IzSynth

Keywords : Text to Speech. TTS. Speak. Voice Synthesizer. Talk. Say. Synth.

izSynth

TTS/Text To Speech synthesizer, background music overlay assembler and audio file converter for PBX and Home Automation Systems

Description

izSynth is a bash script running under Linux, written to automate the synthesis of voices used into izPBX System or for realtime TTS (Text To Speech) used into Home Automation solutions.
It can use offline synthesis software like Loquendo (using the Wine environment), eSpeak, Festival, VoiceRSS, NaturalReaders, and other popular online web TTS services to synthesize audio voices from ASCII text files and automatically merging the audio with background music (mp3 and wav format are supported), adding silences and fade in and out.

For the official site and latest stable version: http://www.initzero.it/products/opensource/izsynth

Follow and contribute the development on GitHub: https://github.com/ugoviti/izsynth

Key Features

• Install and use within 60 seconds
• Synth and download the voice the first time, and reuse it every time without downloading it again (time and bandwidth consuming saver)
• izSynth can use local engine or web based engines (list supported engines with izsynth -L. NB. some engines need an APYKEY to works)
• izSynth support many languages (list supported languages for the given engine with izsynth -l)
• izSynth can detach the process to give responsiveness to the domoticz commands (use -b option to disable/enable)
• Discover all features with izsynth -h

Domoticz Notes

I developed izsynth to be an easy, fast and reliable solution to make domoticz talking.

This is my reference hardware:

• Raspberry Pi 2
• Domoticz version 3.5065
• The RPi2 3.5mm jack analog sound output is connected to a Rokono (B10) speaker (powered via USB cable)

Installation

Install OS Dependencies

Before installing izsynth, install the dependencies for your distribution:

RedHat based distros like CentOS/Fedora

yum install -y epel-release
yum install -y sox lame mplayer curl espeak wine

Debian based distros like Ubuntu or Raspian on RaspberryPi

sudo apt-get update
sudo apt-get install -y sox lame mplayer curl espeak wget bsdmainutils file mawk coreutils

For windows sapi5 support, install the wine package and download sapi2wav.exe:

wget https://gitlab.mister-muffin.de/josch/novel2audio/raw/master/sapi2wav.exe -O /usr/local/bin/sapi2wav.exe

Install izSynth

Development Release (suggested method)

cd /usr/local/bin
wget https://raw.githubusercontent.com/ugoviti/izsynth/master/izsynth -O izsynth && chmod 755 izsynth

Stable Release

Download latest stable version of izSynth package from http://www.initzero.it/products/opensource/izsynth and uncrompress it into a system binary PATH, ex. /usr/local/bin

Link into domoticz script dir (don't forget)

cd /home/pi/domoticz/scripts
ln -s /usr/local/bin/izsynth

Update izSynth

You can update izSynth using the same command you used to install it.

izSynth Usage

Command line

Before using izsynth, we must configure and try it via command line.

Make a quick test with the following command:

izsynth -e naturalreaders -v Peter -t "Welcome home, mr Stark"

NB. Some engines, like voicerss, need an APY KEY before you can use it, follow the onscreen guide to get your APY KEY:

izsynth -e voicerss -H

izSynth is configurable in 3 ways:

1. From command line specifying the options (izsynth -h to list available options or izsynth -E for example usage)
2. Using an external config file for overriding the izsynth variables (suggested method)
3. Modifying the variables on the izsynth script itself (not suggested)

I suggest to create an external config file to avoid changes to your configurations when you update izsynth script.

Create the following config file: $HOME/.config/izsynth/izsynth.conf using the command:

$ izsynth -C

and edit it using your prefered text editor.

NB. if the $HOME/.config/izsynth directory doesn't exist, izsynth will create it on the first run

REMEMBER: the command line options always win against the config file

Now let's try the various izsynth options available. izsynth allow for example to put a background music in every synthesized file, or you can mix various languages and engines in one (MEGAMIX feature)... use the command izsynth -E to discover.

Domoticz Usage

use izsynth inside Domoticz is pretty simple:

• Create a virtual switch (or use a real switch or hardware sensor):
  • in the On Action: put a command like this: script://izsynth device activated
  • in the Off Action: put a command like this: script://izsynth device deactivated

you can use izsynth command line options too:

script://izsynth -e naturalreaders -v Peter -t "Welcome home, mr Stark"

ATTENTION when not using the -w switch, izsynth reuse the already synthesized file (speeding up the audio output feedback), if you want recreate the audio file (for example using other engine and voice), add the -w switch to the command line, or manually remove the /dev/shm/*.mp3 files.

izSynth Tips&Tricks

Make izsynth say the current time without writing complex scripts:

script://izsynth -t "it's $(date +%H:%M) minute" -r

---

If the soundcard output volume is too low:

Use the -W switch to change the output volume of the hardware sound card:

script://izsynth -W 90 -t "Can you hear me now?"

If you found the correct audio volume, I suggest to modify the config file, so you don't need to specify the -W switch every time:

PLAYBACK_VOLUME="90"

---

Of course izsynth can be used into Domoticz LUA scripts or into other chained scripts.

Enhance the izsynth Tips&Tricks usage editing this wiki page.

---

Use an external device (like bluetooth speaker) as playback output:

First configure the bluetooth stack and connect you domoticz box to the bluetooth speaker, using this wiki page: http://www.domoticz.com/wiki/BTAudio

Now test the output from command line:

izsynth -t "I love domoticz" -d alsa:device=bluetooth

or to another audio card:

izsynth -t "I love domoticz" -d alsa:device=hw=2.0

you can list all local usable alsa devices with:

izsynth -D

---

Create voice announcements with background music:

izsynth -e google -v en -m /path/music.mp3 -p 3 -P 2 -F 2 -t "Welcome home, mr Stark"