[asterisk-users] Opensource Speech recognition for Asterisk

[Paul Belanger]

On Sat, Aug 21, 2010 at 7:09 PM, Zeeshan Zakaria <zishanov at gmail.com> wrote:
> Your experience could be different and better then most, and you have
> certainly complete right of your own opinion.
>
Speech recognition is only as effective as your grammars and they are
never 100%. The require lots tuning and analyzing to be effective.  My
clients I works with simply believe once the IVR is developed and
installed, the work end there; This is never the case.

We actively consult with a Speech Linguist when are in the tuning
phase of the grammars.  Countless hours listening to recorded calls,
rebuilding grammars, in an attempt to increase accuracy.  Lots of hard
work, but when you actively improve grammars, recognition rates
increase.

-- 
Paul Belanger | dCAP
Polybeacon | Consultant
Jabber: paul.belanger at polybeacon.com | IRC: pabelanger (Freenode)
blog.polybeacon.com