I am looking to integrate asterisk with some speech recognition engine
I understand that there are open source solutions such as Sphinx and commercial solutions such as LumenVox available for the same.
The only issue is that I want a speech recognition software for Indian languages (primarily hindi) and Indian accent. I am looking for a fairly
comprehensive grammar, so that I cannot use English acoustic models
and somehow make them work for my case.
Does anyone have any experience with doing the same? How difficult would it be to configure an open source software like Sphinx for a new acoustic model and a different accent.