High quality text to speech GUI

[jacquin.t]

Hello everyone,

I would like to know if some of you have a previous version of ivona english voice who works with sapi4linux ?

Thank you in advance.

[Hydranix]

I could be misunderstanding your posts (both this one and the one which links here: 2111436) but from what I've read is that you are looking for a GUI text-to-speech program.

While searching you were able to get high-quality commandline based TTS to function properly.



Why not just use a script to wrap that commandline TTS with a gui?

I would use YAD for this purpose. http://sourceforge.net/projects/yad-dialog/

#!/bin/sh
# --------------------------------
#  Dirty TTS GUI by Hydranix
# --------------------------------
speak() {
    echo \'${@}\'|festival --tts &>/dev/null
}


QUIT=0


while [ "$QUIT" -eq "0" ]; do
    TEXT="$(yad --width 200 --height 100 --title="Hydranix TTS Wrapper" --center --skip-taskbar \
        --form --dialog-sep --entry \
        --entry-text="Speak this text..." --editable \
        --button="Speak":20 \
        --button="Cancel":1)"
    case $? in
        1) speak "Canceled"; QUIT=1; exit 0;;
        20) speak ${TEXT};;
        255) speak "Canceled"; exit 0;;
        *) speak "Canceled"; exit 0;;
    esac
done

[alexei.colin]

(Sorry, didn't get notifications for this thread for some reason.)

"At last: NeoSpeech Voice (Bridget) installed smoothly and works smoothly."

Could you add some details? I could not find any place to download it...

NeoSpeech voices are for sale, but it appears to be quite hard to find a distributor for it (have to contact company directly). The version that worked for me is a 2010 release of Bridget (UK). I just ran the installer via Wine as described above. The voice was then recognized by sapi4linux, and was selectable in sbreader. I'm still using it successfully, so ask for details as needed.

@alexeicolin2: they say on WineHQ that the missing function has been implemented recently. I will try your advices with Ivona voices when the new version of Wine is out. Correct me if the method is not exactly the one you are suggesting:

1) Install Sapi4Linux as described by frytek
2) Install the latest version of Wine that implements the missing function
3) Set Windows version to Windows 2000 in Wine Configuration (was already set to that for me)
4) Download Ivona voices and then simply run 'WINEPREFIX=~/.winsapi wine ~/my/IvonaVoice/setup.exe' for each voice ?
5) Run sapiconfig -n -s and sapitest with the new voices

Yes. Just to be clear, in Step 4, the argument to wine would be the path to the voice installer, the path and the name might be different from the example above. Please report back once you try the Ivona voices in a wine version that has that function implemented.

[frytek]

Summarizing: to be able to use new Ivona voices, we need wine version >= 1.7.18

To install it we need to add the repo:



sudo add-apt-repository ppa:ubuntu-wine/ppa
sudo apt-get update
sudo apt-get install wine

[dom8]

Hello all, many years have been waiting for TTS on linux
today installed on linux mint 17:
wine beta
Ivona voices
Ivona reader
and it's working!!! well almost
reading rrs,written text
but can't copy text to clipboard monitor or even option paste text not working
any ideas?
option

[dom8]

solved with clipboard manager ,ctrl+v still not working but middle button on mouse do the job
clipboard

[jacquin.t]

Hi everyone,

Well since severals months everything works perfectly, I am able to generate mp3 using wine server sapilektor and everything I had installed following the first post of this thread.

Since today, nothing works. In fact I updated my ubuntu server to 14.04.1 LTS and I have done what I done everytime I restart my server : relaunch install_sapi_server -f

But it doesn't work, at the end of the installtion I try to launch sapiconfig -n or -s and the script doesn't end. When I launch sapilektor I have the error message : No SAPI voices declared

Here the output of the install_sapi_server -f (Everything looks great) :

Changing current directory to /usr/lib/sapi4linux
2 Ivona voices found
No Infovox3 demo voices found
Screen already exists
299:Shutdown OK:
Shutting down wine
wine: created the configuration directory '/home/adraesh/.winesapi'
wine: configuration in '/home/adraesh/.winesapi' has been updated.
wine: cannot find L"C:\\windows\\system32\\winemenubuilder.exe"
Setting Windows version to win2k
Clearing Windows version back to default
Disable sound
Installing ivona_voices/celine.dat
wine: cannot find L"C:\\windows\\system32\\winemenubuilder.exe"
wine: cannot find L"C:\\windows\\system32\\winemenubuilder.exe"
wine: cannot find L"C:\\windows\\system32\\winemenubuilder.exe"
Installing ivona_voices/salli.dat
wine: cannot find L"C:\\windows\\system32\\winemenubuilder.exe"
wine: cannot find L"C:\\windows\\system32\\winemenubuilder.exe"
wine: cannot find L"C:\\windows\\system32\\winemenubuilder.exe"
Ivona voices installed, reinstalling sapi
Setting Windows version to win2k
/usr/bin/install_sapi_server: line 101: 20934 Terminated /bin/bash $0 enter > /dev/null 2> /dev/null
Clearing Windows version back to default
Enable sound
Creating server wrapper command in /home/adraesh/bin/sapi_server.sh
Creating Ivona voice download command in /home/adraesh/bin/ivona_download.sh
Running server

