A text-to-speech engine should support output over the telephone line as well as to a wave-out audio device connected to the computer. Here are a few suggestions:
· Supply at least one text-to-speech mode that supports an 8-kHz sampling rate.
· Support multiple instances of the engine while sending output to different destinations so that text-to-speech can occur over the PC and the telephone line at the same time or several telephone lines can be connected to the same machine.
· Set the flag in the TTSMODEINFO structure that indicates that the mode is designed for the telephone.