Code Monkey home page Code Monkey logo

fsqa's Introduction

Voice Quality Analyzer

This is a set of scripts that integrate Sevana AQuA software with FreeSWITCH. They allow you to receive and record calls, compare the received audio with the reference file, and store the results in the database.

We assume that Sevana AQuA executable and the license file are located at the following locations:

/opt/sevana/aqua-v
/opt/sevana/aqua-v.lic

The installation instructions are designed for Debian Linux, and can be adapted to other OS distributions. The audio analysis scores are written to local syslog and visible in /var/log/user.log, and to a local MySQL database.

The FreeSWITCH dialplan is supposed to call the run_proc script with the file name of the audio recording in WAV format. The file name is expected to contain several important parts dividied by underscore: 1) UNIX Epoch timestamp; 2) UUID of the call; 3) call source indicator string. The source indicator is a string that allows you to identify how this recording has been made. For example, it can contain the caller and callee IDs. The source indicator is stored as-is in the database, and shoudl not exceed 256 characters.

The script forks a child process and returns the control immediately to FreeSWITCH, so that the channel can be removed from FreeSWITCH memory. The forked children make sure that only one execution of AQuA binary happens at a time.

The FSQA_SOUNDFILES SQL table contains the names of reference WAV files. By default, the audio is compared against the 2-minute ITU-T reference recording, and an alternative reference audio can be supplied for a specific source indicator.

Installing the scripts and prerequisites

apt-get update && apt-get install -y git mysql-client mysql-server \
  libdbd-mysql-perl libuuid-perl

cd /opt
git clone https://github.com/voxserv/fsqa.git

mysqladmin create fsqa
mysql fsqa </opt/fsqa/fsqa_dbtables.sql 

mkdir -p \
  /var/spool/fsqa/audio_archive \
  /var/spool/fsqa/audio_rec \
  /var/spool/fsqa/temp

chown freeswitch:freeswitch \
  /var/spool/fsqa/audio_archive \
  /var/spool/fsqa/temp


cat >>/etc/fstab <<'EOT'
tmpfs /var/spool/fsqa/audio_rec tmpfs defaults,size=1g 0 0
EOT

mount /var/spool/fsqa/audio_rec

mkdir /opt/sound/
cd /opt/sound/
wget http://www.voxserv.ch/media/ITU-T_P_50_BRITISH_ENGLISH.wav
wget http://www.voxserv.ch/media/ITU-T_P_50_BRITISH_ENGLISH_30s.wav

Example dialplan

This repository contains fsqa.xml which can be copied as-is into /etc/freeswitch/dialplan (FreeSWITCH dialplan directory if you installed FreeSWITCH from precompiled packages):

cp /opt/fsqa/fsqa.xml /etc/freeswitch/dialplan/

The following call processing is defined in fsqa.xml:

  • If the caller ID number is in E.164 format, the leading plus sign is removed.

  • Extension pstn_XXX (here XXX is the DID number that receives the call from PSTN) plays the 2-minute reference file and records the inbound audio channel, and then executes the quality analysis script. The source indicator string is a combination of caller and callee ID's.

  • Extension pstnm_XXX (muted) plays the silence stream and records the inbound audio, and sends it to the quality processing after the call ends.

  • Extension fsqa_YYY (here YYY is an arbitrary string) sets YYY as the source indicator, plays back the reference audio, and sends the recording to quality analysis.

  • Extension fsqam_YYY does the same as above, but the silence stream is played back.

  • Extension speaker just plays the reference audio and hangs up.

So, normally you would define a PSTN gateway in sip_profiles/ in your FreeSWITCH configuration, and send the incoimg calls to pstn_123456789 extension (here 123456789 stands for your DID number) in fsqa context. The easiest way to do this is to define a separate SIP profile and bind it to a dedicated UDP port, so that all inbound calls are sent to fsqa context.

This repository includes sip_profiles--fsqa.xml which defines a SIP profile binding to port 5083 and setting the codec preferences to G711 only.

cd /etc/freeswitch
cp /opt/fsqa/sip_profiles--fsqa.xml sip_profiles/fsqa.xml
mkdir sip_profiles/fsqa

# use your own ITSP name and SIP credentials
cat >sip_profiles/fsqa/itsp_example.xml <<'EOT'
<include>
  <!-- DID: 123456789 -->
  <gateway name="example_itsp_123456789">
    <param name="username" value="123456789"/>
    <param name="proxy" value="sip.example.com"/>
    <param name="password" value="secret"/>
    <param name="extension" value="pstn_123456789"/>
    <param name="expire-seconds" value="120"/>
    <param name="register" value="true"/>
    <param name="register-transport" value="udp"/>
    <param name="retry-seconds" value="30"/>
    <param name="ping" value="30"/>
  </gateway>
</include>
EOT

systemctl restart freeswitch

Author and License

Copyright (c) 2016 Stanislav Sinyagin [email protected]

This software is distributed under the MIT license.

fsqa's People

Contributors

ssinyagin avatar

Watchers

James Cloos avatar

Recommend Projects

  • React photo React

    A declarative, efficient, and flexible JavaScript library for building user interfaces.

  • Vue.js photo Vue.js

    ๐Ÿ–– Vue.js is a progressive, incrementally-adoptable JavaScript framework for building UI on the web.

  • Typescript photo Typescript

    TypeScript is a superset of JavaScript that compiles to clean JavaScript output.

  • TensorFlow photo TensorFlow

    An Open Source Machine Learning Framework for Everyone

  • Django photo Django

    The Web framework for perfectionists with deadlines.

  • D3 photo D3

    Bring data to life with SVG, Canvas and HTML. ๐Ÿ“Š๐Ÿ“ˆ๐ŸŽ‰

Recommend Topics

  • javascript

    JavaScript (JS) is a lightweight interpreted programming language with first-class functions.

  • web

    Some thing interesting about web. New door for the world.

  • server

    A server is a program made to process requests and deliver data to clients.

  • Machine learning

    Machine learning is a way of modeling and interpreting data that allows a piece of software to respond intelligently.

  • Game

    Some thing interesting about game, make everyone happy.

Recommend Org

  • Facebook photo Facebook

    We are working to build community through open source technology. NB: members must have two-factor auth.

  • Microsoft photo Microsoft

    Open source projects and samples from Microsoft.

  • Google photo Google

    Google โค๏ธ Open Source for everyone.

  • D3 photo D3

    Data-Driven Documents codes.