SpeechControl: The Speech Recognition Project for Ubuntu is Underway!

In computing, ‘Speech Recognition’ is a technology that lets you interact with your computer using voice commands rather than the standard input devices such as the Keyboard, Mouse etc. Though it is much easier for us humans but implementing it on a computer has proved to be a difficult task.

There are no speech-recognition engines as advanced as the ones you’d seen in movies, such as “iRobot”. But, then again, in recent versions of Windows (starting with Vista) have some pretty impressive voice command recognition technologies (I just spent like 20 minutes with it and got it to open Firefox, Notpad, Explorer, MS Paint etc).

In GNU/Linux, this is still not that widely used and the OS in general does not come with a speech recognition engine with it (as far as I know).

But that’s not to say, there aren’t any open source projects for creating speech control systems, in fact there a few. And by using a speech recognition system called “Sphinx” (GUI is written using Qt), and as Mark Shuttleworth promised that in the near future, speech recognition will come to Ubuntu …

… well, it seems that the project is well underway as there’s already PPA (only has packages for 11.10 and 12.04 at the moment) that you can use to install a GUI tool that lets you do a bit of a speech recognition testing!.

It didn’t really work in my PC though …

However, please remember that it’s still in its early development and I don’t even know whether you could even use it. Because to use speech recognition in a Computer, there are few basic things that the software should be capable of doing:

It might not be able to fully function yet, still looks promising :D

1. First, we gotta install the software (duh ;-)).

2. Then we have to properly setup the application (such as making sure that your audio hardware/microphones are working properly etc).

3. Then in the last step, you have to “train” the application’sĀ acoustic engine, so it would be able to recognize your voice.

A voice training session for improving the “acoustic engine” in Windows 7 …

This usually involves the speech recognition engine (acoustic model) giving you words and you spelling it back to it, so it would get familiar with your voice. And the more you interact with it (as the time goes on) the more efficient the system will be at recognizing your voice.

The “SpeechControl” tool installed quite fine in my Ubuntu 11.10 Oneiric Ocelot. And, it has a very user friendly wizard type setup process too. But it failed to pickup my voice input through the microphone. Now under Pulse audio, my microphone works just fine and even shows volume levels as I speak.

But, in “SpeechControl’s” window, it failed to detect my voice inputs. Then I choose the second option in its audio settings window, and it seemed to pick up the voice inputs.

And the volume bar went up and down even when I’m not talking. Still, I was able to talk few words into it and clicked “Next” button and the next window presented saying: “Accuracy 24%” which is not that impressive (could be because of the microphone issue).

Then it gave me a few lines of words to speak, for training the “acoustic engine”. But I couldn’t go beyond that. So it’s probably a bug.

Finally, you’ll be able to curse your computer! ;-)

This however is not a surprise as it’s in its early stages of development and perhaps these training exercises are not even fully implemented yet. But, if you use Ubuntu 11.10 or 12.04, and can’t wait anymore ;-), then you can give it a go and see what happens.

The app also has few options such as Desktop integration. Although I’m not sure what it does but perhaps things like, opening the Dash, showing the desktop, opening items on it or deleting stuff (yikes!) etc might be available.

Again, I’m pretty sure it’s not fully functional, but if you’re like me fully excited to hear it, then you can install it in Ubuntu 12.04 Precise Pangolin and 11.10 Oneiric Ocelot by using one of its PPA channels.

Using the “stable” PPA: Please remember that the packages aren’t fully available in the stable PPA (as I’m writing this post) thus I couldn’t install it. But you can try it if you like by using the below commands.

sudo apt-add-repository ppa:speechcontrol-devel/stable

sudo apt-get update

sudo apt-get install speechcontrol

Using the “unstable” PPA:

However, after entering the above commands, if you get an error saying “couldn’t locate the package …” then you can use the “unstable” PPA, which should work. For that, please use the below commands.

sudo apt-add-repository ppa:speechcontrol-devel/unstable

sudo apt-get update

sudo apt-get install speechcontrol

After that, search “speechcontrol” in the Dash and you should be guided by the app (a wizard type easy to use interface).

You can visit this official Ubuntu Wiki Page and “SpeechControl” also has a Wiki page of its own (but it currently miss many documentation entries).

But please remember that, it’s in heavy development right now and it might not work in your Ubuntu PC as well. But still, it’s kinda cool ain’t it? ;-).

2 thoughts on “SpeechControl: The Speech Recognition Project for Ubuntu is Underway!

    1. Gayan Post author

      Hello Colin,

      I didn't do anything special. It just piked up my inputs as I spoke.

      But as I've mentioned in the post, I didn't really work and haven't tried it ever since either :)

Leave a Reply

Your email address will not be published. Required fields are marked *

You may use these HTML tags and attributes: <a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <cite> <code> <del datetime=""> <em> <i> <q cite=""> <strike> <strong>