Jurassic Park movie & SGI
Mark Wang
Anyone know what model of SGI was used to produce the special effects in
the movie Jurrassic Park?
--
Mark Wang, Student ('95) * Monta Vista High, Cupertino, CA USA
"Attack me if you dare. I will | "The union of the mathematician with
crush you!" - Ken (SF2) | the poet, fervor with measure, poison
Internet: mw...@walrus.mvhs.edu | with correctness, this is surely the
Plodigy: XWPK78B | ideal." - William James
Disclaimer: Monta Vista probably will never share my cool opinions.
Frans Rip (Wageningen Agric.U., Netherlands, 08370-82098)
>Just to satisfy my curiousity...
>
>Anyone know what model of SGI was used to produce the special effects in
>the movie Jurrassic Park?
The credits mentioned are IRIS INDIGO ELANS and 4D/440 VGXT computer
systems.
Shay & Duncan, The making of Jurassic Park, p. 194.
BOXTREE, London UK, 1993.
Frans Rip Wageningen Agricultural University (WAU), the Netherlands
****************************************************************************
Department of Physical Planning&Rural Development (RPV) - Landscape Section
email: <R...@RCL.WAU.NL>
FAX:31 8370 82166 -TEL:31 8370 82098 -Gen.Foulkesweg 13, 6703 BJ, Wageningen
============================================================================
Christopher Bailey
>In article <1994Mar22.0...@newsserver.rrzn.uni-hannover.de>, mw...@walrus.mvhs.edu (Mark Wang) writes:
>>Just to satisfy my curiousity...
>>
>>Anyone know what model of SGI was used to produce the special effects in
>>the movie Jurrassic Park?
>The credits mentioned are IRIS INDIGO ELANS and 4D/440 VGXT computer
>systems.
>Shay & Duncan, The making of Jurassic Park, p. 194.
>BOXTREE, London UK, 1993.
>Frans Rip Wageningen Agricultural University (WAU), the Netherlands
Also, if you read the article on SGI in Wired, they say they used
"everything SGI makes." Also, I would bet they used stuff a lot more
powerful than Elans and such. Crimson and Onyx are more like it (I don't
think the Challenge was out yet, and I know the Extreme/XZ^2 wasn't yet).
I bet they have a Challenge now though.
--
Christopher Bailey Ride fast, take chances!
DejaView Software, Inc. Work: bai...@gmg.com
Boulder, CO Home: bai...@sosi.com
Bruce Caruthers
>Just to satisfy my curiousity...
>
>Anyone know what model of SGI was used to produce the special effects in
>the movie Jurrassic Park?
If memory serves, the credits claimed that they used an SGI 4D/440vgxt
and an SGI Indigo (it didn't say what kind), plus a CM5. I know that
I saw an SMP deskside unit (it had a purple stripe) in the embryo lab
and an Indigo in the command center (or whatever they called it),
which would support the above.
Just my $.02,
-bkc
"Imagination is the one weapon in the war against reality." (Jules de Gaultier)
--
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
Bruce Caruthers "Reality doesn't always live up to ideals ...
bcar...@Oracle.Com but that doesn't mean that they're not worth having."
b...@cs.Princeton.Edu -- blurb for "Romantic Ideals", Analog 2/94
Bruce Caruthers
>
>Anyone know what model of SGI was used to produce the special effects in
>the movie Jurrassic Park?
If memory serves, the credits claimed that they used an SGI 4D/440vgxt
Bruce Caruthers
>
>Anyone know what model of SGI was used to produce the special effects in
>the movie Jurrassic Park?
If memory serves, the credits claimed that they used an SGI 4D/440vgxt
Dave Brookshire
Bruce Caruthers <bcar...@us.oracle.com> wrote:
>>Anyone know what model of SGI was used to produce the special effects in
>>the movie Jurrassic Park?
>
>If memory serves, the credits claimed that they used an SGI 4D/440vgxt
>and an SGI Indigo (it didn't say what kind), plus a CM5. I know that
>I saw an SMP deskside unit (it had a purple stripe) in the embryo lab
>and an Indigo in the command center (or whatever they called it),
>which would support the above.
Also, I remember spotting a Crimson sitting next to Nedry.
Dave
--da...@umbc.edu
**-------------------------------------------------------------------**
David Brookshire
UMBC Imaging Research Center voice (410)455-3373
University of Maryland Baltimore County fax (410)455-1035
**-------------------------------------------------------------------**
Walt Schilling Jr.
