Speech to Text for inbound calls

General Help
1

hi all,

I want to setup a speech to text for my Support Incoming Calls.

Is there any possibility that I can Convert the audio to text and then send to my analysis team.

2


https://speech-to-text-demo.ng.bluemix.net/


3

jerson,

Do you have an example on how to configure this with FreePBX?

Your tutorial with Polly has saved us many times over :slight_smile:

Thanks!

4

I didn’t give the Polly a try but will give it.

however is there any example how to implement this for calls?

5

FYI, Polly is for Text To Speech not Speech To Text.

I was just mentioning it because jerson happened to write a great post on that.

6

Ok comtech thanks.

Is there anyway we can do speech to text and send it via email like voicemail notifications.

7

It looks like jerson just linked to Google and Watson as two methods of doing STT. You and I are both interested if there is a tutorial on how to implement this.

While waiting for jerson to reply, I found this on Google.

https://zaf.github.io/asterisk-speech-recog/
http://nerdvittles.com/?p=25136
http://nerdvittles.com/?p=593

8

yeah this is implemented by lot of companies so we need to implement in our company as requirement by the management.

9

Look this: https://jrklein.com/2015/08/17/asterisk-voicemail-transcription-via-ibm-bluemix-speech-to-text-api/

10

My example:

mkdir /var/lib/asterisk/sounds/stt/
fwconsole chown

vim /opt/stt.sh

curl -k -u yourusernameibm:password -X POST
–limit-rate 40000
–header “Content-Type: audio/wav”
–data-binary @/var/lib/asterisk/sounds/stt/$1.wav
“https://stream.watsonplatform.net/speech-to-text/api/v1/recognize?continuous=true&model=pt-BR_NarrowbandModel” 1>/var/lib/asterisk/sounds/stt/$1.txt

#Extract transcript results from JSON response
TRANSCRIPT=cat /var/lib/asterisk/sounds/stt/$1.txt | grep transcript | sed 's#^.*"transcript": "##g' | sed 's# "$##g' > /var/lib/asterisk/sounds/stt/stt-$1.txt
sed -e :a -e N -e ‘s/\n/ /’ -e ta /var/lib/asterisk/sounds/stt/stt-$1.txt >/var/lib/asterisk/sounds/stt/transcripted-stt-$1.txt

vim /etc/asterisk/extensions_custom.conf

[speech-to-text]

exten => s,1,Wait(3)
same => n,Set(stt=${STRFTIME(${EPOCH},%d%m%Y)})
same => n,Record(/var/lib/asterisk/sounds/stt/${stt}.wav,3,60)
same => n,System(sh /opt/stt.sh ${stt})
same => n,Hangup()