Open-source text-to-speech software
- eSpeak:
a speech synthesiser for Linux and Windows that supports
many languages,
including English, French, German, Spanish and Portuguese.
Licence: GNU GPL 3. - eSpeak NG Text-to-Speech: a ford of the original eSpeak that is hosted on GitHub.
- The Festival Speech Synthesis System:
a general framework for building speech synthesis systems as well as including examples of various modules
. The distribution available through the download page contains the full C++ source code, voice for British and American English, documentation and a few other things.
Licence: X11-type licence. - FreeTTS 1.2:
“a speech synthesizer written entirely in the Java™ programming language”.
It is based on CMU Flite (festival-lite; see below).
The project also has a
Wikipedia article.
Licence: a BSD-style licence. - CMU Flite:
a small, fast run-time text to speech synthesis engine developed at CMU and primarily designed for small embedded machines and/or large servers
. Flite is written in C only. Is supports 13 US English voices, Hindi, Tamil, Telugu and Marathi.
Licence: ? - Ekho is text-to-speech software
for several Chinese languages (Mandarin, Cantonese and a few others).
It is part of the
eGuideDog project.
eGuideDog should work on both GNU/Linux and Microsoft Windows.
Warning: most of the documentation about eGuideDog is in Chinese.
Licence: GNU GPL 2. - dhvani is text-to-speech software
designed for Indian languages. It supports Bengali, Gujarati, Hindi, Kannada,
Malayalam, Marathi, Oriya and a few other ones.
The software currently works on GNU/Linux.
The project is looking for volunteers to port the software to Microsoft Windows.
Licence: GNU GPL 2 or later. - MaryTTS:
a text-to-speech platform written in Java.
It supports German, British and American English, French, Italian and a few other languages.
The project has a repository on GitHub.
Licence: see the MaryTTS Software User Agreement, which states that the code itself is available under GNU GPL 3 and lists the licences of the other libraries that the project uses. - BgTalk:
text-to-speech software for Bulgarian.
The software is still in a beta phase; the
CVS repository
has not seen any updates since 2007.
Licence: GNU GPL 2. - Voce:
a speech synthesis and recognition library that uses CMU Sphinx
and FreeTTS. It is written in Java and C++ and should work on
Microsoft Windows, Mac OS X and GNU/Linux.
The most recent binary dates from 2006.
The latest changes to the source code date from 2013.
You'll need to check the readme file for build instructions.
Licence: BSD 3-clause licence and GNU LGPL 2.0. - Praat: a softwar package
developed for doing phonetics research that also supports speech synthesis.
The source code is in C and c++;
binaries are available for Mac OS X, Microsoft Windows, GNU/Linux and a few
other operating systems.
See also Praat's GitHub repository.
Licence: GNU GPL 2 (1991).
The following projects have not yet reached beta status or have been inactive for at least three years.
- A TTS program
for Japanese, Korean, Chinese and English written in C#,
with no updates since February 2012 (i.e. probably abandoned).
Licence: no licence (in other words, do not reuse the code!). - A TTS program
for Mandarin Chinese written mostly in C++ (with some C and Python),
with no updates since February 2012 (i.e. probably abandoned).
Licence: Creative Commons GNU GPL (i.e. a non-existing licence; do not reuse the code before asking the developer for clarification). - festival-in is
a text-to-speech system for Indian languages based on Festival.
The project is still in a pre-alpha phase; no files are available yet.
Licence: GNU GPL 2. - nextens - open source text-to-speech for Dutch:
this project has been archived.
People who want to continue working on it can
export it to GitHub.
Licence: GNU GPL 3. -
Gnuspeech - Articulatory Speech Synthesis
(Gnuspeech project homepage):
This project started in 2003 and is still not in beta.
It is not clear whether there is any support for languages other than English.
In spite of this, the project has a
Wikipedia article.
Licence: GNU GPL 3 or later.
Free text-to-speech software
Thef following text-to-speech engines and programs are free but not available under an open source licence.
- The MBROLA Project:
a speech synthesizer based on the concatenation of diphones
. The system supports many languages, including French, British Englis, American Englis, German, Spanish, Portuguese and Japanese. The download page gives access to mandatory download (the MBROLA binary and voices) and several optional downloads. The binary is available for many operating systems; the source code is not available for download.
Licence: MBROLA licence, which does not allow that it be used for commercial or military purposes.
Advice, Tips and Tutorials
- Install Eloquence TTS On NVDA [HOW TO] , 8-minute YouTube video (November 2014) that shows how to download Eloquence, install it, and change the NVDA preferences to use Eloquence.