How to configure Qtspeech - was configuring speech-dispatcher - festival TTS works!!

Gustav Degreef gustav97 at gmail.com
Thu Oct 10 07:23:38 BST 2019 

Hey Jeremy,

That is awesome! Appreciate your efforts on this.  I look forward to
getting the new version.  Regarding svox/pico, there are no opensuse
packages.  After much looking and sorting out the various software
versions/naes I found - *install-picotts.sh*
<https://gist.github.com/jfacorro/734d153463e74fc10adea96e489a433a#file-install-picotts-sh>
which is described as an Installer for SVOX Pico TTS on non-Debian
platforms.  It is at
https://gist.github.com/jfacorro/734d153463e74fc10adea96e489a433a
Do you think this would be a suitable code for an rpm system.  If so can
you help me with instructions on getting the code from github?  I have a
spare laptop which I can try to install it on, and I would ask the
opensuse list for help in getting it installed.   I am very happy I got
Festival to work, the learning experience gives me a lot of confidence
to experiment.  The Festival voice is better in quality, but harder to
understand (among other limitations), so I was already thinkine which
alternative to try next.

I've also thought to post my experience on some kind of blog.  I've
looked at the opensuse forums and the kde accesibility site, but both
seem to have limitations.  Any thoughts where I cold post?  I don't
think I oculd do a HowTo, but some pointers, etc.

Thanks a lot!  Gustav.

