Contents

Using MRCP Tools

Handler developers who want to take advantage of Media Resource Control Protocol (MRCP) can use the existing TTS tools, Play String, Play Text File, Record String, Record Text File, and the extended versions of each of those tools.

Non-Extended Tools

Each of the non-extended TTS tools automatically takes advantage of MRCP if Interaction Administrator indicates the default TTS provider is MRCP. The tools use the default language set for a server and use it to select a compatible TTS voice on the MRCP server to synthesize its text. If no compatible voice is found for a language, most MRCP servers use a default voice.

Extended Tools

The extended versions of the TTS tools can be used to take advantage of more advanced MRCP settings. You can use these tools to control common features such as voice name, volume, and rate. To specify more advanced properties, you can use name:value pairs in the tool step's Optional Parameters field.

The following table lists the name:value properties. When specifying a name:value pair, separate the name and value with a colon (:), for example "property1:value".

Name	Values	Description
MRCP	None	Instructs the server to use MRCP for this particular play. IMPORTANT: This parameter is required for extended versions of TTS tools. It needs to be the first parameter specified in the optional parameter field.
mrcp.audio.voice.gender	The gender of the voice when configuring MRCP TS servers in Interaction Administrator.	Enter one of the following values: Neutral Male Female
mrcp.audio.server.name	MRCP server name as specified in PureConnect.	Uses the specified MRCP server to synthesize the text to speech.
mrcp.audio.voice.name	MRCP voice name as specified in PureConnect.	Uses the specified MRCP voice to synthesize the text to speech.
*mrcp.audio.voice.rate	SSML prosody rate of a voice. The value can be an absolute float multiplier (i.e., 2.0 means twice the rate or 0.5 means half the rate), or the following predefined values which indicate a relative change: x-slow slow medium fast x-fast default
*mrcp.audio.voice.volume	SSML prosody volume. The value can be an absolute number in the range 0-100. (Higher values are louder and zero is silence.) Or, use one of the following predefined values to indicate a relative change: silent x-soft soft medium loud x-loud default
*mrcp.audio.language	Voice language identifiers as specified in RFC 3066 (language-country).
mrcp.audio.vendor	The MRCP server vendor to use.	This restricts the MRCP server selection choices to ones from a particular vendor as specified in Interaction Administrator.
*mrcp.audio.contentType	The content type of the text being synthesized. Possible values: text/plain (default) application/ssml+xml text/uri-list multipart/mixed

* These parameters are vendor-dependent and may vary between TTS engines.

Example

The following optional parameter used in the Play String Extended tool step specifies that MRCP is used to play SSML text.

"MRCP mrcp.audio.contentType:application/ssml+xml"