Skip to content

Listens in the background for speech, and uses Google's Speech-to-Text API to output text.

License

Notifications You must be signed in to change notification settings

MattWis/constant_listener

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

36 Commits
 
 
 
 
 
 
 
 
 
 

Repository files navigation

Constant Listener

This repository was created as a utility for a home automation system. It listens in the background, and uses Speech-to-Text APIs to give the text said through a Queue. The background noise level is calibrated based on the past 30 seconds of data to enable dynamic power thresholding.

Initilization:

Pick your STT engine and follow the instructions there. You pass in a profile argument to background_speech_to_text that is a dictionary containing the relevant credentials.

Google: Map 'key' to a valid Google Developer Key.

Follow "Step 1" here (Also, concerns raised in the article are valid.)

If there is no key specified, a default key will be used, which was gotten from https://gist.github.com/alotaiba/1730160. There's a surprising amount of discussion on that page, some of it even useful.

AT&T: Map 'ATT_APP_ID' to your APP ID, and 'ATT_APP_SECRET' to your APP SECRET You can sign up for these things here.

Sphinx: If you are using sphinx, you need profile to have "words" mapped to a list of possible outputs. There also may be some additional setup.

Usage:

q = Queue.Queue()
background_speech_to_text(q, {}, "google")
while True:
  print q.get()

About

Listens in the background for speech, and uses Google's Speech-to-Text API to output text.

Resources

License

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages