25

The first time I ever played with software speech synthesis on a microcomputer (not hardware synthesis, like in TI's Speak & Spell) was around 1983, using S.A.M for the Commodore 64.

A year later, I remember hearing the speech synthesis from the Macintosh unveiling, and being struck by the notion that it sounded the same as S.A.M on my C64. I suppose I thought that the SID chip and whatever the Macintosh used for sound would lead to much different results. Two years later, I got an Amiga, which came with built-in speech synthesis just like the Macintosh. Same. Voice. Again.

But the biggest surprise was the first time I heard Stephen Hawking speak. This just didn't seem right, since Professor Hawking should obviously warrant something far better than a "C64 under-the-bonnet" for his translator.

Additionally, the modern (2013) SpeakJet synthesizer chip also seem to have the same voice, while certainly being far removed from the 1980s hardware above.

Why do they all have that same strange voice?

9
  • 3
    They don't! retro-kit.co.uk/page.cfm/content/…
    – Chenmunka
    Commented Mar 9, 2018 at 8:56
  • 1
    The Wired article "How Intel Gave Stephen Hawking a Voice" claims that Stephen Hawking's first speech synthesiser was based on Speech Plus for the Apple II, so it's no surprise he sounds like an 8 bit speech synthesiser! Remember that this was built in 1985 when natural-sounding speech synthesis was still decades away.
    – pndc
    Commented Mar 9, 2018 at 9:17
  • 16
    Because Stephen Hawking has a secret deal with all the speech synth vendors, moonlighting as their voice coach. (But seriously, he has been offered - and declined - upgrades, because he's had that voice for well over 30 years now). Commented Mar 9, 2018 at 10:55
  • 2
    Because it is cool to make your computer sound WOPR in the movie War Games. Anybody remember commands like this? Load "Zork",8,1 Ah, those were the days. Commented Mar 10, 2018 at 1:01
  • 2
    @DavidEisenbeisz +1000 if I could. Forget the Turing test...all computers should have to pass the WOPR test. If you can't talk like WOPR...you're garbage.
    – cbmeeks
    Commented Mar 12, 2018 at 19:00

1 Answer 1

41

The ones you list are all S.A.M, from what is now SoftVoice Inc. It was developed first for the Apple II, Lisa, Atari 8-bit machines and the Commodore 64; based on awareness of those versions, the company was contracted directly to supply the bundled speech generators for the Macintosh and Amiga.

So they all sound the same because they are ports of the same software from the same company.

This shared heritage was also confirmed by former Apple & Amiga engineer @hotpaw2 who comments in a related question: "While working at Apple, I purchased a SAM for my Apple II, and demonstrated it to the Mac team, as well as taking it with me to Hi-Toro/Amiga. Both those teams subsequently contracted with the SAM developers to do a 68000 port."

4
  • 8
    Hawking's voice is a DECTalk unit, unrelated to SoftVoice. They sound similar because of the relatively low sample rate and limited palate modelling available to 1980s-era computers
    – scruss
    Commented Mar 10, 2018 at 3:29
  • 1
    I'm surprised the tech failed to show much improvement going from a 1MHz 6502 with a PSG to a 8 MHz 68K with high-speed DAC. @scruss seems to imply there were actual hardware technical limits to be overcome.
    – Brian H
    Commented Mar 10, 2018 at 17:00
  • 3
    This is a reasonable answer with a lot of votes, but doesn't address Prof. Hawking or SpeakJet or any other unifying technological considerations, besides someone probably porting off of S.A.M
    – Brian H
    Commented Mar 12, 2018 at 3:19
  • 10
    After he received his first voice he didn't want to change it.
    – Zak
    Commented Mar 14, 2018 at 13:33

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .