[Tts-project] Fwd: speech dispatcher and voxin embedded voices

[D.J.J. Ring, Jr.]

The old address was given by the spd-conf script in Slackware 15.



---------- Forwarded message ---------
From: D.J.J. Ring, Jr. <n1ea at arrl.net>
Date: Sat, Feb 11, 2023 at 10:41 PM
Subject: speech dispatcher and voxin embedded voices
To: <speechd at bugs.freebsoft.org>, Gilles Casse <contact at oralux.org>,
Samuel Thibault <samuel.thibault at ens-lyon.org>


I tried to install voxin embedded voices in Slint / Slackware 15.

spd-say "hello" says very very quickly "hello".

voxin-say "hello" | aplay the voxin tool says "hello" very clearly.

I know spd-conf is out of date, so according to Internet advice I
changed some of the default answers.

Instead of accepting the default module of "Default output module
[espeak-ng] :" I entered "voxin" and when the spd-script complained
about this and told me it was not one of the choices, I chose NOT to
correct it.

SLINT has the option of logging into the console before Xorg starts.
This is what I use but in order to have speech at log in, the user
needs to run spd-conf and configure a system wide spdhd conf.

NOTEwhen running as root, this creates a file in userspace that's
owned by root:root.

-rw-r--r-- 1 root root 14K Oct 30 18:29
/home/djringjr/.config/speech-dispatcher/speechd.conf

The installation seems to give errors but voxin works perfectly.

I'm preparing a HOW TO for the slint.fr email list.  I've posted what
I've written so far below my signature.  It's a draft, I haven't
checked it completely.  There is one small typo in the voxin installer
script about language codes. I am going to message Gilles separately
about this.

This is the voxin script error - it's a tiny one letter error.

Default language (two/three-letter iso language code like "en" or
"cs", possibly with a region qualification.) [en-US] :

Unfortunately "cs" is an error in the script, there is NO language
with cs as a code, they probably meant "es" for Spanish.

If you have a Spanish Latin American Voxin voice, you would type
."es-la" without the quotes. If you're an American like myself and my
language is an American speaking English, you enter:

en-US

Best regards.

David

I just installed zoe-ml-embedded-high,en,US which I bought last year
from oralux.org.

I boot into console so there is one more step needed to install a
Voxin Embedded voice than just running the installer.

First find your purchase receipt from Oralux.org.  Use the link to
download the latest version of your voice.

This is the file name for my Zoe ML High voice.
voxin-american-english-zoe-ml-high-3.3rc6.tgz

I untar it.

tar xvf voxin-american-english-zoe-ml-high-3.3rc6.tgz

Unarchiving this file creates a folder:

voxin-3.3rc6

Enter that folder.and you will find (for example) another folder:
voxin-american-english-zoe-ml-high-3.3rc6

Enter tat folder and you will see:along with other files and folders:

voxin-installer.sh

Run this as root.  If you are your user account, just use "su -m" to
stay in the same folder and become root.

Run "voxin-installer.sh" by running this command:

./voxin-installer.sh

./ tells the command line that the command is in the same folder as
you are in and not in the $PATH of the computer.

When you do this you will see on the screen (or hear).

Do you want to install a new language?
If yes, press the ENTER key.

Do you agree the licenses?
If yes, press the ENTER key.

The response to both questions is to press the ENTER key.  I have
noticed that the second dialog asking me to press the ENTER key often
needs me to press the ENTER key again so it registers.

In a second or so you will hear your Voxin Eloquence voice speaking.

First let's make sure we have speech-dispatcher running as a daemon we
enter this command as root without quotation marks and making sure we
use a capital D and not a lower case d. "speech-dispatcher -D":

speech-dispatcher -D

The script will respond:
shome/djringjr/.config/speech-dispatcher/speechd.confpeechd: Speech
Dispatcher 0.11.4 starting


If you just want to use this voice in ORCA you're all finished except
for selecting that voice in ORCA preferences but if you want to use
this voice on the command line and especially if you log into SLINT in
console mode - you need to configure Speech 'Dispatcher for this.  Pay
attention because the installation is a bit "special" because spd-conf
is not up to date.

You should still be root.

Run this command:

spd-conf

That script will respond:

Speech Dispatcher configuration tool

Do you want to setup a completely new configuration? [yes] :

So you press ENTER to accept the default [yes] answer.

If you log into SLINT using console log in, you need to make both a
user and a system speech dispatcher configuration and happily the
answers are the same.

The script will respond:
Do you want to create/setup a 'user' or 'system' configuration? [user] :

This time we are going to change the default answer to "system" so
instead of pressing ENTER we type in "system" without quotation marks.

system

Don't be alarmed, the script will tell you that you have to be root
even if you are already root.

Warning: You must be root or under sudo to do this.
Configuring system settings for Speech Dispatcher
Default output module [espeak-ng] :

Here we have to make another change, instead of just pressing ENTER to
accept espeak-ng we have to type in without quotes "voxin" in all
lower case.

voxin

The script won't like this answer but it is the correct answer.  it
will print out:

The value you have chosen is not among the suggested values.
You have chosen 'voxin'.
The suggested values are ['espeak-ng', 'espeak', 'festival', 'flite',
'ivona', 'pico', 'espeak-ng-mbrola-generic', 'espeak-mbrola-generic',
'swift-generic', 'epos-generic', 'dtk-generic', 'ibmtts', 'cicero',
'kali', 'mary-generic', 'baratinoo']
Do you want to correct your answer? [yes] :

The script is asking us if we want to correct this answer, and we must
type in without quotation marks "no" to NOT change the answer.

no

The script prints:

Default language (two/three-letter iso language code like "en" or
"cs", possibly with a region qualification.) [en-US] :

Unfortunately "cs" is an error in the script, there is NO language
with cs as a code, they probably meant "es" for Spanish.

If you have a Spanish Latin American Voxin voice, you would type
."es-la" without the quotes. If you're an American like myself and my
language is an American speaking English, you enter:

en-US

Now the script prompts you:

Default audio output method [pulse] :

Instead of accepting pulse, type in "libao" without the quotation marks.

libao

The script will accept this value without complaint and prompt you with:

Default speech rate (on the scale of -100..100, 0 is default, 50 is
faster, -50 is slower) [0] :

Just accept the defaults for this and the next few questions about the
speech rate and pitch.

I pressed ENTER and the script prompts:

Default speech pitch (on the scale of -100..100, 0 is default, 50 is
higher, -50 is lower) [0] :

I press ENTER again and the script prompts:

Default speech pitch range (on the scale of -100..100, 0 is default,
50 is higher, -50 is lower) [0] :

I press ENTER again and the script prompts:

Do you want to start/restart Speech Dispatcher now and run some tests? [yes] :

I press ENTER again and the script prompts:

Warning: You must be root or under sudo to do this.
Starting Speech Dispatcher in system-mode
Is your system using an /etc/init.d/speech-dispatcher script? [yes] :

I press ENTER again and the script prompts:
-------------- next part --------------
A non-text attachment was scrubbed...
Name: speechd-debug-11-2-2023.tar.gz
Type: application/gzip
Size: 18838 bytes
Desc: not available
URL: <http://alioth-lists.debian.net/pipermail/tts-project/attachments/20230211/d6f544d4/attachment-0001.gz>