> 360 URLs
WELCOME  to williamson-labs.com     --HOME
> 90 Subjects on CD
The-Color Wheel-Camera
I B M-&-Sarkis Zartarian's-"Com-Club"
   Preface:

The color wheel camera and the system that it was developed for can only be appreciated if one considers the state of the technology at the time (circa 1985).

Consider that video cameras at that time were mostly BW and used vidicon tubes: CCDs were just reaching the industrial level, were expensive, and of not very high quality.

The available color cameras (NTSC, single tube = << 160 TVL) were expensive, and of low resolution.

Flat bed scanners were prohibitively expensive (>> $1K) and were mostly BW; there were a few hand scanners, but of limited utility & quality).

To capture high quality color images into a computer would have required 3 tube studio/industrial cameras costing many thousands of dollars; even then, with few exceptions that camera would have had the limitation of being NTSC resolution (330 TVL = 440 pixels).

Our color wheel camera consisted of an inexpensive vidicon BW camera with a resolution of ~ 550 TVL (TV Lines); adding a rotating color filter controlled by a special frame grabber card in the PC we could capture a 730 pixel (H) x 485 pixel (V) 64-K colors (16 million if we had used a 24 bit capture card).

Compared to the single tube color camera which required 3 pixels (R,G,B [1]) to equal 1 color pixel, the color wheel system was perceptually 5 to 6 times greater.

Also consider: because the color wheel system used only one tube, there was no color registration required as was the case with all 3-tube cameras (a time consuming ritual done several times a day).

To put this camera into today's (1997) prospective: given the COSTS, there are still no cameras of equal or better performance available; however, that gap is closing as I write this!    --gaw 1997

 [1] Color pixel arrangement: R, G, B, R, G, B, R, G, B, ...

I ran across an article in the Summer 1997 issue  of Invention & Technology (p-52) entitled: "The Color War goes to the Moon." It related how CBS's 1953 sequential color TV technology had been used on the Apollo moon landings. And, though Westinghouse had developed the early cameras, RCA delivered the final version to NASA. The irony was that in 1953, the FCC chose RCA's NTSC color TV system to be the U.S. television standard over CBS's sequential color system. 

The color wheel in the article looked very familiar; it looked a lot like the camera I designed and built for Sarkis Zartarian's "Com Club" project while at IBM (1984/85).

Filters
RCA used: 
Cyan, Magenta & Yellow (green x2)
We used: 
Red, Green & Blue
(green x1)

Next generation: 
Clear, Yellow & Cyan 
(green x3)

RCA Color Wheel
 
Our Color Wheel
Since I had designed our frame buffer (using NEC's new serial video ram, 224-kbytes) I was able to incorporate the sequential buffering as well as READ/MODIFY WRITE to use other filter (complementary) colors. The clear, yellow, cyan would have been a GAS!

To implement the color wheel and avoid "reinventing the wheel," we went to NASA for assistance, who obligingly sent us a copy of the Space Shuttle manual (odd numbered pages only) which gave a cursory discussion of the camera. We then tried to contact the, then, camera's prime contractor--RCA, with them stone-walling us. Finally, NASA arranged a conference call with two of the RCA engineers responsible for the design: getting any information was like pulling teeth--mine. It was as hard as getting the people at IBM-Boca to talk to us about PCs. 
I finally had to buy a book written by one of the original designers of the PC to get any information.  "Interfacing to the IBM Personal Computer," 1st Ed., 1984,  Eggebrecht, Howard W. Sams  ISBN: 0-672-22027-X
 

So we Improvised!
The color wheel shape we choose was very much like the 6 phase pattern that I once used in an electrostatic resolver used for non-contact reading of needle position on a remote power meter.   --Huh?
 We were using vidicons--as did RCA, et al. --with their undesirable lag and "wipe-down" readout properties.  CCD cameras were starting to be practical but not in the price range IBM would consider for this type of product. However, the next generation version or the one after that--we planed to use CCDs.
Our competition (Charlie K____) within--IBM--was doing his stuff in NTSC and needed 8 full-sized cards in an expansion unit to do what we were doing with our single card; and we were working in RGB space to do overlay.

He had written a paper on inserting overlay directly into NTSC, using an algorithm that determined the proper subcarrier phase for the overlaid video--mumble, mumble (hardware of course--at that time--we were all working on XT platforms, Ugh!).  It needed a highly accurate xtal timebase--I have no idea why. His excuse for why it looked like SHIT was "...the oscillator's drift, (more) mumble mumble..."

We demoed our stuff and the camera to three site managers at IBM who were there to determine our status--to CAN or not to CAN--our Ass, and who just happen to be good friends of our "NTSC" competitor. 

I will never forget the scene: They walked in, obligatory 'pleasant chit chat,' and all three lined up in front of the SONY KV-1270-Q hi-res RGB monitor. I turned on the light box which held a colorful transparency, the camera spun-up, and finally I turned on the monitor; as the monitor warmed up, I walked behind and to the side of these three guys and waited. As the image came on the screen, there was an audible gasp from all three, and--so Help-me-GOD--at the same time they all rose up on their tip toes: talk about body language! If they had risen any higher they would have been Levitating! 

The image was perfect! 

To prove we weren't "hosin-em," I went over and put in several other transparencies. 

They said absolutely Nothing! 

They sat through the rest of the demo: voice mail; store & forward video mail; video conferencing in the form of live whiteboard annotation; voice reco--that worked without a head set (keyboard commands). It was Outstanding!. 

They thanked us politely for our time, and left.

The project was CANCELED!   --Sarkis took early retirement.

Sarkis M. Zartarian

1931 - 2004
Sarkis Zartarian was a Damn Genius!

He pulled it all together and It Worked!

His performance through the entire project was what TRUE Management is really all about (or should be about).  He Delegated to each person the work they were best at, and most importantly, WANTED to do.  He defined the mission, gave his people the necessary resources, shielded them from outside interference, and got the Hell-outa-the-way! 

His weakness--which doomed him: he was not a "glad-hander," he couldn't or wouldn't "Network" within IBM. Sarkis was great at selling it to the USER, but NOT to IBM management.

History is replete with proof that the best ideas don't necessarily win: it's not a technology question as much as it's a people/political question--Bummer!

Sarkis was a Visionary! 
Back in the early 1980s he saw the future; many of the things that we use in our everyday lives and take for granted, he had working on a desktop PC and PC AT.
His dream was that of early technology being used for Remote Medicine. He wrote and lectured widely an the subject. IBM chose not to listen...

System Elements: (circa 1984/85)
Video Conferencing 
Voice Reco sans headset
Store & Forward Voice Mail 
Store & Forward Video Mail 
Color Wheel Camera: Very High Resolution
     R G B color wheel 
NEC Serial Video RAM Frame Buffer (224-kbytes/IC)  >60FPS
Next Generation plans:
CCD Camera
Color Wheel: Clear, Yellow, Cyan color wheel = 3 x Green, 2 x Red, 2 x Blue
24 bit/32 bit Frame Buffer
Modem speed enhancement: "Shotgunning"

 Copyright 1999 - 2011   Comments or Questions about the site:  Contact  Webmaster

Suggestions are Solicited, P l e a s e !