Using LumenVox with Avaya Interactive Response

The LumenVox speech recognition software is compatible with the Avaya Interactive Response platform (for other platforms, including Voice Portal, see Working with Platforms). This is done via the MRCPv1 connection, so before you begin you will need to have the following installed:

• A working Avaya IR platform (we have tested with IR 3.0; the Speech Engine may work with other versions but at this time we cannot say for sure). You must have already enabled speech recognition on this system — you may refer to our more complete Avaya IR/LumenVox Application Notes for more information on this.
• The LumenVox MRCPv1 Server (see our installation instructions)
• The LumenVox Speech Engine and License Server. These may run on the same machine as the MRCPv1 Server or different machines. (See our installation and licensing instructions for Windows or Linux.)

To configure Avaya, you will use the Web-based administration tools and choose the Administration setting under the Speech and DPR Administrator menu choice:

1. Click the Assign New Recognition Type button.
2. Set the Enginetype to mrcp-ibm
3. Fill in the settings as appropriate. Ensure that the IP Address points to the IP address of your MRCPv1 Server.
4. Make sure that the Server Name is the MRCPv1 Server's IP address, with /media/recognizer on the end of it. So if your MRCPv1 Server were on a machine with the IP address of 10.0.0.1, you Server Name would be 10.0.0.1/media/recognizer
5. Once the server is configured, you will need to assign services and appliations to it as you would any other ASR server in Avaya. Again, for detailed instructions on this, refer to the Avaya IR/LumenVox Application Notes.

Once you have configured your Avaya IR machine to point to the MRCPv1 Server, you will need to make two changes to the LumenVox MRCP configuration file:

1. The mrcp_server_ip must be set to the actual IP address, not the loopback. You cannot use 127.0.0.1 for this and must put in the actual IP address of the machine hosting the MRCPv1 Server.
2. The resource_string must be set to media/recognizer and not left at the default recognizer.
3. Optional: if your Speech Engine is running on a different server from the MRCPv1 Server, you must change sre_ip to the IP address of that machine.