Now the output of a CRTL C on the never end script sapiconfig -n :

^CTraceback (most recent call last):
File "/usr/bin/sapiconfig", line 185, in <module>
getVoices(sock)
File "/usr/bin/sapiconfig", line 111, in getVoices
a=f.readline()
File "/usr/lib/python2.7/socket.py", line 447, in readline
data = self._sock.recv(self._rbufsize)
KeyboardInterrupt

Again, everything works very great before the ubuntu system upgrade.

Thank you in advance.

[frytek]

Hi, jacquin.t,

please edit the

/usr/bin/install_sapi_server

and change the display=:20 to display=:0

(you will need sudo for this)

this is the simplest method to see what the script is actually doing. (you can use vncviewer connected to display :20 if you know how to do it). on my system it appeared that while being run, wine wants first to download some .net and gecko components, and these unexpected dialog boxes make the script freeze. use your mouse to install them - this is needed only once. you will probably need to run the script again, but this time the installation should go as usual.

once you succeed you will probably want to change display to :20 to run the script in the background next time.

i checked one newly-downloaded demo voice (august 2014) and it worked (although it took a considerable time to run and looked as frozen).

so i guess there will be no problems with this system for a while (until some newer versions are released again...).

[jacquin.t]

Hello frytek !

Thank you very much for your answer.

I edited the display value from 20 to 0 and run the install_sapi_server -f script, here the output :

Changing current directory to /usr/lib/sapi4linux
2 Ivona voices found
No Infovox3 demo voices found
Screen already exists
299:Shutdown OK:
Shutting down wine
wine: created the configuration directory '/home/adraesh/.winesapi'
wine: configuration in '/home/adraesh/.winesapi' has been updated.
wine: cannot find L"C:\\windows\\system32\\winemenubuilder.exe"
Setting Windows version to win2k
Clearing Windows version back to default
Disable sound
Installing ivona_voices/celine.dat
wine: cannot find L"C:\\windows\\system32\\winemenubuilder.exe"
wine: cannot find L"C:\\windows\\system32\\winemenubuilder.exe"
wine: cannot find L"C:\\windows\\system32\\winemenubuilder.exe"
Installing ivona_voices/salli.dat
wine: cannot find L"C:\\windows\\system32\\winemenubuilder.exe"
wine: cannot find L"C:\\windows\\system32\\winemenubuilder.exe"
wine: cannot find L"C:\\windows\\system32\\winemenubuilder.exe"
Ivona voices installed, reinstalling sapi
Setting Windows version to win2k
/usr/bin/install_sapi_server: line 101: 9557 Terminated /bin/bash $0 enter > /dev/null 2> /dev/null
Clearing Windows version back to default
Enable sound
Creating server wrapper command in /home/adraesh/bin/sapi_server.sh
Creating Ivona voice download command in /home/adraesh/bin/ivona_download.sh
Running server

But again after that when I run the sapiconfig -n or -s the script always freeze ...

And for sapilektor always : No SAPI voices declared

Thank you again.

[frytek]

we're speaking about GUI here, but i thought i could paste a command line for producing decent mp3 files with sapilektor.

this is basically what i use:

cat "text.txt" | tr "–" "," | sapilektor -a -r -s 1.2 -b 1.0 -p 1.0 -v sapi_voice_name - | \
sox -t raw -e signed-integer -c 1 -r 22050 -b 16 - -t wav - vol 1 bass -30 treble -15 contrast 90 2>/dev/null | \
lame --verbose --preset cbr 56 -r -m m -s 22050 - "filename.mp3"


tr is used because ivona reads "-" out loud as "dash"...

speed, pause and breath are up to you, of course.

as you can see i use sox to remove most of the advertised quality of the voices. it's a matter of personal taste; i just don't like bass because it makes the text less understandable. from the other hand: treble produces hiss that i don't like either. i have volume parameter, too, but the value of 1 changes absolutely nothing. contrast is a nice filter to try. you can skip the treble/bass/contrast/volume filters but you will probably still need sox to convert the sapilektor output to something that lame can work with.

and you can, of course, use some more decent preset than cbr. files made with presets voice and mw-us are quite good for speech.