There are a few problems with non-local voices (at least in my testing as they are implemented in Chrome at the moment):
- they don’t generate
boundaryevents, needed for word highlighting - they don’t generate
endevents when the TTS has finished speaking, which is required to switch the play/pause toggle button back to play mode or to turn to the next page - for long passages, the TTS just cuts off mid sentence
I’ve just tried again to see if I could switch on these non-local voices for a more basic TTS experience without word highlighting, but the browser support isn’t good enough at the moment so it feels too buggy. Sorry.