Emacspeak discussion list

[Dhruv <darbaga10 AT gmail.com>]

Hello,

I'm attempting to add pitch-shift functionality (the way that Emacspeak speaks links etc for example) to Whistler, my addon which bridges between Emacspeak and NVDA. Looking at the way that the espeak tts server works, it seems to tag pitch-shift messages with something like <prosody pitch=93%> and whatnot. Looking at the SSML specification though, this tag is eventually supposed to be closed; but the espeak tts server never closes it. What's happening here? does "d" (dispatch to queue) reset engine state before the next queue of commands gets to run? is something else happening on the espeak side?

(At the moment, I'm just shimming servers/espeak for my own usecase, which means that emacspeak is sending tts data as though it were sending to espeak. I might eventually write my own speech server if needed, but pretending to be espeak is probably okay for now?)