Thread: voice recognition to text
View Single Post
mwiktowy is offline mwiktowy
2007-02-24 , 18:32
Posts: 373 | Thanked: 56 times | Joined on Dec 2005 @ Ottawa, ON
#8
The section of that website that caught my attention was:

You can also download telephone-bandwidth models separately. To use these with raw audio data you need the following extra command-line options:

-nfft 256
-nfilt 31
-lowerf 200
-upperf 3500
-samprate 8000
Since the 770 and N800 seem to capture audio at 8000 Hz sampling rate (based on the maemorecorder abilities), this voice model might be the way to go. Plus it is only 8 MB or so rather than the 25 MB that you speak of.

They are available here:
http://www.speech.cs.cmu.edu/pockets...linterp.tar.gz