SpeechRecognitionConnector can be bound to an AudioVideoFlow, and provides a stream of audio data compatible with Microsoft.Speech.
Assembly: Microsoft.Rtc.Collaboration (in Microsoft.Rtc.Collaboration.dll)
One such stream is provided every time Start() is called on the SpeechRecognitionConnector. The stream will contain the audio up to the point at which Stop() is called on the connector. Calling Start() again will then generate a new stream for the new interval, and so on.
SpeechSynthesisConnector should be used only in conjuction with Microsoft.Speech.