Shop OBEX P1 Docs P2 Docs Learn Events
FINAL BUY OPPORTUNITY: SPO-512 Text to Speech Chip — Parallax Forums

FINAL BUY OPPORTUNITY: SPO-512 Text to Speech Chip

ercoerco Posts: 20,256
edited 2014-02-12 11:25 in General Discussion
Anybody try this out yet? Is it totally new, or just new to Speechchips? I just emailed Ken (@speechchips) and waiting to hear back from him.

http://www.speechchips.com/shop/item.aspx?itemid=22
«13

Comments

  • LeonLeon Posts: 7,620
    edited 2012-06-05 14:46
    That looks like a dsPIC or PIC32 with speech synthesis software.
  • ercoerco Posts: 20,256
    edited 2012-06-05 16:26
    Reply from Speechchips.com:

    Hi,

    I'll post samples and the datasheet tomorrow.
  • kwinnkwinn Posts: 8,697
    edited 2012-06-05 19:56
    IIRC this is a golden oldie though I can't recall any details. I think I even bought one when they first came out but never had time to do anything with it.
  • Phil Pilgrim (PhiPi)Phil Pilgrim (PhiPi) Posts: 23,514
    edited 2012-06-05 20:06
    Kwinn,

    This appears to be a brand-new chip. You might be thinking of the SPO256 that Radio Shack sold eons ago:

    -Phil
  • ercoerco Posts: 20,256
    edited 2012-06-05 20:22
    Righto, PhiPi! More from Ken:

    Like I said, I'll post samples tommorow but please note that this chip is called the RoboVoice. It sounds very similar to a speakjet or soundgin. It's very robotic. The big feature is that it takes English text and not phoneme codes.
  • mindrobotsmindrobots Posts: 6,506
    edited 2012-06-05 20:51
    Folks think their Twitter printer hacks are all the rage....how about a Twitter Talker!!

    "OMG! @justinbieberdreams #lame you retweet like my mom! lol #trashtalk"
  • kwinnkwinn Posts: 8,697
    edited 2012-06-05 22:57
    Phil Pilgrim (PhiPi) wrote: »
    Kwinn,

    This appears to be a brand-new chip. You might be thinking of the SPO256 that Radio Shack sold eons ago:

    You are absolutely right Phil. Thanks for the reminder.

    -Phil
  • ZetsuZetsu Posts: 186
    edited 2012-06-06 06:17
    I have speech to text on my phone. If it had text to speech I could get away with never texting someone again... I would just convert all text messages to audio conversations
  • ercoerco Posts: 20,256
    edited 2012-06-06 11:32
    It's not going to replace Siri or Microsoft Sam, but it's an option for simple speech. Info from Ken attached here. I have sample audio files but the forum software won't let me post it...
    pdf
    pdf
    SP0512_Proto.pdf
    30K
    pdf
    pdf
    SP0-512-Datasheet.pdf
    130K
  • LeonLeon Posts: 7,620
    edited 2012-06-06 11:56
    It does use a dsPIC, as I suggested. It's the dsPIC33FJ128GP802 or something similar, as that includes the 16-bit DAC mentioned in the data sheet.

    I don't think much of the sample circuit, it doesn't include any decoupling capacitors.
  • Dave HeinDave Hein Posts: 6,347
    edited 2012-06-06 12:01
    Zetsu wrote: »
    I have speech to text on my phone. If it had text to speech I could get away with never texting someone again... I would just convert all text messages to audio conversations
    Encoded and decoded speech has been around for a long time. It's called voice-mail. Even before that we used these archaic devices called answering machines that used magnetic tape. And even before that there was a time when you could talk into a phone and the person at the other end could hear your voice in realtime.
  • ercoerco Posts: 20,256
    edited 2012-06-06 12:11
    Two test samples of audio generated by new TTS chip SP0-512. First is raw text of countdown, 10, 9, 8,7... second is text with some control codes mixed in.


  • RobotWorkshopRobotWorkshop Posts: 2,307
    edited 2012-06-06 12:21
    I've heard this chip was in the works but didn't know it was for sale yet. From the initial datasheet I don't see any data about these topics (which are important to me)

    - What kind of handshaking exists with the new chip, (Ideally dedicated lines using spare I/O as well as status sent back serially to the host) That was people can use either one.

    - What is the size of the input buffer? (Don't want to overflow it!)

    - Is is possible to generate sounds similar to the R2 and dial tones that are in some of the previous speech chips? Those are great for robot projects.

    Robert
  • LeonLeon Posts: 7,620
    edited 2012-06-06 12:27
    Intelligibility is very poor, the original SP0256 was much better.

    We developed a speech synthesis system at Racal Communications for helicopter pilots using an OKI chip set, it generated very high-quality speech. The chips weren't all that expensive, but the development system cost a lot of money. The distributor loaned one to us for a few weeks.
  • RobotWorkshopRobotWorkshop Posts: 2,307
    edited 2012-06-06 12:32
    Leon wrote: »
    Intelligibility is very poor, the original SP0256 was much better.

