[Tts-project] Pulseaudio and ALSA support on festival and speech tools

[Sergio Oller]

Dear all,

First of all, sorry for not having contributed more often to maintaining
Festival and Speech Tools. I should have worked a bit more on it in the
past, but I don't have much free time.

Anyway, I recently checked out the latest git commits to speech tools and I
found this:

http://anonscm.debian.org/gitweb/?p=tts/speech-tools.git;a=commit;h=aca5d42fbc5083fa2e70d159b664f2fb9ce54272

Peter says:
Stop building deprecated EsounD support. It has been deprecated by Gnome
for years and reduces our dependencies. I will look to adding Pulseaudio
(not as default in a later patch).

Some years ago I found a native pulseaudio module for festival, and I kept
using it for some time. Here it is the link with the patch:
https://bugzilla.redhat.com/show_bug.cgi?id=471047

In case you (Peter) are planning to work on PulseAudio support for
Festival, this could be a nice starting point.

On the other hand, I will look again to the ALSA audio module soon (see bug
#638394). I don't know much about ALSA internals, but I found the python
module "pyalsaaudio" which implements ALSA bindings for python and it seems
quite simple to read and understand how it works. I will try to "compare"
the pyalsaaudio implementation to the festival implementation in order to
find differences, limitations or problems with the implementation (we
already know there are some "buffer underrun" problems).

Finally, I don't know if the tts mailing list is working or not, so please
forgive me if you receive this email twice.

Best regards,

Sergio
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.alioth.debian.org/pipermail/tts-project/attachments/20121121/b4c0263f/attachment.html>