Tuesday, 14 May 2013


SPEECH RECOGNITION FOR THE KINECT: Kinect speech media controller

First things first, sorry this has been a long time coming. I have had a lot of emails asking me for a working demo of how I did the speech recognition for Kin-educate so I thought I would do one better and make a little demo application for people to download and build on themselves. 

The Kinect speech media controller turns your windows media player into a voice controlled media player. Forget remote controls. Forget waving your arms and hands around to control your media player. Control it purely by voice.


Not only that, but to make it a bit more interesting, you get to choose the words for the commands. While this is fun, I found this to be necessary. Kinect speech works great, but it has a few problems with some strong accents. 

When working on the voice controlled maths game for Kin-educate the Kinect refused to recognise the word 'four' in a British accent. After trying to spell it different ways (fore, for) and having no luck I gave up. I just made sure the answers could never be four. Problem solved. But not really.

The real solution? We all speak differently and use different phrases, so why not choose what words and phrases you use to control the Kinect? Give your Kinect a unique name, and when you call it, the Kinect will activate and start listening for commands. 

Try different phrases and words to see which work well for you, and have a bit of fun with it too. This way the Kinect will be more responsive to your particular way of speaking. You can change the volume, play / pause, rewind / fast-forward  open and close the media player and browse for files all through voice control.

This is a really simple demo that could be expanded in to much more. With a quick change to the code this can be used to control any windows based program. A whole range of other voice commands could be added. 


After a few seconds of silence the Kinect will go into standby mode. It will stay like this until you call it by its name. When you are not using it call the 'hide' command to minimise the app to the task bar. 

Check out this video of the app in action. As you can see it works pretty good. You could have a lot of fun changing the commands around. 

I'll make the source code available to download next week. In the meantime remember to have a look at the speech recognition tutorial. I'll post a link to the download soon.

Contact: michaelpalmer.mp@gmail.com
Website: www.michaelpalmerwebdesign.com

Or leave a comment on my YouTube page

No comments:

Post a Comment