Powered by Universal Speech Solutions LLC


Azure SR Plugin

Administrator Guide


Revision: 10

Distribution: Red Hat / Cent OS

Created: October 26, 2017

Last updated: March 16, 2021

Author: Arsen Chaloyan


Table of Contents


1  Overview.. 3

1.1         Applicable Versions 3

1.2         Supported Endpoints 3

1.3         Supported Distributions. 3

1.4         Authentication. 4

2  Installing RPMs Using YUM.. 5

2.1         Repository Configuration. 5

2.2         Repository Verification. 5

2.3         Azure SR Plugin Installation. 6

3  Installing RPMs Manually. 7

3.1         Package List 7

3.2         Package Installation Order 8

4  Obtaining License. 9

4.1         License Type. 9

4.2         Node Information. 9

4.3         License Installation. 9

5  Obtaining Service Credentials. 10

5.1         Service Subscription. 10

5.2         Installation of Credentials 10

Speech Service. 10

Custom Speech Service. 11

Speech Service Container 11

6  Configuring Server and Plugin. 13

6.1         Plugin Factory Configuration. 13

6.2         RTP Configuration. 13

6.3         Logger Configuration. 14

6.4         Azure SR Plugin Configuration. 14

7  Validating Setup. 15

7.1         Launching Server 15

7.2         Launching Client 15



1       Overview

This guide describes how to obtain and install binary packages for the Microsoft Azure Speech Recognition (SR) plugin to the UniMRCP server on Red Hat-based Linux distributions. The document is intended for system administrators and developers.

1.1      Applicable Versions

Instructions provided in this guide are applicable to the following versions.


UniMRCP 1.5.0 and above

UniMRCP Azure SR Plugin 1.0.0 and above

1.2      Supported Endpoints

The plugin supports the following Speech Service endpoints.


Service Endpoint


Speech Service (regional)

Azure SR 1.6.0 and above

Custom Speech Service (regional)

Azure SR 1.9.0 and above

Speech Service (containerized)

Azure SR 1.14.0 and above

Custom Speech Service (containerized)

Azure SR 1.14.0 and above

1.3      Supported Distributions

UniMRCP RPMs are currently available for x86_64 (64-bit) architecture only.


Operating System


End of Support

Red Hat / Cent OS 7

April 2018


Red Hat / Cent OS 8

January 2021



Note: packages for other distributions can be made available upon request. For more information, contact services@unimrcp.org.

1.4      Authentication

UniMRCP binary packages are available to authenticated users only. In order to register a free account with UniMRCP, please visit the following page.




Note: a new account needs to be verified and activated prior further proceeding.


2       Installing RPMs Using YUM

Using the Yellowdog Updater, Modifier (yum), a command-line package management utility for Red Hat-based distributions, is recommended for installation of UniMRCP binary packages.

2.1      Repository Configuration

The content of a typical yum configuration file, to be placed in /etc/yum.repos.d/unimrcp.repo, is provided below.



name=UniMRCP Packages for Red Hat / Cent OS-$releasever $basearch








name=UniMRCP Packages for Red Hat / Cent OS-$releasever noarch







The username and password fields included in the HTTPS URI must be replaced with the corresponding account credentials.

2.2      Repository Verification

In order to verify that yum can properly connect and access the UniMRCP repository, the following command can be used.


yum repolist unimrcp

yum repolist unimrcp-noarch


where unimrcp and unimrcp-noarch are names of the sections set in the yum configuration file above.


In order to retrieve a list of packages the UniMRCP repository provides, the following command can be used.


yum --disablerepo="*" --enablerepo="unimrcp" list available

yum --disablerepo="*" --enablerepo="unimrcp-noarch" list available


2.3      Azure SR Plugin Installation

In order to install the Azure SR plugin, including all the dependencies, use the following command.


yum install unimrcp-azure-sr


In order to install the additional data files for the sample client application umc, the following command can be used.


yum install umc-addons


Note: this package is optional and provides additional data which can be used for validation of basic setup.

3       Installing RPMs Manually

UniMRCP RPM packages can be installed manually using the rpm utility. Note, however, that the system administrator should take care of package dependencies and install all the packages in appropriate order.


The RPM packages have the following naming convention:





·         packagename is the name of a package

·         universion is the UniMRCP version

·         packageversion is the RPM release version

·         rhelversion is the Red Hat version

·         arch is the architecture (x86_64, i686, …)

3.1      Package List

The following is a list of UniMRCP RPM packages required for the installation of the Azure SR plugin.


Package Name



Azure SR plugin to the server.


