最後更新: 2015-08-18

asterisk text to speech engine

# Using Google Translate

http://zaf.github.io/asterisk-googletts/

# Festival

http://www.asteriskdocs.org/en/3rd_Edition/asterisk-book-html-chunk/ExternalServices_id291618.html

# espeak

 

目錄

• Asterisk-espeak Installation
• Asterisk-espeak Example
• espeak

 

---

Asterisk-espeak Installation

Installation is a 3-part process:

- installation of Asterisk's development header files,
- installation of the eSpeak TTS engine, and
- installation of the asterisk-espeak module.

Pre-install Package

apt-get install espeak

Manual Install

https://github.com/zaf/Asterisk-eSpeak/archive/master.zip

* Asterisk header files
* libsndfile libraries and header files
* libsamplerate libraries and header files

libsndfile1 - Library for reading/writing audio files

apt-get install libsndfile1 libsndfile1-dev

libresample1 - real-time audio resampling library - shared libraries

apt-get install libresample1 libresample1-dev
apt-get install libsamplerate0 libsamplerate0-dev

Espeak libraries and header files

apt-get install libespeak-dev

/usr/include/libresample.h

make
make install
make samples

Test

asterisk:

boxA*CLI> module load app_espeak.so

boxA*CLI> core show application espeak
 

---

Asterisk-espeak Example

 

extensions.conf

<1>

exten => 1234,n,Espeak("This is a simple espeak test in english.",any)

<2>

;;Read a text file from disk (relative to the channel language)

;;and play it with espeak using the asterisk channel language.

exten => 1234,n,ReadFile(MYTEXT=/path/${LANGUAGE}/myfile,200)

exten => 1234,n,Espeak("${MYTEXY}",any,${LANGUAGE})

 

---

Configure File

 

/etc/asterisk/espeak.conf

[general]
; The cache is never cleared by the application
usecache=yes
; mkdir -p /var/lib/asterisk/espeakcache/
; chmod 777 /var/lib/asterisk/espeakcache/
; file: 7b6e76304f2e9c2c354ca113545a6d81.sln
cachedir=/var/lib/asterisk/espeakcache/

[voice]

voice=en
;speed=130
;volume=100
;wordgap=1
;pitch=40

; 修改完 Setting 後

boxA*CLI> module reload app_espeak.so

 

---

espeak

 

espeak uses speech engine in the libespeak shared library

homepage: http://espeak.sourceforge.net/

Debian Package:

espeak-data - /usr/lib/x86_64-linux-gnu/espeak-data

testing

espeak "This is a test"

espeak -f <text file>

# Writes the speech output to a file in WAV format

保存 file

-w <wave file>

讀其他野

-v <voice filename>

; Sets a Voice for the speech
; -v <voice filename>[+<variant>]
; +m1 +m2 +m3 +m4 +m5 +m6 +m7 for male voices
; +f1 +f2 +f3 +f4 which simulate female voices

espeak -vaf

espeak woman voice

espeak -ven-us+f1 "This is a test" -w test.wav

P.S.

<voice filename> is a file within the espeak-data/voices directory.
# Get the list

espeak --voices

zh-yue   Cantonese Chinese
en       standard default English voice.
en-us    American English.
..........................

 

---

廣東話

 

http://espeak.sourceforge.net/data/zhy_list.zip

Download this file and extract the zhy_list file into eSpeak's dictsource directory.
From within the dictsource directory, do:

  espeak --compile=zh-yue

Full dictionary is not installed for 'zhy'

Can't access (r) file 'zhy_rules'

apt-get install libpulse-dev

wget http://sourceforge.net/projects/e-guidedog/files/eSpeak-Chinese/1.47.11/...

rm -rf espeak-1.47.11-source
tar xJvf espeak-1.47.11-source.tar.xz
cd espeak-1.47.11-source/dictsource

espeak --compile=zhy

Full dictionary is not installed for 'zhy'
Using phonemetable: 'zhy'
Compiling: 'zhy_list'
        33744 entries
Compiling: 'zhy_rules'
        79 rules, 27 groups (0)

測試:

espeak -vzhy 123 -w test.wav