    It may be too early to tell with the limited demo. I suspect that using the control codes or using the phonemes make make a big difference. It would be nice to try one out and see how it sounds myself with a little tweaking. May make a good replacement for the SC-01 chip.

    Robert
  • LeonLeon Posts: 7,620
    edited 2012-06-06 12:41
    I remember playing with the Votrax SC-01 chip, as well. I interfaced one to my TRS-80 Model I.

    I've been thinking of trying speech synthesis with that dsPIC chip myself. I've designed a little development system for it, which can take different daughter boards. It could also be used for speech recognition.
  • ercoerco Posts: 20,256
    edited 2012-06-06 12:49
    I'll def get one of these to try. Lord knows I've been asking for a simple one-chip solution for a while!

    http://forums.parallax.com/showthread.php?126586-Text-to-Speech
  • blittledblittled Posts: 681
    edited 2012-06-06 13:05
    Time to blow the dust off my old SPO256 and get it working again :). I didn't buy the special crystal but a color burst crystal works pretty good.
  • hover1hover1 Posts: 1,929
    edited 2012-06-06 13:37
    Like these? Unearthed a month ago. Still trying to see what I can working from stuff that was in the Tennessee flood.

    I used to have the 22/44 board hooked up to my CoCo2. My kids where 5-7 and that the time, and the used to spend hours typing words into the CoCo to see what the board spit out. Hilarious at times.

    DSC00578.JPG
    DSC00579.JPG
    DSC00580.JPG


    Another for erco:

    DSC00581.JPG


    Looking for one of these? :)

    Jim
    DSC00578.JPG
    640 x 480 - 150K
    DSC00579.JPG
    640 x 480 - 148K
    DSC00580.JPG
    640 x 480 - 147K
    DSC00581.JPG
    640 x 480 - 144K
  • ercoerco Posts: 20,256
    edited 2012-06-06 14:23
    Nice collection, Jim! I have much the same collection as you. My vintage Archer, Micronta, Lafayette and Allied treasures will either make me rich or find their way into the perfect future application.

    BTW, I stocked up on LM3909s from http://www.ebay.com/itm/IC-LM3909N-NSC-DIP-8-/250881986100?pt=LH_DefaultDomain_0&hash=item3a69bb4e34 TTYTT, I stashed them away without testing them. Someone warned me they may be fake, like 555 timers repainted 3909!
  • softconsoftcon Posts: 217
    edited 2012-06-06 14:26
    Wow. I was thinking of purchasing a couple of these for some of my projects, but man, even for me, these chips are horrible. The propeller is easier to understand.
    I'll give these a pass.
    Thanks for the demo though.
  • Phil Pilgrim (PhiPi)Phil Pilgrim (PhiPi) Posts: 23,514
    edited 2012-06-06 14:33
    One thing I've learned from my own attempts at speech synthesis is that you never really know how bad it sounds until someone else who hears it for the first time tells you. :)

    -Phil
  • ercoerco Posts: 20,256
    edited 2012-06-06 15:06
    I suspect that there's a "learning curve" to understanding these things. Converting text to speech is very complicated. Synthesizing speech is very complicated. So converting text directly to synthesized speech is very very VERY complicated. There are a number of control codes in the PDF that can improve the inteligibility (see how it pronounces THAT word...) if you want to key them in.

    So agreed, they're not ready for prime time, but for someone who wants to make your project talk (for them, if no one else) it's the most direct and cheapest option. After I get a sample I'll make some videos.
  • RS_JimRS_Jim Posts: 1,765
    edited 2012-06-07 05:58
    I think that Joe Grand is working on a speech synth for Parallax!
    Jim
  • GordonMcCombGordonMcComb Posts: 3,366
    edited 2012-06-07 12:20
    RS_Jim wrote: »
    I think that Joe Grand is working on a speech synth for Parallax!

    Yes, it's based on the EMIC chip, and is an updated version of one Joe did for Parallax a couple of years ago. I imagine it'll cost more than the $25 or so for Kenneth's new chip.

    -- Gordon
  • ercoerco Posts: 20,256
    edited 2012-06-07 13:15
    Word on the street from Huggy Bear: Ken may be having a kickoff sale soon. I wouldn't buy just yet. ;)
  • GordonMcCombGordonMcComb Posts: 3,366
    edited 2012-06-07 14:00
    This guy is dealing in speech chips now?

    huggybear.jpg
    huggybear.jpg
    388 x 264 - 27K
  • ercoerco Posts: 20,256
    edited 2012-06-07 15:03
    Yo Gordo, you my homie, Dawg!
  • ercoerco Posts: 20,256
    edited 2012-06-27 13:40
    erco wrote: »
    Word on the street from Huggy Bear: Ken may be having a kickoff sale soon. I wouldn't buy just yet. ;)

    Ken's sale is on, $21.99. http://www.speechchips.com/shop/item.aspx?itemid=22

    I got one for evaluation. You ain't heard the last of me!
  • ercoerco Posts: 20,256
    edited 2012-06-27 16:54
    Another price drop today. Only $15.99! Just TRY to say no at that price! Great summer project!

    And I want a buck from everyone who orders one, same deal as the Walmart eReader... :)
Sign In or Register to comment.