Post by Keith ThompsonPost by Eugene MiyaPost by Charlie GibbsPost by Derek SimmonsIt was not exactly a computer but the other night I was watching the
last Starfighter and noticed a Heathkit HERO robot running about in a
scene.
By the way, the extras on the DVD has a feature recounting the making
of the film with some of the special effects people working with Cray
super computers. In it the show one of the Cray's used to generate the
film and animators working at CALMA workstations.
I heard that halfway through production a Y-MP became available to them
and they redid a lot of scenes because they could now render a lot more
polygons.
Y-MPs only came available in 1988.
LSF was way before that.
Had to have been an X-MP.
I can't remember anymore what Whitney-Deimos had (a 1 or an X).
Mostly run in a scalar mode.
Very few Crays in open motion picture film business.
They can't afford the maintenance.
According to <http://www.imdb.com/title/tt0087597/trivia>, The Last
Starfighter was
The first movie to do all special effects (except makeup and
explosions) on a computer. All shots of spacecraft, space, etc
were generated on a Cray X-MP computer.
--
<http://www.ghoti.net/~kst>
San Diego Supercomputer Center <*>
<http://users.sdsc.edu/~kst>
We must do something. This is something. Therefore, we must do this.
I was there, see "Larry Luther" in programmer's credits.
We ended up with a Cray XMP.
Production was late, scenes were redone only if they were faulty.
We had a lot of vectorization in the renderer, written in Fortran
(Cray's only viable offering).
Hardware:
* Digital Equipment Corporation Vax (11-780) was used as the central hub.
The Cray was used as a graphics rendering peripheral (2 million 36 bit
words).
Ramtek frame buffers were the display devices.
Users communicated with the machines using ASCII terminals.
Film recorders came from III.
The production pipeline was:
* Encoding started with Evans & Sutherland PS-300 vector display machines
which communicated with our VAXs at 64KB/sec and had to be programmed
with data flow techniques (very odd but someone thought theoretically
elegant).
Communication to machines was slow and unreliable.
* Motion choreography wasn't originally planned for!! Ken Dozier and I
saw the hopelessness of that and quickly created a motion choreography
program for the PS-300. This quickly proved the concept to the art
director Ron Cobb who loved seeing and editing the motion before it was
rendered (which took a long time). We eventually got IMI's UNIX based
(System V) machines which communicated over ethernet and had a far more
standard graphics architecture where you created a graphics assembly
program
with a few well understood commands. The Prevue program remained
split between the Vax supporting the user's interface and the IMI vector
display.
* Technical directors were responsible for motion choreography and for
lighting etc. They used Ramtek frame buffers 1280x1024x24 (2 buffers)
which cost about $100,000 for displaying the pictures.