voice-driven-ui team mailing list archive
-
voice-driven-ui team
-
Mailing list archive
-
Message #00009
pocketsphinx and accuracy
Hi all,
I have installed pocketsphinx (from the latest upstream not the packaged
version) and got it kind of working with a view to helping with the
voice driven UI concept. I am struggling a bit with the accuracy, this
is the output of me counting from one to ten four times in a row and
using pocketsphinx_continuous
number two to review goal will it says clinton thinks million tons
more than two the jury whole lot of six seven and knowing time
will need to jury pool hall is six seven and knowing them
one two three full five six that pregnant and
I was kind of hoping for a bit better, not sure if my levels are wrong
or something. I could go through the acoustic model training and
language model definition but I am not sure if I have got the basics
right. Is this the level of accuracy you would expect?
--
Alan Bell
The Open Learning Centre
Web: http://www.theopenlearningcentre.com
Mob: +44 (0)7738 789190
Tel: +44 (0)844 3576000
The Open Learning Centre is a trading name of Bell Lord Ltd,
a company registered in England and Wales #05868943.
VAT Registration #GB 901 4715 55