Park" by Don Shay and Jody Duncan gives a lot of detail about the making
of the movie itself, and it shows many photos, both of production, and
from the film. A couple are of the computer systems.
Dave Brookshire
(da...@strand.irc.umbc.edu)
wrote: : In article <1994Mar23.0...@oracle.us.oracle.com>,
: Dave
: --da...@umbc.edu
--
----------Walter W. Schilling, Jr.---------------------------------------
| schi...@austin.onu.edu (School Address) |
| dg...@cleveland.freenet.edu (Home Address) |
| "To drive an XT to run software that was never meant to run on an XT."|
John Hawkes
|> In article <1994Mar22.0...@newsserver.rrzn.uni-hannover.de> mw...@walrus.mvhs.edu (Mark Wang) writes:
|> >Just to satisfy my curiousity...
|> >
|> >Anyone know what model of SGI was used to produce the special effects in
|> >the movie Jurrassic Park?
|>
|> If memory serves, the credits claimed that they used an SGI 4D/440vgxt
|> and an SGI Indigo (it didn't say what kind), plus a CM5. I know that
|> I saw an SMP deskside unit (it had a purple stripe) in the embryo lab
|> and an Indigo in the command center (or whatever they called it),
|> which would support the above.
You're talking about two separate issues: (1) the computers that ILM
used to create the effects, and (2) the computers that Spielberg used
in the control room set. For the first question, I don't know enough
to give you an inventory of the SGI equipment that ILM owns (and it
wouldn't be appropriate for anyone outside of ILM to talk about this).
For the second question, SGI loaned equipment to the production company
to populate the control room set. The SGI systems are running real
applications, and the camera is filming real output on the monitors,
not simply replaying canned images. (The monitors were reset to a
scan rate which is compatible with the cameras' frame rate.) It is
my understanding that the CM-5 in the control room was an empty cabinet.
--
John Hawkes haw...@sgi.com
Samuel P. Uselton
It should, of course, be pointed out that in an edition of the SGI
Periodic Table from last July, there is a model labeled "Jurrasic
Classic." It sits between Crimson and Onyx using a single R4400
150mhz processor, and with a VGXT graphics system.
John Peterson
Some folks from ILM spoke at a recent Silicon Valley SIGGRAPH meeting.
When asked about computer equipment, their response was "Everything
SGI makes."
Robert Jones
|> >Just to satisfy my curiousity...
|> >
|> >Anyone know what model of SGI was used to produce the special effects in
|> >the movie Jurrassic Park?
|>
|> If memory serves, the credits claimed that they used an SGI 4D/440vgxt
|> and an SGI Indigo (it didn't say what kind), plus a CM5. I know that
|> I saw an SMP deskside unit (it had a purple stripe) in the embryo lab
|> and an Indigo in the command center (or whatever they called it),
|> which would support the above.
You're talking about two separate issues: (1) the computers that ILM
used to create the effects, and (2) the computers that Spielberg used
[...]
It is
my understanding that the CM-5 in the control room was an empty cabinet.
--
John Hawkes haw...@sgi.com
Yes, we lent them an empty cabinet with an LED panel and a little board
driving them in some random pattern - they used that to mock up a bunch
more cabinets. Normally multiple cabinets are staggered to make a sort of
'lightening bolt' - with a few systems having two bolts side by side with a
little overhead 'bridge' for the network cables. In the movie they had the
cabinets in some sort of semicircle - end on.
The machine should have been placed a little more prominently in my view :^)
The impact on general public awareness of the company from that film was
remarkable
--Rob
Robert Jones jo...@think.com Thinking Machines Corporation
Shixin Zhu
hi, greetings from heidelberg!
i tried to compile ImageMagick 2.3.6 on Aix3.2.5 but I bounced with:
0706-005 ld: Cannot find file: /lib/libMagick.a
Check path name and permissions or
use local problem reporting procedures.
make: 1254-004 The error code from the last command is -1.
indeed I cannot find libMagick.a in my ImageMagick directory. I have gotten the
source from ftp.x.org.
Any ideas what is wrong?
Thanx for help!
shixin
Dan Kinney
It is a fair bet that ILM owns at least one of "Everything SGI makes."
Hey, shouldn't everyone?
-dan
Dan Kinney Telephone: (415)390-2563
Visual Magic Division E-Mail: dki...@sgi.com
Silicon Graphics, Inc.
