[Top][All Lists]

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

WIP audio in server

From: Jeremy Whiting
Subject: WIP audio in server
Date: Thu, 25 Feb 2016 15:34:51 -0700

Hello all,

As I expected there are synchronization problems when using 2
different sockets to send and receive audio data in the server from
the output modules. The sockets call a callback when there is new data
on them and having two separate sockets for the metadata vs data bytes
adds a lot of complexity trying to make sure we know what is coming in
on one socket when the other has data. I have however tested the
existing patch a bit more extensively with different output modules
and the ones I tested all work well with it (I tested pico, flite, and
espeak, since those are all available here).

If anyone has other feedback on the patch itself I would appreciate it.


On Sun, Feb 14, 2016 at 7:33 AM, Andrei Kholodnyi
<andrei.kholodnyi at gmail.com> wrote:
> Sorry, it was just a misleading comment from my side at the beginning of
> this discussion.
> What I meant is that audio shall be processed by the cleint part of the
> speech dispatcher, not by the server itself.
> It is a good option to propagate it to the cleints but only if client needs
> it. E..g. it might be a cloud TTS server.
> 2016-02-14 3:48 GMT+01:00 Chris Brannon <chris at the-brannons.com>:
>> Tim Cross <theophilusx at gmail.com> writes:
>> > having played with writing a few speechd clients, I'm very glad I
>> > don't have to worry about how to actually use/play the audio.
>> Plus one, brother.  I'm a lazy programmer.  If I want to make my
>> application talk, I probably don't want to have to handle the audio I/O
>> on my own.
>> > The sockets v pipes question is an interesting one
>> *snip*
>> > From a security perspective, pipes are probably a better
>> > choice.
>> Well, the default these days is to use Unix-domain sockets.  They're
>> just files in the filesystem.  Access control is protected by standard
>> filesystem permissions, which are well-understood.
>> It's about as secure as pipes.
>> You have to jump through some hoops to use actual Internet sockets with
>> a recent Speech Dispatcher.
>> -- Chris
>> _______________________________________________
>> Speechd mailing list
>> Speechd at lists.freebsoft.org
>> http://lists.freebsoft.org/mailman/listinfo/speechd
> --
> Believe is not enough, you must know
> _______________________________________________
> Speechd mailing list
> Speechd at lists.freebsoft.org
> http://lists.freebsoft.org/mailman/listinfo/speechd

reply via email to

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