Link to home
Start Free TrialLog in
Avatar of Frosty555
Frosty555Flag for Canada

asked on

Voice recognition from a list of possibilities in Linux

Does anybody know of a way to perform voice recognition in linux, and make a best match of the recorded sound to a list of options?

E.g. if you had a contact list

Jane Doe
John Doe
Bill Smith
Anne Frank

And you had a .WAV file with somebody speaking one of these names,

and you end up with something like this:

#1:   John Doe,  95% match
#2:   Jane Doe,  60% match
#3:   Bill Smith,  3% match
#4:   Anne Frank,  1% match

I've looked at Sphinx, but without some sample code or a proof of concept I'm pretty lost, it is way too complicated for me.

I was kind-of hoping there was a third party tool already crafted that does something similar to this that I could invoke from a shell script or something.
Avatar of Phonebuff
Phonebuff
Flag of United States of America image

Have you looked at Cepstral TTS, LumenVox TTS or FEstival TTS ?
Avatar of Frosty555

ASKER

Those are all text-to-speech. I need to go the other way around.
Sorry, I miss read this first time -..  

What you want is a Voice XML solution, the one I have used in the past is VoiceGenie, but I think they were  purchased by someone else.

" ... recognition of spoken and DTMF key input, "

http://voxeo.com/glossary/what-is-voicexml/
Hi Phonebuff,

I've done a bit of research into VoiceXML, it looks like a nice way of describing what your IVR is going to do, what the voice options and paths are etc.

But I don't actually see a PRODUCT that I can actually use to implement it. I see a few hosted products but what do you suggest I sign up for / purchase / run a trial of?
ASKER CERTIFIED SOLUTION
Avatar of Phonebuff
Phonebuff
Flag of United States of America image

Link to home
membership
This solution is only available to members.
To access this solution, you must be a member of Experts Exchange.
Start Free Trial