[Top][All Lists]

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Pause/resume in Speech Dispatcher

From: Hynek Hanke
Subject: Pause/resume in Speech Dispatcher
Date: Wed, 11 Jun 2008 14:43:20 +0200

Dear James,

the pause/resume was a had small bug in Dispatcher which shows
in your type of input in ReadEtextsActivity.py. It should work
now with latest CVS version of speechd. I've tested both with eSpeak
and Festival.

Please note however that in the current implementation
of Speech Dispatcher, pause/resume only works on sentence
boundaries (or their very rough guesses). So after you hit pause,
it will read until the end of sentence and then pause, resume
goes from this place.

The current implementation is not optimal. What makes it
difficult is that Dispatcher handles multiple clients and while
the user can pause the text in your application, other applications
might want to use the synthesizer and the audio output in
the meantime. Also, at the point when you hit pause, Speech Dispatcher
generally doesn't have the whole text synthesized into audio
yet, so it can't do pause just on audio level. So there is also
a question about detailed synchronization of audio with the text
being read, which is not easy. You get a rough idea that the whole
thing is actually quite complex and this is one of the reason why,
again, the true solution will only come with the new implementation
on TTS API Provider, where certain design decision will make it
somehow easier.

As for now, since you already have quite a detailed controll over the
synchronization (for which you pay with sound quality :( , you might
want to consider to do a CANCEL/SPEAK with the appropriate piece
of text yourself instead of using PAUSE/RESUME as offered by the current
Speech Dispatcher.

With regards,
Hynek Hanke

reply via email to

[Prev in Thread] Current Thread [Next in Thread]