The SuperPaint System (1973-1979)

The SuperPaint frame buffer system originated in late 1972 in the Computer Science Lab at the Xerox Palo Alto Research Center (PARC). I designed most of the hardware using pencil on paper, fabricated it (one printed circuit board and the rest Augat wirewrapped prototype boards) with help from a number of others in early 1973, and coaxed it into displaying its first stable picture in early April, 1973. I wrote most of the software too, but am grateful to Larry Clark, Alvy Ray Smith, Bob Flegal, and Patrick Baudelaire who contributed significantly as well.

SuperPaint would never have seen the success that it did without the many contributions of its principal artist, Damon Rarey. Unfortunately, Damon passed away Dec 15, 2002. A memorial to him appears at www.rarey.com.

SuperPaint Lives! A recreation of the SuperPaint system on a Windows PC, with download.

The April-June 2001 issue of IEEE Annals of the History of Computing included an article on SuperPaint. Please note:
"Copyright 2001 IEEE. Personal use of this material is permitted. However, permission to reprint/republish this material for advertising or promotional purposes or for creating new collective works for resale or redistribution to servers or lists, or to reuse any copyrighted component of this work in other works must be obtained from the IEEE."

An article on SuperPaint appeared in Datamation magazine in 1979, and a scan of it can be found here.

I gave a talk on SuperPaint along with my friend and colleague Alvy Ray Smith in January 2000 at the Computer History Museum at Moffett Field in Mountain View. The slides of this talk are available, as is the audio, courtesy of Dr. Dobbs TechNetCast.

Note: This system is not related to the popular "SuperPaint" software program developed and sold by Aldus some years later. Only the name is the same.


SP_backpanel.jpgSuperPaint backpanel 

Judging from the clipleads, the system still had a few bugs at this point.  The frame buffer memory (sixteen cards) held 2Kbit shift register chips totalling 640 x 486 x 8 bits.  We didn't use the brand-new Intel 1103 1K dynamic RAMs because the factor of 2 in density and lower cost were significant.  Using the runcode hardware and careful timing in the software, this shift register buffer was actually faster on most typical tasks than the Alto with RAM, even though the shift registers could only be accessed when the desired scan line rolled around (worst case an entire frame time).

SP_rack.jpgSuperPaint rack 

At the top is the 8-bit video digitizer (~$12,000 as I recall), then a set of fans, then the SuperPaint frame buffer.  The large space is soon to be occupied with another card cage containing the control cards, color tables and D/As, and the interface to the Nova.  Another set of fans and the power supplies are at the bottom.  Originally, there was to have been a second cage of memory cards providing a second 8-bit frame buffer, but it was deemed too expensive.

itworks3.jpgIt Works! 

This is the very first picture captured in the SuperPaint frame buffer when it first came to life in April, 1973.  The system took a standard video signal as input, digitized it to 8 bits, and could capture or combine it with other data.  To take this picture, I pulled a cliplead off of the backpanel using my knees.  The card says "It works! (sort of)".  At this moment the interface to the Nova 800 CPU had not been debugged or even plugged in to the system yet.  In order to preserve this picture, I had to plug the interface card in and get it working with the power on.  Later, I was able to clean up some of the missing bits with a heuristic program and some fiddling.

wagonwheel1.jpgAnti-aliased wagon wheel  twolines.jpgAnti-aliased lines  twoguys.jpgTwo Guys, by Bob Flegal 

Before any paint program existed for SuperPaint, I wrote a "jaggy removal" program to display smooth straight lines.  Today we would call this anti-aliasing.  The algorithm (written in Nova assembler) simply computed an approximation to the intersection of an imaginary circular gaussian with the line.  I was in a hurry to implement this, because I'd already submitted a paper on it.  One of the two sets of lines above is ink on paper viewed by a video camera, and the other set is coming from the frame buffer.  It's pretty obvious which is which, because the paper wasn't entirely flat.

blackgirl.jpgBlack Girl, by Fritz Fisher 

In the earliest days of SuperPaint, the runcode hardware wasn't yet working, so all paint brushes were one pixel wide (and any height)!  Artist Fritz Fisher had taken a job as a night guard at our building in order to be near the system, and images like this one often greeted us in the mornings.  When Fritz wasn't guarding or painting, he read Gravitation by Misner, Wheeler, and Thorne to pass the time.

Janet.jpgJanet, by Bob Flegal 

In the early days, we explored various types of brushes and painting techniques, including this pointillist style on black.

slot.jpgSLOT animation, by Bill Bowman 

This was the first full animation done on the system.  It was created one frame at a time by artist and designer Bill Bowman using a real-time video disk system (300 frames of 4 analog video channels) attached to the RGB outputs of SuperPaint.  Unfortunately, this disk had a habit of crashing its heads on a regular basis despite the heroic efforts on the part of the manufacturer, and we were lucky to be able to complete this animation of the Scanned Laser Output Terminal (SLOT) -- Gary Starkweather's first laser printer.

SuperPaint menu, by Dick Shoup 

Eventually, the menu evolved to look like this.  Functions included Paint, Move, Copy, Store, Load, Video In, Text, Lines, Gridding, Fill, Shrink 2x, Expand 2x, and various forms of color table animation.  The various brush shapes are displayed at the bottom, and the color sliders at the top.

housegrand.jpgOver Easy graphics, by Damon Rarey 

One of the first broadcast uses of SuperPaint was in the PBS series "Over Easy", produced in the late '70s at KQED in San Francisco.  The show's art director wanted some new and different animation for the series, and heard about something strange going on at Xerox PARC.  They came to visit, and this is how Damon and I first met and became collaborators.   Damon produced dozens of graphics and animations which were used regularly in the show.

SP_at_Ames.jpgSuperPaint on location 

In December of 1978, we had our first real tryout during NASA's Pioneer Venus mission.  We transported the entire system to Ames Research Center during the encounter and set it up near the NASA press room.  Damon produced dozens of colorful graphics and animations illustrating mission progress and on-board science which were fed to NASA-TV via standard NTSC video.  This was a godsend to NASA Public Relations, because the Pioneer spacecraft had no camera. On the chalkboard is a tally of demonstrations we gave to mission scientists, press, and other visitors.

inserprint.jpgPioneer Venus orbit insertion, by Damon Rarey 

This color table reveal animation showed insertion of the Pioneer Venus spacecraft into orbit around Venus during the planetary encounter in 1978.  This animation was used on the live NASA-TV feed originating from Ames during the mission.  Unfortunately, the director cut to it just at the moment the scientists at Mission Control were cheering the confirmation of orbit, and some news organizations complained mightily.

solarwind2.jpgSolar Wind, by Damon Rarey 

This color table cycle animation showed the flow of particles from the Sun around Venus, as the Pioneer Venus probes were expected to detect and measure. The NASA scientists and especially the public relations people were overjoyed to have a way to explain the otherwise dull science of the Pioneer mission to the public. A year later, in 1979, we were invited to bring SuperPaint back again to play a similar role in the Pioneer Saturn mission. Between the two missions, our animations were seen on nearly every major television network around the world. On one occasion, Damon did a live interview and SuperPaint demonstration on Italian television from Ames.

greenguy.jpgGreen Guy, by Damon Rarey 

Though it survived the plunge through the atmosphere of Venus, one of the probes mysteriously stopped transmitting just after landing on the hot Venusian surface.  Damon and I thought this is perhaps what happened to it.  The hot gasses in the sky flowed by color table cycling.
 
 
 

Entire site © Copyright 2000-2014, Richard Shoup (rgshoup at rgshoup.com), all rights reserved.