Setting modrana to use voice navigation with mbrola

malfunctioning (Talk | contribs)
(This document outlines how to use modrana with mbrola TTS and voices)
Newer edit →

Revision as of 21:49, 30 October 2014

0. If espeak is not installed in your N900, install it now:

sudo gainroot
apt-get install espeak

1. Download the mbrola ARM binary from http://tcts.fpms.ac.be/synthesis/mbrola/bin/armlinux/mbrola.rar 2. Untar the file, and then copy the mbrola binary to somewhere in your execution path, for example /usr/bin/mbrola 3. As root, change the owner of the mbrola executable to root, and make it executable:

sudo gainroot
chown root /usr/bin/mbrola
chmod 755 /usr/bin/mbrola

4. Download any voices you want from http://tcts.fpms.ac.be/synthesis/mbrola.html For example, for a British Male voice, download http://tcts.fpms.ac.be/synthesis/mbrola/dba/en1/en1-980910.zip 5. Make a directory where to store the voice data files:

mkdir /usr/share/mbrola/

5. Extract the contents of the zip file, and then copy the en1 data file (not the whole en1 directory) to the /usr/share/mbrola directory:

cp en1 /usr/share/mbrola/

6. Test the mbrola voice with espeak:

espeak -v mb-en1 -s 150 "this is a test"

7. If the test succeeds, set up modrana to use this voice:

  • From modrana, click on Options > Sound > Voice
  • Set "Voice parameters" to "manual"
  • Click on "Edit voice string", and enter this string:
espeak -v mb-en1 -s 150 -a %volume% -m %qmessage%
(Where -s 150 means "talk at a speed of 150 words per minute")
  • Now, create a route to test the new voice.