... Patrick Jost

By Sean Kelly

DP: How did you come about working with the Intellivision, and what role did you play in its software/hardware development?

Patrick Jost: In 1981, I'd been working for Pacific Telephone for about a year and a half. This was my first real job after leaving graduate school. I'd messed around with the music industry, done a little "international consulting", some of the typical things one does when one does not know what to do. Anyway, Pacific Telephone was fun. I was working with electronic switching, international testing (I got to call Lybia once), programming custom services, various things. They had lots of Unix machines to play with, so it was also a sort of immersion course in Unix computing.

I started to get bored. I'd gone to most of the schools, I'd worked on various interesting projects. I was spending a lot of time and money at Opamp Technical Books in Hollywood (still in business, still a great place), and I was beginning to want to do something more -- well -- interesting. Mattel was running huge ads in the paper. At the time, my main concern was the commute. I lived about 10 minutes from the Pacific Telephone facility in Hollywood, Hawthorne seemed far away. After a while, I got over this concern, and went to one of Mattel's job fairs (back in those days, LOTS of companies were having them). I got along with the people right away. Intellivision was an established product, they wanted to do more with it. They wanted to add voice synthesis. They were looking for someone with a linguistics background (that's what I majored in!) and who understood computers (thank you, Pacific Telephone).

This was Saturday. They asked me to come back Monday. I talked with some more people, and filled out the application. They were talking good money, and it sure sounded interesting. By the time I got back to Hollywood, I had a message on my machine, they offered me the job that day. I gave notice at Pacific Telephone, gave myself about a week off, and started to work.

My first day was Monday... and already things were getting interesting. I had to fly to New York the next day to help with the speech for the first game. This game grew up to be Space Spartans, but, at the time, all anyone knew was that it was a space game of some sort. It was supposed to be a short trip; it turned out to be several weeks. I recall that due to the short notice I got to fly first class, and sat right behind Count Basie and a member of his band.

DP: How was the speech was made for these games?

Patrick Jost: Along with the game idea, a script was written. I transcribed the script (into phonetic transcription) and made sure there were no critical words that would be "transformed" too badly by the speech synthesis process.

After the script was written, auditions were held. I used my contacts in the music industry to find good agents and a good recording studio. We looked for good voices, good acting, and actors that could work with some of the odd requirements of speech synthesis (not too many 'hissing' ess [s] sounds, no loud popping p's and so on). The only (voice actor) name I feel comfortable mentioning is Phil Proctor of Firesign Theatre fame. Much credit goes to Fred Jones of Fred Jones Recording for his assistance with the recordings - excessive background noise, printthrough, and 'pops' cause problems). I finally developed a pretty good ear for which voices would synthesize well.

After the recording, the voices were sampled. We used a Hewlett-Packard 1000 series machine with the ILS signal processing package STI in Santa Barbara for editing (hi Steve). This editing consisted of cut and paste of sounds, digital filtering of samples, etc.and a large amount of custom software. The sampled speech was fed to the synthesis software for the Intellivoice speech synthesizer, the General Instrument SP-256. Synthesized speech could be generated quickly. Selected recordings were sampled, and played back for comparison with the script and transcription. Finished product synthesized speech was also transcribed and compared to the master transcriptions and game script. The problem is that automatically generated speech took up a lot of space (that could be used for more speech or game code). This was a big problem! The other problem is that the automatic speech synthesis didn't always sound that good... some of it was actually pretty bad.

The solution to both problems was manual editing of the original waveform before the speech was synthesized. This was done with a good, but somewhat primitive editor. Segments to be used for synthesis could be marked, and speech could be deleted. The resulting files could be submitted for synthesis; the results were usually speech that took up less space that the automatic speech and that sounded good.

For the first six months or so, I did almost everything - work on scripts, transcriptions, auditions, recording sessions, speech editing - at General Instrument's site in Hicksville, NY (GI made the speech synthesizer). I did almost every millisecond of speech you hear in Space Spartans and B-17 Bomber myself. The other games were done at the Mattel speech lab.  By the time Bomb Squad came along, Mattel wanted to be more organized. A formal speech group was set up; I trained the editors, largely on what you hear in Bomb Squad. The last speech game was Tron Solar Sailor, I did not have much to do with that one. I went on to work on some other things for Mattel: consumer musical productions, and advanced technologies for the games, specifically a rapid prototyping environment. For a while Mattel was also very interested in entering the European marketplace, so I worked on Spanish, German, French, and Italian versions of Space Spartans. That ROM is out there somewhere.

DP: I've heard that Mattel had a "laid back" environment:  it was a fun place to work. Would you say the same?  

Patrick Jost: Fun place to work? Sure, especially if you liked video games. I didn't, and still don't. But remember, this was during the time when it seemed like there was a Pac-Man machine everywhere. Mattel had some very good people. Most of us were about the same age... late 20s, early 30s, I guess. Many common interests apart from the games. I played Geddy Lee-style bass in an informal group called the Redi Spuds (named after a sign on a nearby building) that played sort of a new wave rock; yes, a total mismatch of styles, but fun. I shudder to think of what it would sound like now, with my more Percy Jones influenced style.

You could always find someone interesting to talk to, even though I don't think they planned it, there was quite a lot of synergy. In speech, we were doing things with audio on minicomputers that are commonplace now in this age of samplers... but we solved the problems years ago.

Laid back? Well, the games programmers didn't work on much of a fixed schedule. I was interested in seeing what could be done with natural language processing technology. I should also say that I'm probably NOT a very laid back type of person! I was never really all that happy in California, and my lack of laid back inclinations may explain why I'm one of the few people I know of who moved from Los Angeles to Washington, DC.

DP: Would you know of any unfinished hardware or software that Mattel may have been working on (besides the previously mentioned foreign ROM)? Video game collectors just love this kind of thing. :-)

Patrick Jost: Unfinished games... there were probably lots and lots of them, things came crashing down pretty fast.  ROMs?  I don't know, probably not many of them had been made into ROMs yet. There was a thing called "Decade" which was a 68000 based system that could have been Macintosh like, had they completed it.  There were prototype wireless remote controls  for Intellivision.  There were plans for all sorts of interfaces... Apple II, IBM PC, and so on. You may have seen the Synsonics drums, four touch pads and some buttons with some rudimentary programming/memory capability.  There were also a Synsonics guitar, with "strum bars" for your right/picking hand and a neck full of switches for your left/fretting hand.  I don't think this ever saw production, but I've seen things like it in the COMB and DAMARK catalogues.

DP: Thanks for the interview, Patrick. I appreciated it.

Patrick Jost: No problem.

 Return to Digital Press Home