<div dir="ltr"><div><div><div>Hi,<br><br></div>I was able to achieve this using:<br><br><a href="http://zaf.github.io/asterisk-speech-recog/">http://zaf.github.io/asterisk-speech-recog/</a><br><br></div>I needed to change code, so it wasn't working out-of-the-box. I did this couple of years ago and unfortunately I do not have code anymore. But it wasn't too difficult.<br></div><div><br></div><div>With kind regards,<br></div></div><div class="gmail_extra"><br clear="all"><div><div class="gmail_signature" data-smartmail="gmail_signature"><div dir="ltr">Jurijs<br></div></div></div>
<br><div class="gmail_quote">On Wed, Dec 6, 2017 at 4:33 PM, Jonathan H <span dir="ltr"><<a href="mailto:lardconcepts@gmail.com" target="_blank">lardconcepts@gmail.com</a>></span> wrote:<br><blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex">Briefly: I want to be able to have "press or say (number)", with<br>
Asterisk listening for a spoken number, but accepting a DTMF digit,<br>
too.<br>
<br>
I'm posting everything I found so far, here, partly to show working,<br>
but also in case anyone else finds it useful. So, moving on....<br>
<br>
This looked hopeful for a moment until I realised that it doesn't do DTMF:<br>
<a href="https://wiki.asterisk.org/wiki/display/AST/Asterisk+15+Application_SpeechBackground" rel="noreferrer" target="_blank">https://wiki.asterisk.org/<wbr>wiki/display/AST/Asterisk+15+<wbr>Application_SpeechBackground</a><br>
<br>
So then there's<br>
<a href="https://wiki.asterisk.org/wiki/display/AST/Asterisk+15+Application_Record" rel="noreferrer" target="_blank">https://wiki.asterisk.org/<wbr>wiki/display/AST/Asterisk+15+<wbr>Application_Record</a>,<br>
which can terminate on any DTMF key with "y", but according to the<br>
docs, "RECORD_STATUS" only sets a flag of "DTMF" (A terminating DTMF<br>
was received ('#' or '*', depending upon option 't')).<br>
So, I don't get to know which key was pressed via that method, either.<br>
<br>
There's very little information I can find about the built-in<br>
functions for speech recognition.<br>
<a href="https://wiki.asterisk.org/wiki/display/AST/Speech+Recognition+API" rel="noreferrer" target="_blank">https://wiki.asterisk.org/<wbr>wiki/display/AST/Speech+<wbr>Recognition+API</a><br>
doesn't actually explain how to integrate the actual speech engines.<br>
<br>
In this previous forum post,<br>
<a href="https://community.asterisk.org/t/asterisk-15-jack-streams-speech-recognition-so-many-questions/72108/2" rel="noreferrer" target="_blank">https://community.asterisk.<wbr>org/t/asterisk-15-jack-<wbr>streams-speech-recognition-so-<wbr>many-questions/72108/2</a><br>
, jcolp explained that most people don't use the speech interface<br>
anyway, because<br>
"Asterisk modules are written in C, and it’s more difficult to do<br>
things in that fashion. Using the Record and ship it off using Python,<br>
etc, is just easier and gets the job done for a lot of people to where<br>
they find it acceptable.<br>
So, AGI it is! But I'm still stuck on how I record for speech AND get<br>
a DTMF if it was dialled.<br>
<br>
Regarding speech in general, even "Asterisk - The Definitive Guide" just says:<br>
<br>
"Asterisk does not have speech recognition built in, but there are<br>
many third-party speech<br>
recognition packages that integrate with Asterisk. Much of that is<br>
outside of the scope<br>
of this book, as those applications are external to Asterisk" - helpful!<br>
<br>
The speech-rec mailing list at<br>
<a href="http://lists.digium.com/pipermail/asterisk-speech-rec/" rel="noreferrer" target="_blank">http://lists.digium.com/<wbr>pipermail/asterisk-speech-rec/</a> hasn't been<br>
posted to since 2013<br>
<br>
Someone else asked about speech recognition and unimrcp in this post:<br>
