Google Voice Recognition and Watchout

[Lloyd Stewart]

This is just a shot in the dark, but Google Voice Recognition (GVR) has gotten to be quite good.  In fact I can't recall that it has ever gotten anything wrong for me.  So now I wonder if there is any way a "Cookbook" connection could be written between GVR and Watchout, like was written between Kinect and Watchout?

 

Consider: If I can ask Google via GVR where the best hamburger in Linköping, Sweden is and it tells me Stangebro Gatukok, then why can't I use GVR to start a timeline or conditional layer?

 

Ultimately, what I want to do is record responses to oft asked questions, put those responses into Watchout, and then have those responses run based on what someone says or asks, as interpreted by GVR.

 

We tried using Kinect and got it working to a degree.  But it is no where near as accurate in understanding human speech as GVR is.  So that led me to wonder if somehow we could use the GVR engine to initiate responses in Watchout?         

[Mike Fahl]

There seem to be a specification for doing this in a web browser:

 

   https://dvcs.w3.org/hg/speech-api/raw-file/9a0075d25326/speechapi.html

 

There's a simple example of accessing the API here:

 

   https://github.com/Daniel-Hug/speech-input

 

Requires Chrome to run, as this API doesn't seem to be supported in other browsers at this time.

 

So if you're prepared to pipe your user's input through Chrome, for using it as a speech-to-text converter (sending it to Google's servers along the way), this should be quite doable.

 

Mike