Wednesday, June 16, 2010

Have your IMs spoke to you with good voices

When using Pidgin, and you receive an IM, you will hear it spoken to you, even if Pidgin is buried under many windows or on another workspace. Following these instructions will get you the good sounding voices, the pidgin plugin, the files necessary to use mbrola English voices with festival since the Pidgin plugin uses Festival. The last command just creates a .festivalrc file in your home directory that will fix issues with errors about /dev/dsp and not hearing the text being spoken. I've just done this today after much research, so I've saved you the time. I've been playing with and using Text to Speech for a long time, and I've been in the process of creating my own software. I could use some donations!!! Yes, I do have a working product, but I am expanding it big time and have great plans for it. Enjoy this. Be sure to go to your list of plugins in Pidgin when done and enable pidgin-festival, and choose between the voices that you have installed. us1 is a female voice btw. It will look like this when you go to configure plugin.



mkdir /tmp/mbrola
cd /tmp/mbrola
sudo apt-get install mbrola festival mbrola-en1 mbrola-us1 mbrola-us2 pidgin-festival
wget -c http://tcts.fpms.ac.be/synthesis/mbrola/dba/us1/us1-980512.zip
wget -c http://tcts.fpms.ac.be/synthesis/mbrola/dba/us2/us2-980812.zip
wget -c http://tcts.fpms.ac.be/synthesis/mbrola/dba/us3/us3-990208.zip
wget -c http://www.cstr.ed.ac.uk/downloads/festival/1.95/festvox_us1.tar.gz
wget -c http://www.cstr.ed.ac.uk/downloads/festival/1.95/festvox_us2.tar.gz
wget -c http://www.cstr.ed.ac.uk/downloads/festival/1.95/festvox_us3.tar.gz
for x in *.zip;do unzip $x;done
for x in *.tar;do tar -xf $x;done
sudo mkdir -p /usr/share/festival/voices/english/us1_mbrola/
sudo mkdir -p /usr/share/festival/voices/english/us2_mbrola/
sudo mkdir -p /usr/share/festival/voices/english/us3_mbrola/
sudo mv us1 /usr/share/festival/voices/english/us1_mbrola/
sudo mv us2 /usr/share/festival/voices/english/us2_mbrola/
sudo mv us3 /usr/share/festival/voices/english/us3_mbrola/
sudo mv festival/lib/voices/english/us1_mbrola/* /usr/share/festival/voices/english/us1_mbrola/
sudo mv festival/lib/voices/english/us2_mbrola/* /usr/share/festival/voices/english/us2_mbrola/
sudo mv festival/lib/voices/english/us3_mbrola/* /usr/share/festival/voices/english/us3_mbrola/
printf “;use ALSA\n(Parameter.set ‘Audio_Method ‘Audio_Command)\n(Parameter.set ‘Audio_Command \”aplay -q -c 1 -t raw -f s16 -r \$SR \$FILE\”)\n” | tee -a .festivalrc

When you are done, go lift 300 lbs like I did!!!

3 comments:

  1. Looks very promising. I am working along similar lines for quite awhile now. Hat's off, man!

    ReplyDelete
  2. The tar files are .tar.gz files, and the "z" option also needs to be added to the tar command.

    The four quotation marks in the last line got changed to MS-SmartQuotes, which as usual messes things up.

    ReplyDelete
  3. I've principled done this today after much inquire into, so I've saved you the previously. I've been playing with and using Printed matter to Lingo for a great once upon a time, and I've been in the handle of creating my own software. I could use some donations!!! Yes, I do suffer with a working output, but I am expanding it big continuously and comprise abundant plans for it. Use this.

    ReplyDelete