On 10/10/19 3:56 AM, Jeremy Whiting wrote:
> Heya Gustav,
>
> Good to hear you got festival working. Another option to try if you
> haven't yet (and if OpenSuse has packages for it) is svox pico.
> Android used to use it before the past couple of releases. It sounds
> pretty natural to most people and has voices available in a few
> languages (English, Spanish, French and German iirc). I think it's
> kind of abandoned at this point from looking at it's git history, but
> it works well if packaged right. Speech-dispatcher has a module for it
> called "pico" you can set as your default module if it's available
> etc. to try.
>
> In other news the okular changes got merged to okular so will be in
> the 19.12 release of KDE Applications once it comes out this december.
> Adds a configuration to select either the flite or speech-dispatcher
> QtSpeech backend in okular's settings dialog and adds an action to
> pause/resume that you can assign a keyboard shortcut to if you need.
>
> thanks,
> Jeremy
>
> On Fri, Oct 4, 2019 at 9:34 AM Gustav Degreef <gustav97 at gmail.com
> <mailto:gustav97 at gmail.com>> wrote:
>
>     Hi Folks,
>
>     I got festival to work!!  I also posted on the opensuse list, and
>     with
>     their help and all the input I got here, I have a new voice.  It
>     is an
>     improvement.  As Simion said, it is only in English, but it is a huge
>     step for me.  I will eventually try to get other voices to work.
>
>     In addition to the below steps, I found out that ESC :wq ENTER is the
>     right command to exit and save the crontab command I had asked help
>     with.  After that I used the spd-conf script repeatedly to figure out
>     what was missing.  In opensuse (Leasp 15) festival needs an
>     additional
>     package to work -   "festival-freebsoft-utils" . I got it as a
>     tarball from
>
>     https://freebsoft.org/festival-freebsoft-utils
>
>     The install instructions are in the readme file and were easy to
>     follow.  Using the spd-conf script then helped me iron out the
>     rest of
>     the details.
>
>     I am very grateful to all of you who gave input in helping me
>     finally solve this.  The sound of the espeak voice never bothered
>     me.  It bothers my partner and recently turned off a very dear
>     friend who first heard it.  It made me realize that it gives a
>     very bad impression of linux.  I am very much encouraged by
>     people's help, and I am going to try to find even bette voices. 
>     In the process of getting festival to work, I saw that the code
>     has essentially not changed in 13 years!  I recently read  an
>     article on the state of the art of linux TTS, it gave me a push to
>     find a voice that sounds more human and gives a better impression
>     to others.  Thanks!  Gustav
>
>
>
>
>     On 9/29/19 3:53 PM, Simion wrote:
>     >
>     > Hi, crontab command will use your default CLI editor, so is probably
>     > opening with vim, vim is notorious for beeing imposible to exit
>     from,
>     >
>     .....
>     >
>     > I can't help with festival
>     >
>     > On 9/29/19 10:18 PM, Gustav Degreef wrote:
>     >> Thanks for the explanation.  I get stuck in the same
>     place.  When I run
>     >> the "crontab -e" command, I get a window where I can past the
>     command
>     >> specified, but then I don't know how to proceed from there. 
>     I read the
>     >> man page for crontab and it says that once the commands are
>     specified,
>     >> you have to exit the editor that is specified (I guess in my
>     system) and
>     >> then the crontab file becomes active.  I have no idea where
>     to proceed
>     >> after I past the command into the window. I tried typing exit,
>        q and
>     >> control q but nothing happens.  The menu of the window says
>     to save the
>     >> output, but when I try that it gives a blank box to type a name
>     to save
>     >> it to.  Any help would be appreciated.  Some years ago, I
>     was also stuck
>     >> in this place.  Is there any simple way to make festival an
>     active
>     >> server?  Gustav.
>     >>
>     >> On 9/26/19 10:32 PM, Jeremy Whiting wrote:
>     >>> Hi Gustav,
>     >>>
>     >>> Just out of curiosity, which distribution and version of linux
>     are you
>     >>> using?
>     >>>
>     >>> To answer your question though, crontab is an application that
>     runs
>     >>> things at different times on a unix based system. crontab -e is a
>     >>> command to edit the cron table of commands. Running crontab -e and
>     >>> adding the line you mentioned:
>     >>>
>     >>> @restart /usr/bin/festival --server
>     >>>
>     >>> will make festival --server run at each boot of the computer I
>     believe.
>     >>>
>     >>> Hope that helps. Another possibility according to arch linux's
>     wiki
>     >>>
>     https://wiki.archlinux.org/index.php/Festival#Can't_open_/dev/dsp is
>     >>> to add some lines to the festival configuration to allow it to
>     open
>     >>> the /dev/dsp
>     >>>
>     >>> BR,
>     >>> Jeremy
>     >>>
>     >>> On Thu, Sep 26, 2019 at 7:25 PM Gustav Degreef
>     <gustav97 at gmail.com <mailto:gustav97 at gmail.com>
>     >>> <mailto:gustav97 at gmail.com <mailto:gustav97 at gmail.com>>> wrote:
>     >>>
>     >>>     Hello folks,
>     >>>
>     >>>     I edited this thread some to make it easier to follow
>     my current post.
>     >>>     It is not directed at Simion, though his comments are
>     helpful to
>     >>>     make my
>     >>>     post clearer. 
>     >>>     As a result of these discussions I have more clearly
>     understood
>     >>>     the role
>     >>>     of speech-dispatcher.  Since will be used with
>     Qtspeech, it makes
>     >>>     sense
>     >>>     to work with it.   So,  Perching with
>     google  I understood how
>     >>>     central
>     >>>     speech-dispatcher is to linux TTS.  I tried
>     again to configure
>     >>>     Festival
>     >>>     with speech-dispatcher.  I want to try and
>     configure a different
>     >>>     voice. 
>     >>>     I wan to try and experiment with different
>     voices.  Espeak is OK,
>     >>>     but I
>     >>>     feel I have to move forward.  But again I am
>     stuck.  Here is
>     >>>     where I am at.
>     >>>
>     >>>     I could never get speech output even on the CLI.
>     >>>
>     >>>     The error I would get was -festival: can't open /dev/dsp
>     >>>
>     >>>     It is due to a bug.   The workaround is:
>     >>>     Create ~/.festivalrc with the following content
>     >>>
>     >>>     ;use ALSA
>     >>>     (Parameter.set 'Audio_Method 'Audio_Command)
>     >>>     (Parameter.set 'Audio_Command "aplay -q -c 1 -t raw -f
>     s16 -r $SR
>     >>>     $FILE")
>     >>>
>     >>>
>     >>>     I then needed to configure speech-dispatcher, so I ran:
>     >>>
>     >>>     spd-conf
>     >>>
>     >>>     as a regular user.  It creates the speech-dispatcher
>     global server
>     >>>     configuration file:
>     >>>
>     >>>     in ~/.config/speech-dispatcher/|speechd.conf
>     >>>
>     >>>     This configures speech-dispatcher (interactively) with
>     a number of
>     >>>     variables, the default voice is espeak.
>     >>>
>     >>>     I then edit speechd.conf, which allows for many different
>     >>>     voices.  I change:
>     >>>
>     >>>     |#AddModule "espeak" "sd_espeak" "espeak.conf"||
>     >>>     |
>     >>>     |AddModule "festival" "sd_festival" "festival.conf"|
>     >>>     |
>     >>>     |
>     >>>     |and|
>     >>>     ||||
>     >>>
>     >>>     |#DefaultModule espeak |
>     >>>
>     >>>     |DefaultModule festival
>     The next step .... "|We  need to run |festival| as a server in
>     order to
>     make
>     >>>     |speech-dispatcher| use it as default. We can do that
>     by adding
>     >>>     the following line at the end of the file that's open
>     when we use the
>     >>>     command: |sudo crontab -e
>     >>>
>     >>>     |@reboot /usr/bin/festival --server
>     >>>
>     >>>     |||I can't understand the last step.  I've
>     never used crontab
>     >>>     fore, any
>     >>>     help would be appreciated, Gustav.|
>     >>>     ||
>     >>>
>     >>> 
>     >>>     On 9/22/19 9:29 AM, Simion wrote:
>     >>>
>     >>>
>     >>>     snip
>     >>>
>     >>>     > As I said I suggest to get used of using faster and
>     faster voices,
>     >>>     > your family will understand and you will read same
>     texts 2 times
>     >>>     > faster so you gain time, at that speed any voice will
>     sound robotic,
>     >>>     > this would not work if you want other to listen with
>     you (but
>     >>>     you can
>     >>>     > have a normal speed and a fast speed talker in jovie)
>     >>>     >
>     >>>     >
>     >>>     > On 9/22/19 2:54 PM, Gustav Degreef wrote:
>     >>>     >>
>     >>>     >> snip
>     >>>     >>
>     >>>     >>  From what you and others are saying and from what I've
>     >>>     read, I think I
>     >>>     >> understand better.  Correct me if I am wrong, but don't
>     >>>     you have
>     >>>     >> to tell
>     >>>     >> speech-dispatcher which voice (e-speak, festival,
>     etc.) to
>     >>>     use?  If I
>     >>>     >> can figure out how to tell speech-dispatcher which
>     voice to use
>     >>>     then
>     >>>     >> Jovie (or Qtspeech) will then output my choice?  I have
>     >>>     looked
>     >>>     >> into it,
>     >>>     >> and there are several (about 6 different) voices I
>     can use in
>     >>>     Linux.
>     >>>     >>
>

More information about the kde-accessibility mailing list