<!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 4.01 Transitional//EN">
<html>
<head>
<meta content="text/html; charset=ISO-8859-1"
http-equiv="Content-Type">
</head>
<body bgcolor="#ffffff" text="#000000">
I just installed pocketshpinx (everything that search in synaptic
came up with for "pocketsphinx") and the gst stuff installed fine.
To make sure I ran "python" in a terminal and then "gst import" and
that was good.<br>
So after installing everything with synaptic I ran
"pocketsphinx_wsj" and it ran. I'm not sure how it can work for
anything remotely complex speechwise.<br>
AS I don't have an internal mic I then hooked up my usb logitech
video cam and mic and ran the command again. It worked perfectly and
upon the READY prompt I spoke clearly. Pocketsphinx heard this ( and
several other attempts):<br>
<br>
<i><b>"What is the date?" got:</b></i><br>
000000000: THE ADDED EIGHT EDITOR ERA EIGHT (-90853210)<br>
READY....<br>
Listening...<br>
Stopped listening, please wait...<br>
INFO: cmn_prior.c(121): cmn_prior_update: from < 52.78 -4.94
0.21 1.90 -3.85 -0.74 -1.96 0.11 0.24 -0.10 -0.88 0.29 -0.09
><br>
INFO: cmn_prior.c(139): cmn_prior_update: to < 53.30 -4.87
0.59 2.10 -3.96 -0.77 -1.97 0.10 0.16 -0.05 -0.93 0.33 -0.07
><br>
INFO: ngram_search_fwdtree.c(1450): 1998 words recognized
(12/fr)<br>
INFO: ngram_search_fwdtree.c(1452): 498588 senones evaluated
(2986/fr)<br>
INFO: ngram_search_fwdtree.c(1454): 414984 channels searched
(2484/fr), 71821 1st, 76167 last<br>
INFO: ngram_search_fwdtree.c(1458): 5464 words for which last
channels evaluated (32/fr)<br>
INFO: ngram_search_fwdtree.c(1461): 23496 candidate words for
entering last phone (140/fr)<br>
<br>
<i><b>Again, "Can you hear me?" got:</b></i><br>
000000001: THEN YOU HEAR ME (-28521979)<br>
READY....<br>
Listening...<br>
Stopped listening, please wait...<br>
INFO: cmn_prior.c(121): cmn_prior_update: from < 53.30 -4.87
0.59 2.10 -3.96 -0.77 -1.97 0.10 0.16 -0.05 -0.93 0.33 -0.07
><br>
INFO: cmn_prior.c(139): cmn_prior_update: to < 53.19 -4.62
0.69 2.01 -3.96 -0.80 -2.05 0.07 0.07 -0.03 -0.92 0.34 -0.06
><br>
INFO: ngram_search_fwdtree.c(1450): 3488 words recognized
(23/fr)<br>
INFO: ngram_search_fwdtree.c(1452): 395780 senones evaluated
(2656/fr)<br>
INFO: ngram_search_fwdtree.c(1454): 368169 channels searched
(2470/fr), 61806 1st, 99997 last<br>
INFO: ngram_search_fwdtree.c(1458): 6458 words for which last
channels evaluated (43/fr)<br>
INFO: ngram_search_fwdtree.c(1461): 20215 candidate words for
entering last phone (135/fr)<br>
<br>
<i><b>"testing, 1, 2, 3" got:</b></i><br>
000000002: SEEING WANTS TO RE (-28710067)<br>
READY....<br>
Listening...<br>
Stopped listening, please wait...<br>
INFO: cmn_prior.c(121): cmn_prior_update: from < 53.69 -4.78
0.44 1.85 -3.88 -0.67 -1.96 0.12 0.06 -0.08 -0.94 0.30 -0.04
><br>
INFO: cmn_prior.c(139): cmn_prior_update: to < 53.83 -5.04
0.26 1.77 -4.00 -0.51 -1.66 0.11 0.05 -0.13 -1.06 0.23 0.01
><br>
INFO: ngram_search_fwdtree.c(1450): 4423 words recognized
(17/fr)<br>
INFO: ngram_search_fwdtree.c(1452): 785596 senones evaluated
(3057/fr)<br>
INFO: ngram_search_fwdtree.c(1454): 759449 channels searched
(2955/fr), 109015 1st, 182918 last<br>
INFO: ngram_search_fwdtree.c(1458): 11348 words for which last
channels evaluated (44/fr)<br>
INFO: ngram_search_fwdtree.c(1461): 45597 candidate words for
entering last phone (177/fr)<br>
<br>
<i><b>"I am a man" got:</b></i><br>
000000003: ALL I AM A PLAN (-44454889)<br>
READY....<br>
Listening...<br>
Stopped listening, please wait...<br>
INFO: cmn_prior.c(121): cmn_prior_update: from < 53.83 -5.04
0.26 1.77 -4.00 -0.51 -1.66 0.11 0.05 -0.13 -1.06 0.23 0.01
><br>
INFO: cmn_prior.c(139): cmn_prior_update: to < 54.65 -4.53
0.03 1.19 -3.88 -0.39 -1.73 0.25 0.11 -0.05 -1.08 0.10 0.01
><br>
INFO: ngram_search_fwdtree.c(1450): 1407 words recognized
(11/fr)<br>
INFO: ngram_search_fwdtree.c(1452): 271057 senones evaluated
(2151/fr)<br>
INFO: ngram_search_fwdtree.c(1454): 226017 channels searched
(1793/fr), 45466 1st, 45724 last<br>
INFO: ngram_search_fwdtree.c(1458): 3404 words for which last
channels evaluated (27/fr)<br>
INFO: ngram_search_fwdtree.c(1461): 13482 candidate words for
entering last phone (107/fr)<br>
<i><b><br>
And finally the famous "Hello World" got:</b></i><br>
000000004: THE LOAN WORLD (-19829280)<br>
<br>
<br>
On 04/09/2012 04:41 PM, Sam Noble wrote:
<blockquote cite="mid:20120409224102.GB10473@thepromisedlan.org"
type="cite">
<pre wrap="">On Mon, Apr 09, 2012 at 04:22:31PM -0600, Steve Katona wrote:
</pre>
<blockquote type="cite">
<pre wrap="">Thanks for the comments. Still looking...sk
</pre>
</blockquote>
<pre wrap="">
</pre>
<blockquote type="cite">
<blockquote type="cite">
<pre wrap="">But I futzed with it for a while and haven't yet managed to get a
working version of the gstreamer plugins for pocketsphinx going.
Plus I have no idea how it well it would work if you did get it built or
found some working binaries. Sphinx is pretty good at reading back "GO
FORWARD 10 METERS" but I've never really seen it in action on live
conversational speech.
</pre>
</blockquote>
</blockquote>
<pre wrap="">
Hmm, it looks like Ubuntu packages the pocketsphinx gstreamer plugins,
so maybe just install gaupol and give it a try.
_______________________________________________
nmglug mailing list
<a class="moz-txt-link-abbreviated" href="mailto:nmglug@lists.nmglug.org">nmglug@lists.nmglug.org</a>
<a class="moz-txt-link-freetext" href="http://lists.nmglug.org/listinfo.cgi/nmglug-nmglug.org">http://lists.nmglug.org/listinfo.cgi/nmglug-nmglug.org</a>
</pre>
</blockquote>
</body>
</html>