Model Adaptation using pocketsphinx

[palmero76]

Hi, since i've got a lot of registration where noise removal is applied (so,
there are sentences sometimes a little bit confused), could i use them to
adapt my model? The main fact is that the sentences to be recognized will be
the same as the ones for the model adaptation (so, with noise removal
applied). I think that trying model adaptation with those files will be
efficient to recognize other sentences where noise removal is applied, or
there's something wrong in my thought?
Thank you!

[Nickolay V. Shmyrev]

Hello

I think that trying model adaptation with those files will be efficient to
recognize other sentences where noise removal is applied, or there's something
wrong in my thought?

This is correct.

[palmero76]

Thank you. Is it necessary to follow ALL the steps in the tutorial (http://cm
usphinx.sourceforge.net/wiki/tutorialadapt) for best adaptation? Or there's something that i can
jump? More steps means more accuracy? Thanks again in advance.

p.s: i can't find, in the API documentation, the description and what a single
function can do (sincerely, i don't see for example the ps_decode_raw function
description in the doxygen). Maube i've got to read directly from the source
code? Thanks again!

[Nickolay V. Shmyrev]

Is it necessary to follow ALL the steps in the tutorial (http://cmusphinx.s
ourceforge.net/wiki/tutorialadapt) for best adaptation? Or there's something that i can jump?

You need to complete all the steps.

i can't find, in the API documentation, the description and what a single
function can do

I see the description here

http://cmusphinx.sourceforge.net/api/pocketsphinx/pocketsphinx_8h.html#ac9a9a
0df5cebda28876c1e95bd197321

[palmero76]

THANK YOU very much (and sorry for the documentation thing) ;)