Take the 2-minute tour ×
Ask Ubuntu is a question and answer site for Ubuntu users and developers. It's 100% free, no registration required.

Although I like face recognition, I'd prefer to use voice recognition login. Is it possible? If it's possible how do I setup the software?

share|improve this question
21  
Upvoted because I'm interested as well, but be warned, face recognition can often be defeated by a good picture, and voice recognition can sometimes be defeated with a voice recorder. Ideally, voice recognition would involve reading back a randomly chosen set of words. –  Jeff Welling Mar 31 '12 at 2:52
2  
There is a VoiceAuth PAM module, but the link from linux-pam.org/modules.html goes to nowhere. Wayback machine has a copy of the software from 2006: wayback.archive.org/web/… It compiles ok (you need to install at least sudo apt-get install libpam-dev libasound-dev), and comes with some documentation. But I don't have a microphone to try it. –  taneli Mar 31 '12 at 21:00
    
From the readme, it doesn't as of then support challenge response. It only supports a single voice password, but it is a start. –  hbdgaf Apr 1 '12 at 1:15
6  
It's been mentioned on Mark Shuttleworth's blog that he'd like to see voice recognition as a part of the Unity HUD. If it's implemented successfully there, we might see it become adopted into other areas of the desktop (like the login screen) as well, but that's just pure speculation on my part. –  Christopher Kyle Horton Apr 1 '12 at 11:40
2  
I hope you don't need login to work when you have a cough or cold! :D –  Sepero Sep 9 '12 at 8:34

3 Answers 3

No, with the current set of software for Ubuntu, it is not possible to login by voice.

There is no software to allow you to do this. Voice recognition is still a very new and imperfect technology. Everyone can agree it would be cool and high-tech to walk up to your computer, say something, have it say "Access Granted" and log us in. But right now, we can only dream.

To create a secure voice login would require an extraordinary amount of computer power to accurately analyze the voice. Apple's Siri sends the data to servers to have them report back what was said, because it does not have the power to do it on it's own.

However, I think it would be extremely cool and is a great idea. If you would like to get the ball rolling, head over to http://brainstorm.ubuntu.com site and start the discussion there. You'll get feedback from lots of people including Ubuntu developers, and there if enough people think it's a good idea it'll happen.

share|improve this answer
1  
I agree mostly. 'would require an extraordinary amount of computer power to accurately analyze the voice' is not necessarily so depending on how it's implemented. For handhelds yes. For desktops, probably not so much. I think part of pushing all the requests to Siri backend is Apple crowd-sourcing a HUGE voice corpus, although it's definitely required for handhelds. –  hbdgaf Apr 6 '12 at 16:20
    
You have a point. However, it depends much on the desktop. A one-size fits all solution would not work in this case, due to the large variety of computers running Ubuntu. But, as a result, some systems would be more secure then others. If the same amount of computing power was required to log in to login to all Ubuntu systems, some requires would take several minutes to log in, annoying the user. –  William Apr 6 '12 at 19:05
    
Valid point. I default to 'use netbook edition' in the latter case. –  hbdgaf Apr 6 '12 at 20:50
    
Yes, an option to enable/disable the login in the System Settings window would be good as well. –  William Apr 6 '12 at 20:53

There are various kinds of voice authentication. For example, doing speech-to-text and then verifying the decoded text; voice pattern analysis; singing a tune; analyzing the frequency spectrum of the voice and so on. Depending on what you are looking for, you are early or very early in the game.

However,

You could allow automatic login with or without a password, and then start some software tool that would prompt and check your voice, speech, anything. For proof of concept purposes, you may start the software from your .login or .bash_rc script, but beware that they may be terminated from the keyboard. If authentication fails, execute logout.

For voice recognition software, check this Wikipedia article or project Julius.

The alternative (the hard way) would be to find or implement a Pluggable Authentication Module (PAM), similar to the fingerprint authentication for TI laptops. Although I can not point you to a ready to use module, it is technically possible.

share|improve this answer

this maybe helpful: voice recognition in ubuntu

share|improve this answer
10  
Welcome to Ask Ubuntu! Whilst this may theoretically answer the question, it would be preferable to include the essential parts of the answer here, and provide the link for reference. –  Bruno Pereira Apr 6 '12 at 6:06

Your Answer

 
discard

By posting your answer, you agree to the privacy policy and terms of service.

Not the answer you're looking for? Browse other questions tagged or ask your own question.