UniMRCP edition of the libevent library.


Sample en-US data files used with umc. [Optional]


Node information retrieval tool, required for license deployment.


Shared library and application of the server.


Shared libraries and sample applications of the client. [Optional]


Set of demo plugins to the server. [Optional]


Data common for the client and the server.


UniMRCP edition of the Apache Portable Runtime (APR) library.


UniMRCP edition of the Apache Portable Runtime Utility (APR-Util) library.


UniMRCP edition of the Sofia SIP library.


3.2      Package Installation Order

Note that all the RPM packages provided by UniMRCP are signed by a GNU Privacy Guard (GPG) key. Before starting the installation, you may need to import the public key in order to allow the rpm utility to verify the packages.


rpm --import https://unimrcp.org/keys/unimrcp-gpg-key.public


Packages for the APR, APR-Util and Sofia-SIP libraries must be installed first.


rpm -ivh uniapr-$aprversion-$packageversion.el$rhelversion.$arch.rpm

rpm -ivh uniapr-util-$apuversion-$packageversion.el$rhelversion.$arch.rpm

rpm -ivh unisofia-sip-$sofiaversion-$packageversion.el$rhelversion.$arch.rpm


Then, a package containing common data for the client and the server, and a package for the server should follow.


rpm -ivh unimrcp-common-$universion-$packageversion.el$rhelversion.$arch.rpm

rpm -ivh unimrcp-server-$universion-$packageversion.el$rhelversion.$arch.rpm


Next, a package containing the utility tool unilicnodegen, required for license deployment.


rpm -ivh unilicnodegen-$toolversion-$packageversion.el$rhelversion.$arch.rpm


Next, a package containing the libevent library.


rpm -ivh unilibevent-$libeventversion-$packageversion.el$rhelversion.$arch.rpm


Finally, a package containing the Azure SR plugin should follow.


rpm -ivh unimrcp-azure-sr-$universion-$packageversion.el$rhelversion.noarch.rpm


4       Obtaining License

The Azure SR plugin to the UniMRCP server is a commercial product, which requires a license file to be installed.

4.1      License Type

The following license types are available:

·         Trial

·         Production

·         Test and Development

4.2      Node Information

The license files are bound to a node the product is installed on. In order to obtain a license, the corresponding node information needs to be retrieved and submitted for generation of a license file.


Use the installed tool unilicnodegen to retrieve the node information.




As a result, a text file uninode.info will be saved in the current directory. Submit the file uninode.info for license generation to services@unimrcp.org by mentioning the product name in the subject.

4.3      License Installation

The license file needs to be placed into the directory /opt/unimrcp/data.


cp umsazuresr_*.lic /opt/unimrcp/data

5       Obtaining Service Credentials


Instructions provided in this section are applicable to Azure SR 1.6.0 and above. Support for the old format remains in-tact, but is no longer documented.


In order to utilize either the deprecated Bing Speech API or the new Speech Service API, a corresponding service subscription key and an authentication endpoint need to be retrieved from the Microsoft Azure portal and further installed to the UniMRCP server.

5.1      Service Subscription

Navigate to the Microsoft Azure dashboard and create a new resource.


1.       Navigate to the Dashboard.


2.       Create a new resource (+).

3.       Type ‘Speech’ in the text box. (Note: for the deprecated Bing service, type ‘Bing Speech’)

4.       Select and create the resource by filling out all the required parameters based on your needs.

5.       Go to the created resource.

6.       Collect one of the two keys (1) and the region-based authentication endpoint (2).


5.2      Installation of Credentials

Create a text file cognitive.subscription.key in the directory /opt/unimrcp/data.


nano /opt/unimrcp/data/cognitive.subscription.key


Proceed with the instructions for one of:

·         Speech Service (default)

·         Custom Speech Service

·         Speech Service Container

Speech Service

Place the collected key and the authentication endpoint in the following JSON format. Leave the service endpoint empty by default.



            "auth-endpoint": "*******************************",

            "auth-key": "*******************************",

            "service-endpoint": ""



The field auth-endpoint is composed based on the following pattern.


Authentication Endpoint


Speech Service (regional)



The field service-endpoint is implicitly composed based on the following pattern.


Service Endpoint


Speech Service (regional)



Where $region is one of westus, eastus, westeurope and others. See Speech Service Regions.

Custom Speech Service

Place the collected key and the authentication endpoint in the following JSON format. Note that the field service-endpoint is supported since Azure SR 1.9.0 and must be explicitly specified when a custom speech endpoint is used.



