Fancom SAPIUser Guide

Overview

Fancom Sapi is a complete collection of classes that wrap Microsoft Speech API (SAPI) 5.4 when running Fantom on a JVM.

Speech

Making your computer speak couldn't be simpler than:

SpVoice().speak("It's time to kick ass 'n' chew bubble gum!")

A more complete example that initialises proper COM threading, lists available voices, and speaks in the background is:

static Void main(Str[] args) {
  afFancom::ComThread.initSta

  spVoice := afFancomSapi::SpVoice()

  Obj.echo("Available voices:")
  spVoice.getVoices.each {
    Obj.echo(" - ${it->getDescription}")
  }

  name := spVoice.voice.getDescription.split('-')[0]
  spVoice.speak("Hello, I'm $name", SpeechVoiceSpeakFlags.SVSFlagsAsync)

  concurrent::Actor.sleep(3sec)

  afFancom::ComThread.release
}

Speech Recognition

Speech recognition is a bit more involved as you need to initialise an input stream, register some grammar to listen for and set up an event sink to recieve callbacks. Never the less, a complete example is given below:

using gfx
using fwt
using afFancom
using afFancomSapi

class SpeechRecognition {

  static Void main(Str[] args) {
    ComThread.initSta

    spReco   := SpInProcRecognizer()
    recoCtx  := spReco.createRecoContext

    // initialise the input stream / microphone
    // not needed with an SpSharedRecoContext
    category := SpObjectTokenCategory()
    category.setId(SpeechStringConstants.SpeechCategoryAudioIn)
    token := SpObjectToken()
    token.setId(category.default_)
    recoCtx.recognizer.audioInput = token

    // register some commands to listen for
    grammar  := recoCtx.createGrammar
    rule   := grammar.rules.add("awesome", SpeechRuleAttributes.SRATopLevel)
    rule.initialState.addWordTransition(null, "Kick Ass")
    rule.initialState.addWordTransition(null, "Chew Bubblegum")
    grammar.rules.commit
    grammar.cmdSetRuleState("awesome", SpeechRuleState.SGDSActive)

    // register an event sink
    recoCtx.withEvents(SpeechRecognition())

    window := Window {
      it.size = Size(320, 240)
      it.title = "Say Kick Ass!"
    }.open

    ComThread.release
  }

  Void onRecognition(Int streamNumber, Variant streamPosition, SpeechRecognitionType recognitionType, ISpeechRecoResult result) {
    utterance := result.phraseInfo.getText.capitalize
    if (utterance.contains("gum"))
      Obj.echo("Chewing gum.")
    else
      Obj.echo("Hur hur, you said, 'Ass'!!!")
  }
}

See ISpeechRecoContext (Events) for a list of possible callback events.

Release Notes

v1.0.2

• Enums with values were not auto-generated with a Variant surrogate fromVariant() static factory method

v1.0.0

• Initial release