Speech Synthesis (draft)

Update: I happened to publish this when I was intending to save a draft. I’ll trim it down a bit nd keep it here, write a more interesting text another time. No undo, no surrender!

I think the first song I heard a computer sing in was Cylob’s ‘Rewind the Track’ (1998). That was pretty cool. Well, as it turns out, Max Mathews made this 26 years earlier. What the…? In fact, it is easier to make a computer sing than talk. “Talkboxes” were around before computers, but were pretty hard to handle, because of all the tonality of human talk. Voder!

We can distinguish two main types of making machines talk: constructed (consisting of short sampled phonems) and synthesized (the sounds are generated by software). Here’s a little list of low-fi sounding speech.

1962: bicycle for two @ v/a: Music For Mathematics
1963: misc bell stuff @ v/a: first philadelphia computer music festival
1963: misc bell stuff @ some 7″ release?
1980: Stratovox (arcade game, using SN76477) first?
1983: Star Wars (Atari arcade game, using TMS5220-chip)
1983: Discs of Tron (Atari arcade game, using TMS5220-chip)
1985: Indiana Jones (Atari arcade game, using TMS5220-chip)
1985: VironCybernet’s homemade
1990s: Dr.Sbaitso
2010: NES-singing using the (external) FDS-channel here + utube

Demoscene Speech Synthesises

  • Viznut – Who Cares (C64-demo 2004) sampled phonems, ie – run on 6581!
  • Dalezy – Error23 (C64-song 2005)

Tools

  • the Apple speaks, by Bob Bishop (software or tutorial?) here (1979)
  • Commodore Speech Demo (c64 1984)
  • MacInTalk (Mac 1984)
  • SAM (c64)
  • Atari 520ST Speech Synthesizer (here)
  • Say in AmigaOS (1985)
  • Speech Demo 4, among others (c64 1985)
  • Speech Demo Kennedy Approach (c64)
  • Audiosculpture (Amiga/Atari 1999x)
  • Dalek Speech, GUI for SAM (c64, 1994)
  • LSDj (Gameboy 200x)
  • RADWAR Speech Synth (c64, 2007)
  • C64 SPO256 Speech Emulation (c64, 2008)

Misc

  • Speech Plus Inc for misc PCs (1984) youtube
  • Synthetiseur Vocal for C64. French and rather articulate. info + mp3
  • Currah Microspeech (SP0256)

Vocoderoidz

  • Alih – Chase That Feeling (C64-demo 2009). Two channels of singing – the vocoder-approach. info
  • Frantic – Birdburner (C64-thing 2010). Replaying bird speech with triangle voice on/off. Yes, birds talk too, you specieist.

42 Responses to “Speech Synthesis (draft)”

  1. valmetchildren Says:

    (core) also has a similar speech synth (and is also by viznut) for the Atari 2600:
    http://pouet.net/prod.php?which=30236

  2. Viznut Says:

    Hey, thanks for the visibility again (: An adaptation of my speech routine can also be found in the Atari 2600 demo “(core)” by Wamma.

  3. kometbomb Says:

    Since Amiga Say was mentioned, I have to add there was something similar on the Atari ST as well. It’s notable because I think at least U96’s Das Boot uses it (see video). I’m not sure about the year but as an Atarian, I’ll make a completely biased guess it was earlier than the Amiga speech synth.

  4. brytburken Says:

    the book “How to wreck a nice beach” seems to be interesting.

    It’s all about the history of speech synthesis. Check

    http://www.mhpbooks.com/book.php?id=309

    • chipflip Says:

      thanks for the link! vocoders is cheating though, because it needs that rotten human voice input. speech synthesis is full-on anti-humanism!

  5. Dubmood Says:

    afaik the das boot song uses STspeech aka STSpeech520 and STtts.prg. It is from early 1985. I only have version 2.0 of it, once a drunk french told me that the first version had glitches and that there was a french version of it but I never had that confirmed.

