Speech-dispatcher and the -generic modules for Dectalk and Swift voices

[Paul Hunt]

Hi,

Thanks for the tip of commenting out all modules you're not actually 
using. I was having trouble getting speech-dispatcher to work at all 
with the swift module uncommented.

In my case I also had to add the name of my voice to the swift-generic 
config file before it would work.

I now have sd working through pulse with espeak, viavoice and swift. I 
have the output mode in sd set to oss then run speech-dispatcher with 
the padsp command.

Thanks.
Paul

On 06/06/09 15:57, Garry Turkington wrote:
> Hi,
>
> Following up on my own post as I solved my problems and wanted to share in
> case it's helpful to others.
>
> Firstly, never underestimate the power of user error.  I'd screwed up a
> symlink to the Dectalk say executable so speech-dispatcher couldn't find
> it.  The log files in /var/log/speech-dispatcher (one per module) are
> really really helpful.
>
> I also found that reducing the number of speech-dispatcher modules added
> to the absolute minimum needed helped.  I was getting a broken pipe error
> when trying to use Cepstral but it seems to have been caused by a
> different module that I wasn't using.  Uncommenting only the espeak,
> dectalk and cepstral modules meant I could get all working.
>
> The ALSA/OSS problem did arise so installing the aoss tool from the
> alsa-oss package was the answer.  I changed the relevant module conf file
> to use that when calling the executable and again that worked out.
>
> I tried various sound output and for Dectalk at least I was getting minor
> clicks at the end of each utterance when using PulseAudio.  For this synth
> I found the ALSA output with the aoss as described above worked best.
>
> Because the Swift module actually generates a wav file and then sends that
> to the sound system I find the swift module has too much lag for my
> tastes.  Which is a shame as the Cepstral voices are superb.
>
> Hope that's useful to someone,
> Garry
>
>