Movies did not achieve wide popular success until artists
replaced engineers as the primary creators.
- Paul Heckel
Paul Martz
--
-paul "Eno is the one, Eno's the one to take
pma...@dsd.es.com One hundred percent for your stomach's sake...
Evans & Sutherland Bubbly bubbly Eno!"
Daniel Robert Goldman
>What software did ILM use to produce these images? Was it custom?
According to CG World, ILM used SoftImage, Alias, RenderMan, Parallax
Colorburst (now called Matador, or something like that), and "hundreds"
of custom programs.
Gavin Bell
Daniel Robert Goldman <dgol...@Xenon.Stanford.EDU> wrote:
>pma...@dsd.es.com (Paul Martz) writes:
>>What software did ILM use to produce these images? Was it custom?
>Colorburst (now called Matador, or something like that), and "hundreds"
>of custom programs.
The last color plate in "The Inventor Mentor" has a picture of one of
their custom programs in action-- an Inventor program they use to
assign RenderMan shader parameters to surfaces.
(to clarify-- no, Inventor doesn't support RenderMan out of the box--
ILM extended Inventor to do what they needed it to do).
--
--gavin (ga...@sgi.com, (415)390-1024)
Paul Martz
> According to CG World, ILM used SoftImage, Alias, RenderMan, Parallax
> Colorburst (now called Matador, or something like that), and "hundreds"
> of custom programs.
Could not the same software run on machines that don't have hw
acceleration for graphics? I know some kind of million-vecs/sec
machine would be needed to prototype the models and animation
sequences. But doesn't most of the software mentioned do most of its
fancy rendering in software, and then just draw pixels? In this case,
wouldn't a dumb 24bit framebuffer suffice?
Larry Edwards
|> [...]
|> Could not the same software run on machines that don't have hw
|> acceleration for graphics? I know some kind of million-vecs/sec
|> machine would be needed to prototype the models and animation
|> sequences. But doesn't most of the software mentioned do most of its
|> fancy rendering in software, and then just draw pixels? In this case,
|> wouldn't a dumb 24bit framebuffer suffice?
Well SGI's machines are quite fast all around (our Indigo^2 is maybe 3-4 times
as fast as our SUN Sparcstation 2), but yes for the custom rendering effects
the 3D hardware acceleration isn't too useful, although its nice to have for
conceptual design and prototyping as you say.
Larry Edwards
Michael S. Huang
=
= What software did ILM use to produce these images? Was it custom?
Alias' SoftImage as well as Inventor source code were both used heavily
in the production of Jurassic Park CGI.
>>mikey
--
Michael S. Huang
mhu...@calpoly.edu
Paul Martz
> Well SGI's machines are quite fast all around (our Indigo^2 is maybe 3-4 times
> as fast as our SUN Sparcstation 2), but yes for the custom rendering effects
> the 3D hardware acceleration isn't too useful, although its nice to have for
> conceptual design and prototyping as you say.
Agreed. I was seeing so much emphasis put on the *hardware* used to
make Jurassic Park, I was afraid people were getting the impression
they could buy an expensive hw accelrated graphics workstation and see
Jurassic Park images generated in real time before their eyes. I
wanted to make sure people (especially newcomers in the field) knew
the final images were generated without any hw accelerated graphics.
Michael S. Huang
= In an earlier article, pma...@dsd.es.com (Paul Martz) writes:
= =
= = What software did ILM use to produce these images? Was it custom?
= Alias' SoftImage as well as Inventor source code were both used heavily
= in the production of Jurassic Park CGI.
Excuse me while I bite by tounge.
Alias and SoftImage produce two completely different products. Both
companies claim that their applications were used in the creation of CGI
for Jurassic Park.
Incidentially, a source at SGI says that Inventor source code was given
to ILM and around 90% of the CGI was based on their use of Inventor
source.
John J. O'Leary
well, i'm not sure exactly which model was used. however, in the
sep/oct issue of wired(1993) on page 75, there is a picture of
the ROOMFUL of SGI workstations used. the article claims over
100 of the machines(which are labled "power series") were used
by ILM. I don't think my Indigo could keep up!!!
just my 2cents...
Ethan Gregory Henry
>In article <dgoldman....@Xenon.Stanford.EDU>, dgol...@Xenon.Stanford.EDU (Daniel Robert Goldman) writes:
>> According to CG World, ILM used SoftImage, Alias, RenderMan, Parallax
>> Colorburst (now called Matador, or something like that), and "hundreds"
>> of custom programs.