<a href="http://lists.digium.com/pipermail/asterisk-users/2017-February/290875.html" rel="noreferrer" target="_blank">http://lists.digium.com/<wbr>pipermail/asterisk-users/2017-<wbr>February/290875.html</a><br>
<br>
uniMCRP <a href="https://mojolingo.com/blog/2015/speech-rec-asterisk-get-started/" rel="noreferrer" target="_blank">https://mojolingo.com/blog/<wbr>2015/speech-rec-asterisk-get-<wbr>started/</a><br>
<a href="http://www.unimrcp.org/manuals/html/AsteriskManual.html#_Toc424230605" rel="noreferrer" target="_blank">http://www.unimrcp.org/<wbr>manuals/html/AsteriskManual.<wbr>html#_Toc424230605</a><br>
This has a Google Speech Recogniser plugin, but it's $50 per channel<br>
<a href="http://www.unimrcp.org/gsr" rel="noreferrer" target="_blank">http://www.unimrcp.org/gsr</a><br>
<br>
*Reasons to use Lex over Google TTS*<br>
• Has just been released in eu-west-1:<br>
<a href="https://forums.aws.amazon.com/ann.jspa?annID=5186" rel="noreferrer" target="_blank">https://forums.aws.amazon.com/<wbr>ann.jspa?annID=5186</a><br>
• Supports 8KHz telepony <a href="https://forums.aws.amazon.com/ann.jspa?annID=4775" rel="noreferrer" target="_blank">https://forums.aws.amazon.com/<wbr>ann.jspa?annID=4775</a><br>
• Is in the core AWS SDK<br>
<a href="http://docs.aws.amazon.com/AWSJavaScriptSDK/latest/AWS/LexRuntime.html" rel="noreferrer" target="_blank">http://docs.aws.amazon.com/<wbr>AWSJavaScriptSDK/latest/AWS/<wbr>LexRuntime.html</a><br>
• Has a number slot type:<br>
<a href="http://docs.aws.amazon.com/lex/latest/dg/built-in-slot-number.html" rel="noreferrer" target="_blank">http://docs.aws.amazon.com/<wbr>lex/latest/dg/built-in-slot-<wbr>number.html</a><br>
 - this means no accidental recognition of "won", "one" or "juan" instead of 1!<br>
<br>
The pricing is definitely right: "The cost for 1,000 speech requests<br>
would be $4.00, and 1,000 text requests would cost $0.75. From the<br>
date you get started with Amazon Lex, you can process up to 10,000<br>
text requests and 5,000 speech requests per month for free for the<br>
first year".<br>
<br>
Amazon Transcribe looks promising too, but is only available for<br>
developer invitation at this time:<br>
<a href="https://aws.amazon.com/transcribe/" rel="noreferrer" target="_blank">https://aws.amazon.com/<wbr>transcribe/</a> <a href="https://aws.amazon.com/transcribe/pricing/" rel="noreferrer" target="_blank">https://aws.amazon.com/<wbr>transcribe/pricing/</a><br>
<br>
But all I need now is the quickest, simplest way to send Lex a short<br>
8KHz file and get a single digit back, as quickly and reliably as<br>
possible.<br>
<br>
Before I travel too far down this road, can someone point me in the<br>
right direction and possibly steer me away from the wrong path?!<br>
<span class="HOEnZb"><font color="#888888"><br>
--<br>
______________________________<wbr>______________________________<wbr>_________<br>
-- Bandwidth and Colocation Provided by <a href="http://www.api-digital.com" rel="noreferrer" target="_blank">http://www.api-digital.com</a> --<br>
<br>
Check out the new Asterisk community forum at: <a href="https://community.asterisk.org/" rel="noreferrer" target="_blank">https://community.asterisk.<wbr>org/</a><br>
<br>
New to Asterisk? Start here:<br>
      <a href="https://wiki.asterisk.org/wiki/display/AST/Getting+Started" rel="noreferrer" target="_blank">https://wiki.asterisk.org/<wbr>wiki/display/AST/Getting+<wbr>Started</a><br>
<br>
asterisk-users mailing list<br>
To UNSUBSCRIBE or update options visit:<br>
   <a href="http://lists.digium.com/mailman/listinfo/asterisk-users" rel="noreferrer" target="_blank">http://lists.digium.com/<wbr>mailman/listinfo/asterisk-<wbr>users</a></font></span></blockquote></div><br></div>