0
$\begingroup$

Rosanswers logo

Hello,

I would like to implement a speech detection which will send sentences to the google API and get the text. This will be permanently listening and doesn't need any vocabulary. so before starting I thought it would be better to ask maybe somebody already did something like that.


Originally posted by Mehdi. on ROS Answers with karma: 3339 on 2014-04-10

Post score: 0


Original comments

Comment by bapclenet on 2014-06-07:
I'm looking for something like you, have you tried the package audio_common? http://wiki.ros.org/audio_common It should catch the sound from the microphones and publish an /audio message. Let us know if it works.

$\endgroup$

2 Answers 2

0
$\begingroup$

Rosanswers logo

A similar thing has been done here https://github.com/Barista-Bot/voice-control in python.

This wasn't on a Nao but our own proprietary robot.

Hope it may be of use.


Originally posted by rmb209 with karma: 80 on 2014-05-27

This answer was ACCEPTED on the original site

Post score: 1

$\endgroup$
0
$\begingroup$

Rosanswers logo

For a simpler library, you may want to check gspeet: it's a small python library that listen (continously or not) on the microphone (using ALSA. It should work on Nao), encode the speech using the speex encoder (well, a google variant of it, actually -- it still requires libspeex), send it to Google speech API, and returns the result.

Integration in a ROS node should be trivial, but the library itself is independant.


Originally posted by severin with karma: 240 on 2014-05-27

This answer was NOT ACCEPTED on the original site

Post score: 1

$\endgroup$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.