Re: Speech dispatcher and audio?

[Jeremy Whiting]

Yes, now most of the modules send their audio back to the daemon so it can do all the audio playback. This will help with sending audio to clients once that's added at some point. but that's not in place yet. This also helps to use one bit of code to do all audio playback in the daemon itself instead of having to have every module do their own playback (usually in the same way using the same code, but still). This should also help (though I'm not sure if it does or not yet) make it so volume control can be per application instead of per synthesis module. I.e. you could turn the Orca speech volume up, while having the Okular or Firefox speech volume lower if you wanted to as a user for some reason. For that to work we may need to have different pulse/pipewire/whatever streams per client in the daemon. I'm not sure if that's in place just yet, but having all the audio playback in the daemon itself makes that possible.

BR,

Jeremy

On Wed, Jul 26, 2023 at 1:47 PM Rastislav Kish <rastislav.kish@protonmail.com> wrote:

Hello,

I noticed sd_espeak-ng, after receiving the speak instruction, doesn't
start speaking on its own, instead it seems it returns audio tag under
code 705 which is afterwards played by speech dispatcher?


Is this some kind of preparation for synthesis to audio functionality?


Previously, I think SSIP used to be a plain-text protocol, the speech
modules were responsible for playing their audio output.


Now, my Chinfusor project seems to be crashing when directly
communicating with other speech modules, because it's unable to utf-8
decode their output after the speak request.


Any hint on what's going on?


Ubuntu Mate 22.04 64-bit, speech-dispatcher 0.11.1-1, espeak-ng 1.50+dfsg-10


Thanks!


Best regards


Rastislav