The media stream consists of the information exchanged over a call. TAPI by itself provides control only for line and phone devices and does not give access to the content of the media stream. To manage the media stream, an application uses Win32 APIs, such as the Communication, Wave Audio, or Media Control Interface (MCI). For example, an application that provides an interface for managing fax or data transmission uses the TAPI functions to control and monitor the line over which bits are sent, but uses the Communications APIs to transmit the actual data.
In the same manner, the media stream in a speech call (where speech refers exclusively to human speech) is produced and controlled not by TAPI, but by one human talking to another. However, the line on which that call is established and monitored, and the call itself, remain in control of the TAPI application. Voice is considered to be any signal that can travel over a 3.1 kHz-bandwidth channel.