Video Hardcopy

Michael Frank Deering: Hardware: Video Hardcopy

Company: Benson

Date of Commercial Product: ~1985

Market Need

By the mid nineteen eighties, a number of companies were selling dedicated CAD/CAM color workstations for both mechanical and electrical design. The major players included IBM, Daisy, and Applicon among others. Their frame buffers were getting fairly powerful for being rendered into, but had very slow read-back access. Also, the networking between individual workstations was not yet standardized (e.g. not all Ethernet yet). What printers were available tended to be monochrome black and white, yet almost all the screens were colored wireframe or shaded color. There was a need for a user to be able to press a “print screen” button and get a faithful hardcopy rendition of what was on their screen in a short time.

Opportunity

Benson (owned at that time by Schlumberger) was a maker of traditional color pen plotters, as well as black and white electrostatic printers (via their acquisition of Varian’s electrostatic printer business). While Benson was working on a high end wide body solution (seeĀ Color Electrostatic Plotter), it was not out yet, and a smaller format higher quality print image solution was needed. Benson was evaluating an early color ink jet printer (drum mounted), and decided to look at building a real-time high resolution video grabber as an interface device between a users workstation and the ink jet printer.

Implementation

I had already written several software implementations of a high quality dither algorithm I had developed (vaguely based on the Floyd-Steinberg algorithm), and had also designed a custom chip to implement the algorithm. One of my software implementations was a all hand coded 6800 assembler loop for real-time video conferencing on monochrome Sun workstations. So Benson came to me with the question: could I put a four color version of this algorithm into off the shelf chips (e.g. bit-slices)? After another team had implemented a genlock and pixel digitization circuit (one pixel delivered per scan line, moving over by one each frame) I set out to write an AMD 29116 based program for the color dithering and ink jet printer control. I wrote my own macro assembler for the machine in Franz lisp, but the only was to load the microcode was to burn it into EEPROMs and physically swap them into the board.

Sales

The product worked, produced great looking pictures, and filled a need that there was only a distant competition from Textronix (using an inferior quality printer). So despite the high cost (ink jet printers alone in these days were $20K), a fair number of sales were made (one printer could service several dedicated $60K workstations). Of course over time better frame buffer read-back and network connectivity eventually made the product unnecessary.