| 1 |
Hi Deedra, |
| 2 |
I am far from expert, but I will give it a shot. |
| 3 |
|
| 4 |
First, for speakup itself make sure that you have software speech compiled |
| 5 |
in as a module. You will also need to create the /dev/softsynth character |
| 6 |
device (I don't remember the major/minor node numbers offhand), will look |
| 7 |
those up and forward. I presume you have alsa already loaded/configured, if |
| 8 |
not do so. I then emerged flite (the only tts I got to work with |
| 9 |
speechd-up). I then emerged speech-dispatcher and speechd-up. Speechd-up |
| 10 |
at the time required an entry in package.keywords to emerge. I manually |
| 11 |
edited a startup script for flite since it doesn't have one. I then did |
| 12 |
rc-update add speech-dispatcher boot |
| 13 |
rc-update add flite boot |
| 14 |
rc-update add speechd-up default |
| 15 |
|
| 16 |
Use a kernel command line which includes speakup_synth=sftsynth. |
| 17 |
|
| 18 |
Making speech-dispatcher and flite start at boot loads them before |
| 19 |
speechd-up which is the proper sequence, getting speechd-up before |
| 20 |
speech-dispatcher results in much pulling of hair. On my system speech |
| 21 |
starts around the time networking is being loaded. flite is/was the default |
| 22 |
tts for speech-dispatcher and I personally didn't have any luck getting it |
| 23 |
to use festival (which was loaded/configured and worked in gnome just fine). |
| 24 |
However I didn't spend a lot of time on it either. I found a dectalk |
| 25 |
express on ebay which i got and as soon as it arrived and worked I disabled |
| 26 |
software speech, except for festival which is still present for gnopernicus |
| 27 |
to use. |
| 28 |
|
| 29 |
As a side note, for ease of troubleshooting I wrote a script that would flip |
| 30 |
speakup back to the hw synth I had so if speech didn't come up I could just |
| 31 |
log in and type hwsynth to get speech. |
| 32 |
------- cut here ------ |
| 33 |
#!/bin/bash |
| 34 |
echo "synth_keyword" > /proc/speakup/synth_name |
| 35 |
----- cut here ---- |
| 36 |
Replace synth_keyword in the above with the appropriate speakup designation |
| 37 |
for your hw synth. |
| 38 |
**** important ***** |
| 39 |
Also check some of the commands above, some used _ instead of - in places, |
| 40 |
and I don't remember if it was in the command names or the names of the log |
| 41 |
files. The critical things are to create the character device, to start |
| 42 |
speech-dispatcher *before* speechd-up and to start your tts with an init |
| 43 |
script during boot. |
| 44 |
Regards, |
| 45 |
Cecil |
Archives