2

I would like to be able to use speech intents with known definitions (0-9) in my glass app.

In reference to glass voice triggers, when a user uses the "OK, Glass" menu they're prompted with a list of available apps with voice triggers that are recognizable. IE "OK, Glass Take a Picture"

They don't require network connectivity or don't call "ACTION_RECOGNIZE_SPEECH" on android which requires internet.

How does GDK handle this?

StanleyZheng
  • 4,008
  • 3
  • 21
  • 24

2 Answers2

2

There are a number of voice triggers that are part of the implementation of Google Glass code ... however other voice recognition does require internet access (either wi-fi direct or through your bluetooth connected phone).

ErstwhileIII
  • 4,829
  • 2
  • 23
  • 37
  • If i create a glass app, and define a 'voice_trigger' it saves that trigger as a recognizable section on the OK glass menu options. It maynot be preferred but if my user does not have network connectivity I want them not to be limited without voice input. – StanleyZheng Mar 10 '14 at 20:49
1

First of all there is a way to verify when glass talk to google server and when does not. For traffic analysis you will need to pair your glass with laptop and have Wireshark running for the bluetooth.

My perception is that the decision "recognize on server" VS "recognize right here in glass" does not depend if it's a voice trigger or voice recognintion activity. The decision depends on how difficult the sentence is or how many points of possible double mapping of the sounds to words.

For 0-9 it should be possible to perform simple wave analysis without google speech recognition.

Pavlonator
  • 879
  • 1
  • 9
  • 21