IVR Development Toolkit (IVR DT)
The Enghouse Interactive IVR Development Toolkit (IVR DT) builds telephony and IVR applications that rely on an external speech engine for recognition and synthesis. Capacity Private Cloud is that engine, connecting to IVR DT over MRCPv2 (SIP). This guide has been tested with IVR DT version 12; for product details, visit enghouseinteractive.com.
Before configuring your IVR DT system to point at the platform's speech engine, you should have already installed, licensed, and tested your ASR (SRE) Server, TTS Server, Media Server, and License Server. This guide assumes some familiarity with Enghouse Interactive IVR DT configuration and that IVR DT is already installed and licensed.
IVR DT connects to the platform using the MRCPv2 (SIP) protocol. Because both applications use SIP, their default configurations assume port 5060. If you intend to run IVR DT and the platform on the same server, you must change the platform's SIP port to a different value to avoid a port conflict.
Configure the Voice Processing Card
First, ensure the Voice Processing Card's firmware setting points to a valid firmware file that supports Continuous Speech Processing. The example shown here uses a Dialogic D/4PCI-U board; your configuration may vary (consult Enghouse for assistance if needed).
For Dialogic boards, open the Dialogic Configuration Manager Utility and select the Device tab to configure the board. In the example below, the Misc tab of the Configure Device option shows the FirmwareFile setting using d4ucsp.fwl, which supports Continuous Speech Processing.
The default firmware file usually does not support continuous speech recognition, so you will likely need to change it. Highlight the FirmwareFile parameter, select the one with csp in its name, and click OK.
Note the SIP_PORT and MRCP_SERVER_IP
Make a note of the SIP_PORT parameter defined in the Media Server configuration, along with the IP address of the Media Server (shown as MRCP_SERVER_IP). You will enter these in the IVRDT.ini file in the next step.
Important: Never leave MRCP_SERVER_IP set to the loopback address (127.0.0.1). It must always be the (internal) static IP address of the speech server.
Configure the IVRDT.ini File
Modify the IVRDT.ini configuration file, located in the root of the Windows directory, so IVR DT knows how and where to communicate with the speech server.
Using the Media Server's IP address and the SIP_PORT noted above, add them to the [MRCPv2Servers] section of IVRDT.ini:
In this example, the platform's internal IP address is 192.168.0.21 and the port matches the Media Server's SIP_PORT (the default is 5060), which is used to negotiate ASR and TTS resources. These must match your own server's settings, so they may differ from the example values (though 5060 is the typical SIP port).
Exclude Non-MRCP Protocols
Disable any non-MRCP voice recognition protocols in the [ProfileExcludeTechnologies] section of IVRDT.ini:
Testing
Following these steps allows IVR DT to fully integrate with the platform's speech resources over MRCPv2. We recommend performing some basic testing to confirm all functionality is working correctly.