Text To Speech

For text to speech we use the library MaryTTS.

It produces AudioInputStreams, whose bytes we copy over into a SourceDataLine from the java AudioSystem. This is important, because like that we don't need to close the SourceDataLine, which doesn't work under most linux systems.

The text to speech class currently has an internal queue of things to say, and pauses for 1 second after each output. While this pause the function isCurrentlyOutputting() still returns true. That is necessary, so Amy doesn't hear herself.

Before a string is passed to MaryTTS, some preprocessing is done. Like substituting °C and °F with their spoken form.