            "auth-endpoint": "*******************************",

            "auth-key": "*******************************",

            "service-endpoint": "*******************************"


Speech Service Container

Speech Service containers are supported since Azure SR 1.14.0.


Install and run a container per instructions available in this guide. Collect and place the container service endpoint in the following JSON format. Leave the auth endpoint and key empty.



            "auth-endpoint": "",

            "auth-key": "",

            "service-endpoint": "*******************************"



The field service-endpoint is composed based on the following pattern.


Service Endpoint


Speech Service (containerized)



6       Configuring Server and Plugin

6.1      Plugin Factory Configuration

In order to load the Azure SR plugin into the UniMRCP server, open the file unimrcpserver.xml, located in the directory /opt/unimrcp/conf, and add the following entry under the XML element <plugin-factory>. Disable other recognition plugins, if available. The remaining demo plugins might also be disabled, if not installed.


  <!-- Factory of plugins (MRCP engines) -->


      <engine id="Demo-Synth-1" name="demosynth" enable="true"/>

      <engine id="Demo-Recog-1" name="demorecog" enable="false"/>

      <engine id="Demo-Verifier-1" name="demoverifier" enable="true"/>

      <engine id="Recorder-1" name="mrcprecorder" enable="true"/>

      <engine id="Azure-SR-1" name="umsazuresr" enable="true"/>



6.2      RTP Configuration

Since the Azure Speech API supports PCM audio sampled at 16 kHz only, the corresponding codecs needs to be specified in the configuration file unimrcpserver.xml under the XML element <rtp-settings> as follows.


<rtp-settings id="RTP-Settings-1">


   <codecs own-preference="false">PCMU/97/16000 PCMA/98/16000 L16/99/16000 telephone-event/102/16000</codecs>




Note that, starting from Azure SR 1.3.0, the plugin supports resampling of audio data from 8 kHz to 16 kHz. In order to support both 8 and 16 kHz audio in the incoming RTP stream, the corresponding 8 kHz codecs needs to be specified as well.


<rtp-settings id="RTP-Settings-1">


   <codecs own-preference="false"> PCMU PCMA L16/96/8000 telephone-event/101/8000 PCMU/97/16000 PCMA/98/16000 L16/99/16000 telephone-event/102/16000</codecs>




For the basic verification test to work, similar settings should be specified in the client configuration file unimrcpclient.xml as well.


<rtp-settings id="RTP-Settings-1">


   <codecs>PCMU PCMA L16/96/8000 telephone-event/101/8000 PCMU/97/16000 PCMA/98/16000 L16/99/16000 telephone-event/102/16000</codecs>





6.3      Logger Configuration

In order to enable log output from the plugin and set filtering rules, open the configuration file logger.xml, located in the directory /opt/unimrcp/conf, and add the following entry under the element <sources>.


    <source name="AZURESR-PLUGIN" priority="INFO" masking="NONE"/>


6.4      Azure SR Plugin Configuration

The configuration file of the plugin is located in /opt/unimrcp/conf/umsazuresr.xml. Default settings should be sufficient for generic use.


Refer to the Usage Guide for more information.


7       Validating Setup

Validate your setup by using the sample UniMRCP client and server applications on the same host. The default configuration and data files should be sufficient for a basic test.

7.1      Launching Server

Launch the UniMRCP server application.


cd /opt/unimrcp/bin



In the server log output, check whether the plugin is normally loaded.


[INFO]   Load Plugin [Azure-SR-1] [/opt/unimrcp/plugin/umsazuresr.so]


Next, check for the license information.




-product name:    umsazuresr

-product version: 1.0.0

-license owner:   -

-license type:    trial

-issue date:      2017-10-26

-exp date:        2017-11-25

-channel count:   2

-feature set:     0


7.2      Launching Client

Note: the optional package umc-addons must be installed for this test to work.


Launch the sample UniMRCP client application umc.


cd /opt/unimrcp/bin



Run a typical speech recognition scenario by issuing the command run bsr1 from the console of the umc client application.


run bsr1


This command sends a RECOGNIZE request to the server and then starts streaming a sample audio input file whatstheweatherlike.pcm to recognize.


Check for the NLSML results to be returned as expected.


<?xml version="1.0"?>


  <interpretation grammar="builtin:speech/transcribe" confidence="0.95">

    <instance>what's the weather like</instance>

    <input mode="speech">what's the weather like</input>




Visually inspect the log output for any possible warnings or errors.


Note that utterances are stored in the var directory, if the corresponding parameter is enabled in the configuration file umsazuresr.xml and/or requested by the client.