for standards. In the process
of describing the need for such standards, he wrote:
>My understanding is that Spitz, Sky Skan, and Evans & SutherlandFirst off, we don't have an "SGI system" per se. In addition, I'd
>have some compatability. I understand that SGI systems (in the Rose
>Center and soon to be in Denver) are not compatible with the
say the Rose Center has managed a fair degree of compatibility with
other systems so far.
To the first point... We have an SGI Onyx that is used as one source
(among many) feeding an array of seven SeOS projectors with geometric
correction and edge-blending done in real time. The integration work
was not done by SGI, to the best of my knowledge; instead, much of
the effort that went into creating a coherent system has been done
in-house and by the folks at Trimension. Also, the SGI is not used
to run shows. It *is* used for astronomical research, for content
creation, and as a real-time feed running software created by NCSA
and other collaborators.
The fulldome system (software plus hardware) we have in place here is
as unique as the Rose Center. I can't speak to Denver's situation,
but I know that their approach differs from ours not insignificantly.
Secondly -- and more importantly -- the system here is "compatible"
with others' insofar as we could easily take an
appropriately-formatted playback program created for Sky-Skan's or
Spitz's or E&S's system and convert it to run on ours. We haven't
done this explicitly, but we have done effectively the same thing,
using "dome masters" to create a program that runs in our theater.
We have obviously done the reverse, taking portions of "Passport to
the Universe" and "The Search for Life" and adapting them to run on
Sky-Skan's SkyVision system in Albuquerque (see my posting from 20
June, "Fulldome at the American Astronomical Society Meeting").
So to what degree is that "not compatible"? Not significantly, in
the quotidian sense. Unless you want to do flexible, real-time
activities with software based on NCSA's CAVE libraries... Then you
need an SGI or something similar. :)
Anyway, dome masters are the closest thing to a standard we have at
this point. A dome master is simply a circular fisheye-ish view of
the entire dome. I think Spitz, Sky-Skan, and E&S can all deal with
dome masters; I know we can.
>I cannot speak to Minolta, Goto or other systems at this time.Minolta's system is theoretically compatible with the others
mentioned so far -- the MediaGlobe should be able to take dome
masters and create a program that will run on the system. At this
juncture, however, we should discuss theory versus practice!
In mid-June, Minolta kindly set up a MediaGlobe on the sixth floor of
the Rose Center, and we used the opportunity to try creating a few
movies for the MediaGlobe to play. With a brief period of
experimentation, we weren't able to create appropriately-formatted
files to run on the MediaGlobe. More time would have helped, and we
hope to try something out soon. The folks at STScI have created
sequences that run just dandy on the MediaGlobe, so I don't think
this is a serious problem -- just a snag.
In our experiments with SkyVision, we've had a few problems with file
formats for the dome masters: currently, each image requires an
additional conversion step before it can be read by the SkyVision
So, what works fine in theory always requires practice to work out the bugs...
>Or at least how much image rectification must be dome to moveHere are some of the challenges, off the top of my head...
>between the various systems.
+ recommended resolutions
+ compression codecs
+ color preservation
+ gamma tests / black levels
+ frame rate
+ file format for dome masters
I'm sure there are many more! And Mickey's absolutely correct in
saying that we need to start discussing standards with some
Thanks for reading!
Ryan Wyatt, Science Visualizer
Rose Center for Earth & Space
American Museum of Natural History
79th Street & Central Park West
New York, NY 10024