Voice commands can be added to an application by updating the resource files, but it’s also possible to add them in the source code if required.

Registering voice commands

Voice commands can be registered by sending a list to WearHF using an intent:

  private static final String ACTION_OVERRIDE =
      "com.realwear.wearhf.intent.action.OVERRIDE_COMMANDS";

  private static final String EXTRA_COMMANDS =
      "com.realwear.wearhf.intent.extra.COMMANDS";
  private static final String EXTRA_SOURCE =
      "com.realwear.wearhf.intent.extra.SOURCE_PACKAGE";

  private void sendCommands() {
      final ArrayList<String> voiceCommands = new ArrayList<>();
      voiceCommands.add("Command 1");
      voiceCommands.add("Command 2");

      Intent intent = new Intent(ACTION_OVERRIDE_COMMANDS);
      intent.putExtra(EXTRA_SOURCE_PACKAGE, getPackageName());
      intent.putExtra(EXTRA_COMMANDS, commands);
      sendBroadcast(intent);
  }

Once registered WearHF will listen for the provided commands until:

A voice command is spoken.
The user presses the ACTION button, which will take the system to the home screen.
The current application exits.
A new OVERRIDE_COMMANDS action is preformed.
A RESTORE_COMMANDS action is preformed.

Once one of these events has occurred the application must re-register it’s voice commands.

Reacting to voice command triggers

Applications can listen for when the speech recognizer has detected a spoken voice command by using a broadcast listener:

  private static final String ACTION_SPEECH_EVENT = "com.realwear.wearhf.intent.action.SPEECH_EVENT";
  private static final String EXTRA_COMMAND = "com.realwear.wearhf.intent.extra.COMMAND";
  private static final String EXTRA_ORIG_COMMAND = "com.realwear.wearhf.intent.extra.ORIGINAL_COMMAND";
  private static final String EXTRA_CONFIDENCE = "com.realwear.wearhf.intent.extra.CONFIDENCE";

  private BroadcastReceiver speechBroadcastReceiver = new BroadcastReceiver() {
      @Override
      public void onReceive(Context context, Intent intent) {
          if (intent.getAction().equals(ACTION_SPEECH_EVENT)) {
              String command = intent.getStringExtra(EXTRA_COMMAND);
          }
      }
  };

Restore normal services

When an application has finished with the speech recognizer it should send a message to restore default behaviour, using the RESTORE_COMMANDS intent.

  private static final String ACTION_RESTORE =
      "com.realwear.wearhf.intent.action.RESTORE_COMMANDS";

  private void restoreVoiceCommands() {
    Intent intent = new Intent(ACTION_RESTORE);
    sendBroadcast(intent);
  }

View Full Source Code

API

Actions

Intent	Description
com.realwear.wearhf.intent.action.OVERRIDE_COMMANDS	Used to take control of the speech recogniser
com.realwear.wearhf.intent.action.RESTORE_COMMANDS	Used to reset the speech recogniser to it’s default behaviour
com.realwear.wearhf.intent.action.SPEECH_EVENT	Sent from WearHF to identify spoken voice commands

Extras

Extra Name	Data Type	Description
com.realwear.wearhf.intent.extra.COMMANDS	String Array	Used to pass in a list of voice commands to the speech recogniser
com.realwear.wearhf.intent.extra.SOURCE_PACKAGE	String	Used to pass the application’s package name to the speech recogniser
com.realwear.wearhf.intent.extra.GLOBAL_COMMANDS	boolean	Specifies whether the speech recogniser should also listen for the global commands
com.realwear.wearhf.intent.extra.ACTION_BUTTON_HOME	boolean	Specifies whether the ACTION button should be used as a way to exit the application and go to home screen
com.realwear.wearhf.intent.extra.COMMAND	String	Passed from the speech recognizer when a voice command is spoken. Identifies the command spoken by the user
com.realwear.wearhf.intent.extra.ORIGINAL_COMMAND	String	Passed from the speech recognizer when a voice command is spoken. Identifies the command spoken by the user, but may be stripped of extra whitespace or unwanted characters
com.realwear.wearhf.intent.extra.CONFIDENCE	int	The confidence level reported by the speech recogniser. The higher the confidence, the more closely the spoken words matched the returned phrase (ranges 4000-8000, where higher is better).

Speech recognizer example