>
>Could not the same software run on machines that don't have hw
>acceleration for graphics? I know some kind of million-vecs/sec
>machine would be needed to prototype the models and animation
>sequences. But doesn't most of the software mentioned do most of its
>fancy rendering in software, and then just draw pixels? In this case,
>wouldn't a dumb 24bit framebuffer suffice?
Well, presumably you have a finite number of machines. You could have
more dumb framebuffers than accelerated machines if you had fixed money,
as opposed to a fixed number of machines... What I think it boils
down to is that for a fixed number of dollars, one accelerated workstation
will run 2 to 3 times as fast as a software driven framebuffer, but
does not cost 2 to 3 times as much. It's a price/performance thing...
also, the movie companies have to pay (somehow) for processor time.
The longer it takes, the more it costs them, so it's a question of
higher fixed costs allowing lower variable costs.
It varies based on what the price on any given technology is...
also, ILM probably had all the hardware on hand already. If you
started up a new project, you might want to use a simpler machine,
like you suggest eg. Star Trek DS9 using Amigas vs SGIs in Jurassic Park.
That is, I think that's the way it all plays out.
Ethan
Adrian Hill
> In an earlier article, pma...@dsd.es.com (Paul Martz) writes:
> = What software did ILM use to produce these images? Was it custom?
> Alias' SoftImage as well as Inventor source code were both used heavily
> in the production of Jurassic Park CGI.
Not to forget Renderman.
--
Adrian Hill Softimage
Tel: +44 71 287 0708 20 Romilly Street
Fax: +44 71 287 0701 London W1V 5TG
Email: te...@softimage.co.uk United Kingdom
Adrian Hill
> Alias and SoftImage produce two completely different products. Both
> companies claim that their applications were used in the creation of CGI
> for Jurassic Park.
Um, I think you'll find that Alias was used for the majority of modelling,
Softimage for the majority of animation, and Renderman for output.
Perhaps someone from kerner would like to verify this.
> Incidentially, a source at SGI says that Inventor source code was given
> to ILM and around 90% of the CGI was based on their use of Inventor
> source.
For fsn maybe. Why re-invent the wheel.
Stephen H. Westin
>
> In article <dgoldman....@Xenon.Stanford.EDU>, dgol...@Xenon.Stanford.EDU (Daniel Robert Goldman) writes:
> > According to CG World, ILM used SoftImage, Alias, RenderMan, Parallax
> > Colorburst (now called Matador, or something like that), and "hundreds"
> > of custom programs.
>
> Could not the same software run on machines that don't have hw
> acceleration for graphics? I know some kind of million-vecs/sec
> machine would be needed to prototype the models and animation
> sequences. But doesn't most of the software mentioned do most of its
> fancy rendering in software, and then just draw pixels? In this case,
> wouldn't a dumb 24bit framebuffer suffice?
> --
>
I think the confusion here is about what is the "hard" part of computer
animation. To the amateur, with severely limited computing resources,
it seems like rendering is both the expensive process and the key
to good animation. Herewith I expound two principles of computer graphics:
1) A good model is 90% of making a good rendering. Accuracy and detail
in geometry, along with judicious texture mapping, will make up for
a lot of rendering shortcuts. To look at it another way, lots of people
know how to make "good" renderings at a certain level; the difference
is made by the quality of the model.
2) Motion is 90% of good animation. Think of Luxo Jr.: a couple of desk
lamps show affection, disappointment, glee, etc., just by the way
they move. Cruder rendering would have made this piece a bit less
effective; cruder motion would have destroyed it.
So folks like Pixar and ILM invest a LOT of time and talent in modeling
and motion editing. To do this well, you need the best interactive display
you can get. SGI does this well; others do, too. SGI has a track record
of staying in the business and staying competitive; this has led lots of
software vendors to support the box.
Once you have the model and the motion, you just want some generic CPU
to crank frames. Not complicated, but possibly expensive. SGI is reasonably
competitive in price/performance here, and it simplifies life a lot to
keep a one-vendor shop. So you use the SGI's for rendering during slack
periods, and maybe buy some of their compute servers so you get the images
rendered before the release date.
So yes, they could have rendered it all on Amigas, or Suns, or IBM 3090s.
The SGI one-vendor solution happens to work for them at the moment.
--
-Stephen H. Westin
wes...@jake.nad.ford.com
The information and opinions in this message are mine, not Ford's.