Image Map Image Map
Page 1 of 2 12 LastLast
Results 1 to 10 of 18

Thread: TRS Voice Synthesizer - Which Speech Chip and Link to Manual

  1. #1

    Default TRS Voice Synthesizer - Which Speech Chip and Link to Manual

    What speech chip was being used? The schematics don't disclose that (the details are in the "proprietary section" of the schematics, and the chips in the device are probably covered in epoxy or at least have their labels scratched off - would be nice if a TRS Voice Synthesizer Owner could take a look).

    The VS100 used the SC-01. It may or not be the case that the SC-01 was also being used in the TRS Voice Synthesizer.

    Also, I am looking for the user manual, and the list of phonemes.

  2. #2

    Default

    Quote Originally Posted by LambdaMikel View Post
    What speech chip was being used? The schematics don't disclose that (the details are in the "proprietary section" of the schematics, and the chips in the device are probably covered in epoxy or at least have their labels scratched off - would be nice if a TRS Voice Synthesizer Owner could take a look).

    The VS100 used the SC-01. It may or not be the case that the SC-01 was also being used in the TRS Voice Synthesizer.

    Also, I am looking for the user manual, and the list of phonemes.
    The Radio Shack Voice Synthesizer also used the Votrax SC-01A chip. It interpreted writes to the last 32 bytes of screen memory as letters representing phonemes, and had a small ROM to translate from the printable characters to the binary codes used by the SC-01A. As a result I think some of the SC-01A phonemes (or maybe just pause codes) were not available.

    Unlike the VS100 it had no "ready" bit to say when it could take a new phoneme, so you had to do the timing yourself... but the Voice Synthesizer also had a FIFO for storing phonemes.

    I have a print copy of the user manual, but alas no scan.

  3. #3

    Default

    Excellent, thanks a lot for the info.
    It would be great to know what these letters were... my goal is to build a hardware emulation of both the VS100 and Voice Synthesizer, without the SC-01 chip of course, and it should be faithful enough to at least approximately run the original voice software (Talking Eliza and Demos and such). For that I would need to know though what these special phoneme characters were (I know the SC-01 phonemes already). And if somebody has the software CAS files or disk images, or knows where to find these files, please let me know as well. Thanks!

  4. #4

  5. #5

    Default

    That's the wrong one, Pete... we are talking about the Voice Synthesizer != VS100.
    Thanks anyhow. VS100 and Voice Synthesizer have the same SC-01 voice chip,
    but completely different hardware interface and driver / software side.

  6. #6

    Default

    Quote Originally Posted by Joe Zwerko View Post
    and had a small ROM to translate from the printable characters to the binary codes used by the SC-01A. As a result I think some of the SC-01A phonemes (or maybe just pause codes) were not available.
    This is the info - the mapping from ASCII to SC-01 phonemes - that is needed in order to reimplement it. Not the SC-01 phonemes.

  7. #7

    Default

    Quote Originally Posted by LambdaMikel View Post
    Excellent, thanks a lot for the info.
    It would be great to know what these letters were... my goal is to build a hardware emulation of both the VS100 and Voice Synthesizer, without the SC-01 chip of course, and it should be faithful enough to at least approximately run the original voice software (Talking Eliza and Demos and such). For that I would need to know though what these special phoneme characters were (I know the SC-01 phonemes already). And if somebody has the software CAS files or disk images, or knows where to find these files, please let me know as well. Thanks!
    Maybe 12 or so years ago I built a clone that was compatible with both the VS-100 and the Radio Shack Voice synthesizer. However it still needed the Votrax SC-01A chip, which would require a Great Big Science Project to duplicate.

    One factoid that's not generally known about the Voice Synthesizer: those 32 final characters of screen memory (called "the window") were all treated exactly the same. The address didn't matter. You could print all 32 characters to the same location if you wanted to, but it was easier to print strings up to 32 characters long... they all went into a 32-byte FIFO. You had to "open the window" by printing a certain character (I think it was the question mark) somewhere in the window. This turned on the red LED. When you were through you "closed the window" with another question mark, after which the Voice Synthesizer would stop listening for phonemes. If you listed a BASIC program that happened to scroll a question mark through the window, you'd get insane-sounding gibberish as the lines scrolled by.

    The Radio Shack Voice Synthesizer also gave no control over the inflection bits (unlike the VS100).

    I've attached a photo of the first version of the device... the second version is on an ISA board with a RS Vox-Box clone as well. The great big chip in the middle is the FIFO.

    IMG_2269b.JPG

  8. #8

    Default

    Quote Originally Posted by Joe Zwerko View Post
    I've attached a photo of the first version of the device... the second version is on an ISA board with a RS Vox-Box clone as well. The great big chip in the middle is the FIFO.
    IMG_2269b.JPG
    Very cool! A new re-implementation has to do without the SC-01 though - there are no more SC-01's available these days, or the prices are astronomical.

    If have built "emulations" of the SP0256-AL2 using the SpeakJet as well as the Epson S1V30120 (which is DECtalk on a chip). DECtalk is versatile enough to do everything (and more) than the SC-01 could ever dream of So that is the way to go for a modern re-implementation.

    My previous speech synth projects were for the Amstrad CPC:

    https://github.com/lambdamikel/LambdaSpeak
    https://github.com/lambdamikel/LambdaSpeak3
    https://github.com/lambdamikel/Speak-SID

    I am planning to do something very similar for the TRS-80. I only need some more details for the mappings. The FIFO and everything is readily done with an MCU and 2 PALs.

  9. #9
    Join Date
    Aug 2013
    Location
    Philadelphia,PA area
    Posts
    126

    Default

    I have the Service Manual to the RS Voice Synth. I thought I got it off archive.org

    Tom

  10. #10

    Default

    Quote Originally Posted by hornbetw View Post
    I have the Service Manual to the RS Voice Synth. I thought I got it off archive.org
    Thanks - but the service manual is useless, and I have it.

    What I am asking for is:
    - links to CAS and disk images with software for the device
    - more importantly, the user manual that details the "one letter character to SC-01 phoneme mappings"

    Or, is it the case that the RS Voice Synth used the phoneme codes from the SC-01 phoneme list? In that case, they should not be single letter. For example, the "ay" in "day" is written like this: "ay", and is mapped to phoneme code 21. But I got the impression that they designed their own phoneme "notation" scheme?

    Well, looking at an example demo program for the RS voice synth might clarify this. If the phonemes there match the phoneme codes in the SC-01 data sheet, then it is likely that they have just adopted that (I guess this was before ARPABET and other phoneme nomenclature systems).
    Last edited by LambdaMikel; March 4th, 2020 at 07:57 PM.

Bookmarks

Posting Permissions

  • You may not post new threads
  • You may not post replies
  • You may not post attachments
  • You may not edit your posts
  •