    Maybe version 3 will be able to pronounce Zabutom zabutom and not zabjucom. =)

  6. ftc Says:

    Regarding that speech synthesis used by U96.. Someone please tell me what speech synthesis that really is? I mean, it is used all over the place… (The same sounds are used in the speech synthesis in Audiosculpture on Amiga for example.) Was the Atari version the original one perhaps, or were this engine/these sounds used somewhere else first?

    To me, it sounds like the same one as used in Model 500’s “Night Drive (Time space Transmat)”:

    …and that was quite early. Don’t know the exact year though. Might have been while the Atari/Amiga versions of this synthesis were out.

  7. ftc Says:

    By the way.. If Vocoders are not counted, then I am not sure Alih’s hack should be counted either, since he is doing something similar. Using pitch info from a human voice as a carrier (for the ringmod, in his case, if your description of his approach is correct).

  8. linde Says:

    I like everything! http://www.sr.se/sida/artikel.aspx?programid=2098&artikel=3433222 <- interesting listening.

  9. gijs Says:

    this one is also nice:

  10. gijs Says:
  11. TRUE CHIP TILL DEATH • TCTD Links for 2010-03-29 Says:

    […] on: Speech Synthesis (draft) https://chipflip.wordpress.com/2010/03/29/speech-synthesis/ […]

  12. chipflip Says:

    it seems like audiosculpture & st speech sound almost exactly the same. audiosculpture was in fact also an atari program – http://dhs.nu/files_msx_digi.php. i guess that all makes it more likely that it is st speech. just remembered that it appears in a polygon window song aswell on surfing on sine waves. some more examples:
    http://www.youtube.com/watch?v=gfB-sDFftQc, http://www.youtube.com/watch?v=rq4YuWz4gpg,

    @ftc: yeah, technically alih’s thing is a vocoder. but everything that is made with a soundchip can be included in anything on this god damn blog. i also considered your birdburner mega software. but i decided that this blog is not about birds.

  13. µB Says:

    I think my first contact with constructed speech was DR. SBAITSO (DOS/1992), which came with my SB card. http://www.x-entertainment.com/articles/0952/
    http://en.wikipedia.org/wiki/Dr._Sbaitso
    I still like the sound of it, although I’m playing around with Shittalker mostly these days. The AT&T one is pretty good, too:
    http://www2.research.att.com/~ttsweb/tts/demo.php

    Looking forward to the full journey from Voder to the uncanny valley of Vocaloid ;)

  14. Rico Says:

    wonderful use of speech synthesis!

  15. otro Says:

    There is also the Lankhor speech synthesis made by Jean-luc Langlois in 1987 for this adventure game (8&16bit computers):
    http://www.lankhor.net/jeux.php?jeu=15&menu=tele
    He pushed it a bit further in 1990 with Maupiti Island http://www.youtube.com/watch?v=Ab7qTE0ZJ30 =)

  16. Bud Says:

    I used to have a program for Apple ][ that did speech, and pre-programmed into were whole routines for doing prank calls. One I still remember started out something like, “Hello, I am a disabled person using this computer to talk. I would like to order a pizza.” It then went on to order ingredients and slowly degenerated into berating and cursing at the pizza restaurant employee.
    I have no idea what it was called or where I obtained it – it must have been some indie release.

  17. linde Says:

    This might be of interest: http://www.youtube.com/watch?v=fSJQsY3icd8

  18. FTC Says:

    Jucke gave me this link, hehe..

  19. chipflip Says:

    Nova Styrene! Apple II rock!
    [audio src="http://www.archive.org/download/Nova_Styrene-Night_Driver_EP/03_Telnet_Heart_vbr.mp3" /]

    Fortran 5 – Bike (Sid Sings Syd) – tlr @ csdb: On Fortran 5’s Blues (1991) there is a song called “Bike (Sid Sings Syd)” where S.A.M. sings a part of the song “Bike” from Pink Floyds 1967 Release “The Piper at the Gates of Dawn” (written by the late Syd Barret).

