HOWTO: Make festival TTS use better voices (MBROLA / CMU / HTS)

[N4RPS]

Is there any way the how-to can be updated to cover how to successfully pull this off in 11.10 Oneric Ocelot?

[gefthebest]

Subscribing too (:

Ubuntu 11.10
Festival 2.1

hts_voice_cmu_us_arctic_slt-1.04

[go_beep_yourself]

Give JSpeak a try. IMHO, it's the best TTS software for Linux =D

http://linuxinnovations.blogspot.com...to-speach.html

[stactrac]

I succeeded in getting festival to work with the HTS voices on Ubuntu 11.04. I simply installed the festival package from lucid [1] and followed the procedure for installing the HTS 2.1 voices. You need one additional dependency from lucid, libestools1.2. Simply download these two packages and install the .debs manually with dpkg -i.

[1] https://launchpad.net/ubuntu/+source/festival/1.96~beta-10ubuntu1/+build/1335123

On 12.04 I used the old setup I did with the HTS & then downgraded Festival manually by installing these in this order. Works fine!

libaudiofile0_0.2.6-8ubuntu1_i386.deb (80.1 KiB)
festival_1.96~beta-10ubuntu1_i386.deb (865.1 KiB)

Figured this might help someone.

Stactrac : http://societystacktrace.com

[needhelpwithfestival]

Hello People!

I have a really urgent question!
But first, the Background of the story.
At the aprenticeship school we have to do IT projects and write adocumentation about what we did.

My Project-Part is Asterisk and it runs very well.
But it looks a little scant in the documentation although I have 26 instead of the requested 15 Pages.
So I decided to make a TTS with festival.
A script will detect if there is any mail for the "admini" user (Mail Program is Postfix), wich come from the nagios.
So the basic order of events is:
checkmail.sh -> notification_text.sh -> originate.sh
checks mail makes TTS with text2wave originates the phonecall
(i know there is a easier way, but i choose this one^^)

It already ran on Tuesday, but unfortunately I decided to change the voice. I installed a voice completly wrong (wrong directory, etc.), but that was not the problem because the TTS still worked with the standard voice.

But then I used a line in the tutorial that sets the installed voice to default voice.. I can't remember the exact line, but it writes the lines into the /etc/festival.scm
(! set somethingsomethingsomething...

Since then not even the default voice worked anymore.

I purged the festival installation and deleted the festival.scm in /etc, restarted the server but nothing worked to get my TTS done.

I reeinstalled it, but still no TTS

Could you help me possibly?

Here the codes from the scripts.
I can't access the server on this weekend, 'cause the serve is at school and the IP is not NATed


mailcheck.sh

echo mail -u admini
if [ "No mail for admini" ]; then
        echo "Alles okay, keine Mails fuer den Admin!" | wall
else
        echo `/etc/asterisk/nagios2voip/notification_text.sh`
fi

notification_text.sh

echo `mail -u admini < /etc/asterisk/nagios2voip/mailscript | grep -A6 Nagios` | text2wave -scale 50  -o /etc/asterisk/nagios2voip/test.wav
sox test.wav -r 8000 -c 1 test.gsm
mv /etc/asterisk/nagios2voip/test.wav /var/lib/asterisk/sounds/en/test.gsm
/etc/asterisk/nagios2voip/originate.sh

originate.sh

echo calling 300! | wall
#asterisk -rvx "originate sip/200 extension 103"
asterisk -rvx "originate sip/300 extension 123"

I don't think that it's the scripts fault, 'cuz it worked earlier...


Thanks!

[ramreddy502]

hello,
i am new to tts. i have the problem with the building of speech tools. unless until festival can not build with out building speech-tools. while building speech-tools i got the following error

../../include/../base_class/EST_TSimpleMatrix.cc:54:5: error: ‘memcpy’ was not declared in this scope, and no declarations were found by argument-dependent lookup at the point of instantiation [-fpermissive]
In file included from /usr/include/string.h:642:0,
from /usr/include/c++/4.7/cstring:44,
from ../../include/EST_String.h:42,
from ../../include/EST_error.h:182,
from ../../include/../base_class/EST_TMatrix.cc:48,
from matrix_i_t.cc:50:
/usr/include/i386-linux-gnu/bits/string3.h:49:1: note: ‘void* memcpy(void*, const void*, size_t)’ declared here, later in the translation unit
make[2]: *** [matrix_i_t.o] Error 1
make[1]: *** [inst_tmpl] Error 2
make: *** [base_class] Error 2


please help me how should i solve this problem

thanks in advance

[ramreddy502]

hi
i am new to edenburgh festival

i have done the building of festival in ubuntu system. in what way i should check weather it is installed correctly or not.

in the same way is also built festvox in ubuntu system. but while checking i got
"oss audio error: failed to open audio device dev/dsp" as an error

please can anyone help me how to solve this output audio error.

thanks in advance

[Ian Clark]

these instructions are really old. i'm running xubuntu 12.04 and would like to know the simplest way to install these better voices for festival. thanks!

[drysdalepete]

hello,
i am new to tts. i have the problem with the building of speech tools. unless until festival can not build with out building speech-tools. while building speech-tools i got the following error

../../include/../base_class/EST_TSimpleMatrix.cc:54:5: error: ‘memcpy’ was not declared in this scope, and no declarations were found by argument-dependent lookup at the point of instantiation [-fpermissive]
In file included from /usr/include/string.h:642:0,
from /usr/include/c++/4.7/cstring:44,
from ../../include/EST_String.h:42,
from ../../include/EST_error.h:182,
from ../../include/../base_class/EST_TMatrix.cc:48,
from matrix_i_t.cc:50:
/usr/include/i386-linux-gnu/bits/string3.h:49:1: note: ‘void* memcpy(void*, const void*, size_t)’ declared here, later in the translation unit
make[2]: *** [matrix_i_t.o] Error 1
make[1]: *** [inst_tmpl] Error 2
make: *** [base_class] Error 2


please help me how should i solve this problem

thanks in advance

Dear fellow tts user,

Judging from your error message you are compiling speech-tools on >=gcc 4.7
or you are compiling on >=clang 3.0 from upstream rather than the lastest
ubuntu raring patched version.

The upstream (CMU/edinborough) version of festival 2.1 needs to be modified
to compile (it needs stricter c++ compliance).
If you look in the raring code you will see a patch unqualifiedmethods.diff in debian/patches. You will need to apply this to your upstream version to compile with
gcc 4.7 or greater.

For compiling with clang 3.0, your upstream will need unqualifiedmethods.diff and also need templatefixes.diff applied which can be found in Debian master branch for speech tools
http://anonscm.debian.org/gitweb/?p=...eech-tools.git.
This patch has not yet been published in Debian or Ubuntu since so few users compile with clang.

best regards
Peter
(a Festival uploader for Debian)

[drysdalepete]

hi
i am new to edenburgh festival

i have done the building of festival in ubuntu system. in what way i should check weather it is installed correctly or not.

in the same way is also built festvox in ubuntu system. but while checking i got
"oss audio error: failed to open audio device dev/dsp" as an error

please can anyone help me how to solve this output audio error.

thanks in advance

Dear fellow tts user,

This also looks like an upstream version of festival. Have a look in
debian/festival.scm in raring for an example of how sound output is set. Note
we modify the build system to build our sound support in Debian and Ubuntu.

Have you considered building and using the raring version as your starting point.
It removes some of these wrinkles and has hts21 compatibility (if you want
nitech voices which seem surprisingly popular in this forum N.B.- alter the nitech voices as described in the Debian README which comes with the raring version.)

best regards,
Peter
(a Festival uploader for Debian)