After calling the startup function, an application can call TextToSpeechOpenInMemory. This call blocks until all previously queued text is processed. After the call returns, all text subsequently queued by the TextToSpeechSpeak call is converted to speech and stored in the memory buffers supplied by the TextToSpeechAddBuffer call.
The TextToSpeechCloseInMemory call blocks until the speech from all previously queued text is processed.
When a memory buffer is completed, the buffer is returned to the calling application .