    SAM in pop culture @ CSDb

    LFO – Freak

    Neuroflip – Jah64 Love V

    Autobot-1000 – First Contact

    Rififi – Dr Acid Mr House
    LFO – We are Back
    Ultradyne – Radar Blips (From “Futurist”, 2000)
    SIMON (game with freaky voice) http://www.youtube.com/watch?v=fyITF-EWyoc
    VOTRAX (singing) http://www.youtube.com/watch?v=swED5Euoj0k
    MEGAMIX http://www.youtube.com/watch?v=aFQOYBNAMHg
    TI-99/4A (“vocoder”) http://www.youtube.com/watch?v=LN4fYDrQXkc
    RB5X Robot http://www.youtube.com/watch?v=IEwdkCeBbJs

  20. kami68k Says:

    as for recent events, someone made a voice synth that is using the FDS Channel of the Famicom, its called FaMiC (ふぁミク):
    http://taka-p.homeip.net/dtm/tools/index.html

    demo:

    singing starts at about 1:00 :-)

    It is not entirely synthesized since the FDS is using very short Waveforms, but well.. ;-)

    • chipflip Says:

      wow, that’s top notch! especially like the we are the world cover. something similar should be possible with the built-in sample-channel aswell. i think more elaborate wavetable-stuff is something we’ll see more of in the future, even on say C64.

  21. dlz Says:

    some more c64 speech: jch’s orcus from 1990 featured some emulated ringmod-based speech.
    same for dalezy – error23 from 2005.

  22. b-hack Says:

    there can only be one:

    http://www.cybersitevisual.net/download/atari/stspeech.zip

    the full name is Atari 520ST speech synthesizer. don’t ask me when it was released or who released it. i’m quite sure it’s pre-90s though.

    demo here: http://www.freesound.org/packsViewSingle.php?id=2688

    there’s also one for the tt called speak_ez.

  23. 1.000.000 soundchips you never heard about « CHIPFLIP Says:

    […] 2191 were found in keyboards, arcade games, toys and synthesizers. Some chips were found in speech devices, domestic robots, mobile phones and other thingies. Afaik, there is no thorough lists of such […]

  24. goto80 Says:

    MAT-101, Amiga Say, 1999. http://www.youtube.com/watch?v=d9CSnXo3DrI

  25. dhoppo Says:

    Some singing synthesis here: http://www.soundcloud.com/mr_hopkinson

  26. goto80 Says:

    Mouse On Mars – Metrotopy. uses SAM or SAY or sth?

  27. goto80 Says:

    1967, mouth/throat animation thingee too: http://www.youtube.com/watch?v=OvWeTJGrTnA

    • FTC Says:

      There are a few weird machines lying around in the basement at the Centre for Languages and Literature at Lund University from 1950 and earlier, with all sorts of weird attempts at creating voice sounds with mechanical means. Too bad there are no video clips and things like that with these machines. Some of them definitely sounds weird and I bet that it would be possible to create some singing-like stuff with at least some of them.

  28. WORLD WIDE INFORMATION SERVICE Says:

    This Gelbart album has some good stuff, especially Dish Washing (organ remix) http://gelbart.bandcamp.com/album/dish-washing

  29. goto80 Says:

    LFO – We are Back

  30. goto80 Says:

    Blake Dexter – Ex (1991). Atari or audiosculpture, probably.

  31. muckleby Says:

    does anyone have a copy of FaMiC? the authors site is down but i want in!

  32. goto80 Says:

    AVIADOR DRO 4000 – Selector De Frequencias 91 (Instrumental)
    | La Fábrica Magnética 913FM35 | ES | 1991

  33. goto80 Says:

    Quinoline Yellow – LCC Basement (Atari’s Rotten.st?)

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s


%d bloggers like this: