Re: [orca-list] punctuation pronunciation and pauses
- From: Kenny Hitt <kenny hittsjunk net>
- To: orca-list gnome org
- Subject: Re: [orca-list] punctuation pronunciation and pauses
- Date: Fri, 4 Apr 2008 20:29:46 -0500
Technically, gnome-speech doesn't know anything about audio output. It just tells the TTS engine
to speak something and lets the engine do it however it wants. What this means for Linux users is all
gnome-speech drivers except espeak use OSS. Without a multichannel sound card, using gnome-speech drivers
other audio output. Even espeak can output to OSS instead of ALSA if it is built against portaudio 18.
Tricks like using the aoss wrapper can work with some sound cars to get gnome-speech drivers to share the
sound card, but it doesn't work with all cards.
On my new box that uses the hda-intel driver, aoss doesn't work for swift. Padsp does work, but I have
with pulseaudio and gxine. Because I use gxine to tune local tv, forcing everything to use ALSA is a
requirement. I'm lucky because the Debian espeak package
uses portaudio 19. That lets me use the gnome-speech espeak driver. I will have to use speech-dispatcher to
get access to any other synth on the system.
If your box has a multichannel card like a sblive, the gnome-speech deiver for Cepstral works really well.
Hope this helps.
On Fri, Apr 04, 2008 at 02:55:13PM -0700, Steve Holmes wrote:
-----BEGIN PGP SIGNED MESSAGE-----
I'm a recent commer to this whole situation with Orca and was lead to
believe by others that Speech Dispatcher was supposedly the way of the
future and gnome-speech was not going forward. Maybe I mis-understood
or something. But one thing I did hear was that gnome-speech was
OSS. What if we want to use ALSA and want to take advantage of ALSA's
multi chanel support? I thought OSS was obsolete now since the 2.6
kernel uses ALSA by default now. I'm considering getting gnome-speech
going to see if it works better with Cepstral but if it is restricted
to OSS and single chanel operation, then I'm not interested.
If gnome-speech is the official line and speech dispatcher is *NOT*
"supported", then why doesn't gnome-speech get up to the more current
ALSA implementation? What synths does gnome-speech work with? I'm
using espeak currently.
] [Thread Prev