[Top][All Lists]

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

retrieving synthesized auio data?

From: Tim Cross
Subject: retrieving synthesized auio data?
Date: Fri, 5 Feb 2010 18:28:34 +1100

If all that is needed is to catch audio output into a .wav file (or similar),
it can be done very easily with ALSA using a .asoundrc file. For example, this
is how I had mine configured so that all sound output sent to the audio card
was captured

 pcm.copy {
     type plug
     slave {
         pcm "hw"
 route_policy copy

this will generate a wav file, which you can redirect to something like lame
and generate an mp3 file to save space. By default, it writes to a file in the
/var hierarchy (see the ALSA plugins docs for more details). 

Its been about a year since I last used this technique, so can't garantee ALSA
hasn't changed a bit, so check the docs to be safe. I was using this technique
to generate mp3 files from speechd and emacspeak reading large buffers of
data in emacs. What I liked about it was that it was simple and independent of
either the speech synth being used or the framework i.e. speech-dispatcher,
emacspeak, orca etc. 

The only issue I had was getting recording levels right. Had to fiddle with
alsamixer settngs a bit to get the right levels etc.



Halim Sahin writes:
 > hi Jacob and Luke
 > @Jacob do you need the audio data for further processing?
 > Or do you need only creating  wave files from the synthesized text?
 > Maybe a good start is to add a dummy audio output driver in speechd
 > which writes it's
 > output data into a fifo.
 > This wouldn't need any api work  and could  be implemented (in my
 > opinion) really fast and without much work!
 > @Luke:
 > On Thu, Feb 04, 2010 at 12:04:00PM -0800, Luke Yelavich wrote:
 > > I intend to write up some roadmap/specification documentation as to
 > > what I would like to work on with speech-dispatcher next. I think
 > > first, we get a 0.6.8 release out the door, then start thinking what
 > > needs major work, to ensure speech-dispatcher is still usable both as
 > > a system service for those who want it, and for the ever changing
 > > multi-user desktop environment. 
 > Consider making pulse optional for ubuntu will solve this problem without
 > any new line code.
 > > One such idea I have, is to consider
 > > dbus as a client/server communication transport layer. This could even
 > > go so far as to solve the issue of using system level clients like
 > > BrlTTY with a system level speech-dispatcher, which would then
 > > communicate with a user level speech-dispatcher for example.
 > Luke! It's only an issue because you and other prefer the wrong audio
 > system. i hope one day you start thinking about other stuff to do for
 > speech-dispatcher than the ..... user session integration.
 > The decision to use pulseaudio (only) for ubuntu produced tons of mails from
 > many unhappy users in orca/speechd/ubuntu accessibility mailinglists.
 > Allmost every day some people asking howto use sd as system service etc.
 > BTW.: it works really well this way!
 > Starting paralel process and let them communicate through dbus will add
 > more and more and more overhead to speechd and it's deppendencies.
 > And it will only produce new issues without bringing really new
 > features instead of complexity.
 > Many other audio apps needs to be rewritten to be compatible with this
 > new approach. Thx to PA for this.
 > Just my two cents.
 > Halim
 > PS.: @Luke it doesn't make sense to ignore the user wishes in this area.
 > Read the mailinglists and talk with the people who are not able to use
 > pulse with speechd.
 > Talk also with other a11y projects and speechd users.
 > _______________________________________________
 > Speechd mailing list
 > Speechd at lists.freebsoft.org
 > http://lists.freebsoft.org/mailman/listinfo/speechd

Tim Cross
tcross at rapttech.com.au

There are two types of people in IT - those who do not manage what they 
understand and those who do not understand what they manage.
Tim Cross
tcross at rapttech.com.au

There are two types of people in IT - those who do not manage what they 
understand and those who do not understand what they manage.

reply via email to

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