The most important member of the TtsEngineSsml class to be implemented is the Speak method.
The Speak method is called by the infrastructures text parser receiving:
A reference to the ITtsEngineSite interface, which provides access to system services such as even queuing and writing audio output.
An array of TextFragment instance produced from Speech Synthesis Markup Language (SSML) input. In addition to text to be rendered as speech, the parsing of the SSML stores information about the requested attributes of the speech in a FragmentState instance associated with each incoming TextFragment object.
A speech synthesizer application can optionally make requests for a specified output format by implementing GetOutputFormat to be called by the platform when it tries to provide the correct audio output.
An implementer can also provide support for managing external definitions of pronunciations, or lexicons, by their implementation of AddLexicon and RemoveLexicon.