As someone who was alive and at least semiconscious during the 1980s, I vividly remember the strange thrill of having a machine talk. For reasons that I think are more emotional than rational, there was a time when the idea of having a talking computer or video game system was a Big Deal, at least in short bursts. Everyone was excited when a Berzerk arcade machine would announce “coin detected in pocket!” or when the computer in WarGames asked if you wanted to play a game. Speech seemed like a big deal, but in hindsight I’m not so sure it was, at least for home consoles. Artificial speech synthesis is a complex thing, something that people have been working on, starting off fully mechanically, for literally centuries. By the 20th century electronic solutions started to become possible, and combined with a developing understanding of the fundamentals of speech — phonemes, allophones and so on — early electrical speech synthesis systems were developed. The most famous of these was likely the Voder, a Bell Labs project that used electronics to synthesize speech. This was not a computerized system — a human was needed to combine the various phonemes manually — but it used electrically generated buzz tones that were modified and “shaped” by other components to create strange speech: In 1961, programmers John Kelly and Carol Lockbaum wrote code to make an IBM 7094 sing the song Daisy Bell, making it the first computer programmed to use synthesized speech. This was the inspiration for that famous scene in the movie 2001: By the 1980s, Texas Instruments pioneered mainstream computer speech with the first integrated circuit — a chip — that talked, the TMS5200. This was at the heart of Speak-and-Spells and those Chrysler talking New Yorkers, and later, in the speech module add-on for the TI 994/A home computer. Soon after that, Mattel Electronics brought out the Intellivoice for its Intellivision console, and then Magnavox introduced The Voice for the Odyssey2 console. These machines represented the second- and third-ranked sellers in the first big boom of cartridge-based home video game systems. It’s worth noting that Atari, the leader with their VCS (later 2600) never bothered to release a speech synthesizer, though it would have been technically possible. In fact, there’s one you can buy today, developed by some very clever hobbyists. Atari’s call to not bother seems to have been the right one; even though these add-on talking boxes were cool, they didn’t sell all that well, really. And Mattel’s insistence on games that worked only with the speech module meant that a number of games just never caught on. The Magnavox approach was different: Speech enhanced games, but wasn’t required. While that could have killed impetus for people to buy the speech module, it did mean that the games wouldn’t be hamstrung by it, either. Also, the Odyssey2's keyboard became an asset here, as a real text-to-speech cartridge was able to be made, which allowed what kids really wanted a computer speech module to do: Say bad words, over and over again. Making a robot voice say “SHIT! FUCK OFF! HEY ASSHOLE” was arguably far more fun than playing some game where the computer tells you your plane’s fuel is low or whatever. I know that’s what my kid likes best. Look, just watch the video and you’ll get even more exciting speech synthesizer facts to make your loved ones wish you’d just shut up about it, forever! Oh, and yes, that is a Changli shirt, and I think I know where you can get one of your own, by the way. They don’t all have paint on them, either.
