Hi folks,

I am trying to use Sphinx with InMoov. I beleive I have set everything up correctly, but Sphinx does not seem to be listening/understanding what I say?

I an running on Ubuntu and using the "sound settings" I can see that the OS is hearing a signal when I speak, but how can I debug what is going on (or not going on)?

The voice just keeps saying "did you say...." every few minutes.




6 years 6 months ago


I have never have much success using Sphinx, it drive me crazy.

try to use webkitspeech reconizer, it's working like a charm


6 years 6 months ago

Hi Andrew,

Yes Sphinx can be a challenge sometimes, and webkit speech is more accurate and responsive. 
You can attempt to debug by setting breakpoints and watching the logfile or console when its running..

Line input quaility affects Sphinx a lot more than webkit speech.

Hope this is helpful.


Most of what I have read seems to point at the webkit, I just didn't want to be tied to a web connection and always a bit paranoid about an open microphone to the web - no different to my phone I guess?



Its the same service as your Phone.
Many people get frustrated with Sphinx, and development seems to have taken as sebatical..
It was a Cargegie Mellon University Project - but has been not very active for the last 6+ years ... 

There's another one called "pocket sphinx" which gained some activity, because of the small size, and it was basically designed to run on phones as an open source alternative.

I wish I had the time to invest in Sphinx to make it better, but I spend all my 'spare' time trying to get things for the next MRL release ..

If I were you, I'd give it a try, even if you go back to Sphinx you have something to compare.


6 years 6 months ago

Hi ! does Webkit ear function can work autonomous ? I see somewhere we must click on button to use it