RFC: Speech Dispatcher output module using a web browser's TTS capabilities

[Luke Yelavich]

On Wed, Oct 29, 2014 at 12:36:08PM AEDT, Alain Kalker wrote:
> Hi list,
> 
> I'm a bit surprised to have found so little information about this, but I
> think it may be very interesting to try and use a web browser as a speech
> synthesizer.

I would rather try and get proper support for those speech synthesizers made 
available, such that we can write output modules for them in a way that we can 
control what is done with the audio. Having said that, this is certainly an 
interesting consideration.
> 
> Getting this to work with Speech Dispatcher involves writing an output
> module, which in turn will need to communicate with the browser.
> I have a little experience with Node.js, so that would be my first choice
> for implementing this.

It would require a proxy of some sort, that would interface with Speech 
Dispatcher, and then talk to a browser based app via something on the local 
system. I have no idea as to how much access one can get to file descriptors 
from within a sandboxed web app, but I would think it would be very little to 
none. An inet socket via localhost could work, but that also introduces 
security concerns, see Festival and its server mode.

If someone is willing to do the work, I'd be happy to review and include it in 
Speech Dispatcher proper.

Luke