It would be good to have some kind off feedback when the system is working (typically between sending a request to the TTS-server and receiving the response).
One of the most straight forward ways of doing this would be to introduce an animation for the play button.
The button animation can be done using FontAwesome (https://jsfiddle.net/Beppe/1tz4238k/), but may need some changing of how it's used, e.g. adding an element (like in the FA examples) rather than using :after.