Festival and setting voice to female

First you need to install festival package

# aptitude install festival festival-freebsoft-utils

In a terminal:

Test festival:
# echo "Hello World"| festival --tts

To copy to wav:
# echo "Hello World" | text2wave -o Hello_World.wav

To read man pages of cron:
#man cron | festival --tts

To read a saved file:
#festival --tts snort-usage.odt

To change voice to a female:

#cd /usr/share/festival/voices/english

#wget http://festvox.org/packed/festival/latest/festvox_cmu_us_slt_arctic_hts.tar.gz
#tar xvf festvox_cmu_us_slt_arctic_hts.tar.gz
#mv festival/lib/voices/us/cmu_us_slt_arctic_hts .

Check if the female voice is loading

#festival

Festival Speech Synthesis System 1.95:beta July 2004
Copyright (C) University of Edinburgh, 1996-2004. All rights reserved.
For details type `(festival_warranty)'
festival> (voice_cmu_us_slt_arctic_hts)    <== Set Voice

 cmu_us_slt_arctic_hts

festival> (SayText "this is a test text for demo") <== Hear the voice
 #<Utterance 0xb7216598>
festival> 

If it works,

PHPAGI for asterisk uses text2wave which is a script based on Festival.

text2wave can be set to the voice using

#text2wave -eval "(voice_cmu_us_slt_arctic_hts)"

so a wrapper script is created in /usr/local/bin/text2wave
#!/bin/bash
/usr/bin/text2wave -eval "(voice_cmu_us_slt_arctic_hts)" "$@"
this script will give sometimes "Segmentation fault"
so a replacement of text2wave is taken from http://trantrietvn.blogspot.com/2011/08/fix-bug-segmentation-fault-text2wave.html
and copied to /usr/bin
(this one is bug-fixed version)

You have a successfully change the voice!

Source: http://www.poespoes.nl/log/installing-festival-and-setting-voice-to-female.html