Using VXML to Implement a Language Not Available in Cisco Unified CCX
MRCP (Media Resource Control Protocol) is an application-level protocol that enables client devices requiring audio/video stream processing to control media service resources like Speech Synthesizers (TTS), Speech Recognizers (ASR), Signal Generators, Signal Detectors, Fax Servers, and so on over a network.
To implement an MRCP ASR and TTS-enabled script for a language outside the set available with Cisco Unified CCX (but within the set available from an MRCP vendor) using VXML, you must do the following:
Procedure
Step 1 | Install and configure the appropriate MRCP ASR and/or TTS language pack(s). See your MRCP vendor documentation for ASR/TTS language pack installation/configuration instructions. | ||
Step 2 | If your VXML script uses prompts, then you need to record the prompts in a suitable format G711 μ-Law or A-Law codec and store the prompts on a server that is accessible to the VXML script. Only G711 prompts are supported by MRCP.
| ||
Step 3 | Provide a VXML script, referencing the recorded prompts (if used) and using any necessary built-in grammars provided by the vendor or script-writer provided grammars which are specified either in the VXML script or a location specified by the "src" attribute of the grammar element. | ||
Step 4 | Provide localized prompts for default event handlers and system-level errors. Cisco Unified CCX does not completely implement the notion of platform-specific audio, as defined by World Wide Web Consortium (W3C) (http://www.w3.org/), since system prompts are played instead. A default script is provided with Cisco Unified CCX which you can associate with the script's trigger to localize default event handlers (See When Do You Need a Language Group?, page 4-2). |