Shop OBEX P1 Docs P2 Docs Learn Events
Speech Recognition — Parallax Forums

Speech Recognition

Cole LoganCole Logan Posts: 196
edited 2009-02-20 01:30 in Propeller 1
I was wondering if anybody has done speech recognition with the prop. Would it be possible to do it all on chip or would it need to have a seperate chip to help with this.

Comments

  • simonlsimonl Posts: 866
    edited 2009-02-19 19:40
    @Cole: I've recently wondered this myself. I've been following this forum for a few years now, but don't remember this ever popping-up before!

    My guess is that a rudimentary speech rec' must be possible - it was done on a 32K BBC Micro many years ago - but you'll probably need to store pre-recorded 'templates' on a SD card or upper memory in the EEPROM.

    If I can figure-out a suitable mic circuit to hook-up to my MCP3802 A/D chip, I might have a play...

    ▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
    Cheers,
    Simon

    www.norfolkhelicopterclub.com

    You'll always have as many take-offs as landings, the trick is to be sure you can take-off again wink.gif
    BTW: I type as I'm thinking, so please don't take any offence at my writing style smile.gif
  • Oldbitcollector (Jeff)Oldbitcollector (Jeff) Posts: 8,091
    edited 2009-02-19 19:45
    I questioned this myself last week with a Project that was listed on the front page,
    but I can't seem to lay my hands on the thread..

    It used a Propeller/Mic/Viewport to view the data from the mic. It's definitely the starting
    point for this project. I just don't have time to pursue it right now. (Too many others cooking)

    OBC

    ▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
    New to the Propeller?

    Check out: Protoboard Introduction , Propeller Cookbook 1.4 & Software Index
    Updates to the Cookbook are now posted to: Propeller.warrantyvoid.us
    Got an SD card connected? - PropDOS
  • LeonLeon Posts: 7,620
    edited 2009-02-19 22:47
    The TRS-80 had a simple speech recognition system that used a number of band-pass filters using op-amps, with the outputs connected to comparators. The comparator outputs were sampled by an input port whilst a PTT button was pressed on a microphone. The system first had to be trained for a particular speaker by repeating each word several times, creating a template for that word. Recognition accuracy was very good, and it could recognise up to 32 words. Nowadays an FFT would be used to determine the speech energy in each frequency band.

    Leon

    ▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
    Amateur radio callsign: G1HSM
    Suzuki SV1000S motorcycle
  • Mike HuseltonMike Huselton Posts: 746
    edited 2009-02-20 01:30
    Speech recognition was covered as recently as last week. Several posts in the last two years.

    I seem to remember Chip Himself chiming in a couple weeks ago.

    Am I in a parallel dimension? Did I have a stroke? What?

    ▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
    JMH

    Post Edited (James Michael Huselton) : 2/20/2009 2:12:05 AM GMT
Sign In or Register to comment.