- The added line is THIS COLOR.
- The deleted line is THIS COLOR.
* [[HTS-2.0 README]] [#a4de251a]
#contents
** Release notes [#kb654e7e]
CENTER:&size(20){The HMM-based Speech Synthesis System (HTS) version 2.0};
CENTER:&size(20){release December 29, 2006};
> The HMM-Based Speech Synthesis System (HTS) (http://hts.ics.nitech.ac.jp/) has been being developed by the HTS working group (see "Who we are" below) and others (see "Acknowledgements" in the separate file). The training part of HTS was implemented as a modified version of the HTK (http://htk.eng.cam.ac.uk/). Major modifications which we made to HTK are listed below:
> The HMM-Based Speech Synthesis System (HTS) (http://hts.ics.nitech.ac.jp/) has been being developed by the HTS working group (see "Who we are" below) and others (see "Acknowledgements" in the separate file). The training part of HTS was implemented as a modified version of the Hidden Markov Model Toolkit (HTK) (http://htk.eng.cam.ac.uk/). Major modifications which we made to HTK are listed below:
- Context clustering based on MDL criterion (instead of ML one)
- Stream-dependent context clustering
- Multi-space probability distribution as state output probability (for pitch pattern modeling)
- Multi-space probability distribution as state output probability (for F0 pattern modeling)
- State duration modeling and clustering
> Related publications about the techniques and algorithms used in HTS can be found at http://hts.ics.nitech.ac.jp/publications.html
> The current version does not include any text analyzer but the
Festival Speech Synthesis System (http://www.festvox.org/festival/) can be used as a text analyzer. HTS version 2.0 includes a small run-time synthesis engine (less than 1 M byte including HMMs). Since the synthesis engine can run without the HTK library, it is suitable for using on the Festival.
Festival Speech Synthesis System (http://www.festvox.org/festival/) can be used as a text analyzer. HTS version 2.0 includes a small run-time synthesis engine (less than 1 Mbytes including HMMs). Since the synthesis engine can run without the HTK library, it is suitable for using on the Festival.
> This distribution comes with a demo script using "CMU ARCTIC US English slt" (http://www.festvox.org/cmu_arctic/dbs_slt.html), which generates "voices" for Festival.
> This distribution comes with demo scripts for training speaker-dependent and speaker-adaptation systems using CMU ARCTIC databases (http://www.festvox.org/cmu_arctic/). These demo scripts generate "voices" for Festival.
> Six HTS voices for Festival 1.95 & 1.96 trained by using "CMU ARCTIC database" (http://www.festvox.org/cmu_arctic/) are also released with HTS version 2.0. Each of HTS voices consists of HMMs trained by the demo script, and can be used as a "voice" of Festival Speech Synthesis System without any other HTS tools.
> Six HTS voices for Festival 1.95 & 1.96 trained by using CMU ARCTIC databases are also released with HTS version 2.0. Each of HTS voices consists of HMMs trained by the demo script, and can be used as a "voice" of Festival Speech Synthesis System without any other HTS tools.
> A demo script using the Nitech database for speech synthesis "Nitech Jp ATR503 m001" is also prepared for training Japanese voices. Voices trained by the demo script can be used on GalateaTalk, which is a speech synthesis module of an open-source toolkit for anthropomorphic spoken dialogue agents developed in Galatea project (http://hil.t.u-tokyo.ac.jp/~galatea/), without any other HTS tools.
> An HTS voice for GalateaTalk trained by the demo script is also released with HTS version 2.0.
** What's new in version 2.0 [#n0947bd6]
- Based on [[HTK-3.4>http://htk.eng.cam.ac.uk/download.shtml]]
- Compilation without [[SPTK>http://kt-lab.ics.nitech.ac.jp/~tokuda/SPTK/index.html]]
- Thousands of fixed bugs
- HRest can generate state duration densities (-g option)
- Model boundaries can be given to HERest (-e option)~
We may specify a part of model boundaries (e.g, pause positions)
- Reduced-memory implementation of context clustering in HHEd (-r option)
- Each decision tree can have a name with regular expression (-p option)~
- Based on [[HTK-3.4>http://htk.eng.cam.ac.uk/download.shtml]].
- Compilation without [[SPTK>http://kt-lab.ics.nitech.ac.jp/~tokuda/SPTK/index.html]].
- Thousands of fixed bugs.
- HRest can generate state duration densities (-g option).
- Model boundaries can be given to HERest (-e option).~
We may specify a part of model boundaries (e.g, pause positions).
- Reduced-memory implementation of context clustering in HHEd (-r option).
- Each decision tree can have a name with regular expression (-p option).~
TB 000 {(*-a+*, *-i+*, *-u+*).state[2]}
TB 000 {(*-sil+*, *-pau+*).state[3]}
As a result, different two trees can be constructed for consonants and vowels, respectively.
- The interface of HMGenS has been switched from ''HHEd-style'' to ''HERest-style''
- The interface of HMGenS has been switched from ''HHEd-style'' to ''HERest-style''.
- Flexible model structures in HMGenS (in the previous version, the first stream is assumed as mcep, and the others are assumed as log f0). Non-left-to-right models and full covariance matrices for state output pdfs can also be used.
- EM-based parameter generation algorithm (-c option), i.e, mixture of Gaussians can be used.
-- -c 0: Cholesky decomposition
-- -c 1: EM (with fixed state sequence)
-- -c 2: EM (phone boundaries can be given with -e option)
- Random generation algorithm (set config. variable RNDPG = TRUE)
- Random generation algorithm is also supported (set config. variable RNDPG = TRUE).
- Speaker adaptation, adaptive training, and semi-tied covariance transforms are supported for multi-stream HMMs/MSD-HMMs.
-- MLLRMEAN, MLLRCOV, and CMLLR-based adaptation.
-- CMLLR-based adaptive training.
-- Decision trees for context clustering can be used for definition of regression classes for adaptation.
-- Decision trees for context clustering can be used to define regression classes for adaptation.
-- HMGenS can read MLLRMEAN, MLLRCOV, CMLLR, and SEMIT transforms for adaptation.
- MAP adaptation is also supported.
- Performance improvements in hts_engine
- Performance improvements in hts_engine.
- Miscellaneous changes.
** Copying [#l9acd2a9]
> The basic core system of [[HTS version 2.0:http://hts.ics.nitech.ac.jp/]] is released as a patch code to [[HTK version 3.4:http://htk.eng.cam.ac.uk/]]. The patch code is released under a free license, without commercial restrictions. However, it should be noted that once you apply the patch to the HTK source code, you must obey the license of HTK.
> The basic core system of [[HTS version 2.0:http://hts.ics.nitech.ac.jp/]] is released as a patch code to [[HTK version 3.4:http://htk.eng.cam.ac.uk/]]. The patch code is released under a BSD-style license, without commercial restrictions. However, it should be noted that once you apply the patch to the HTK source code, you must obey the license of HTK.
> Although the patch code is free, we still offer no warranties and no maintenance. We will continue to endeavor to fix bugs and answer queries when can, but are not in a position to guarantee it. We will consider consultancy if desired, please contacts us for details.
> If you are using HTS version 2.0 in commercial environment, even though no license is required, we would be grateful if you let us know as it helps justify ourselves to our various sponsors.
> The current copyright on the core system is
The HMM-Based Speech Synthesis System (HTS)
HTS Working Group
Department of Computer Science
Nagoya Institute of Technology
and
Interdisciplinary Graduate School of Science and Engineering
Tokyo Institute of Technology
Copyright (c) 2001-2006
All Rights Reserved.
Permission is hereby granted, free of charge, to use and
distribute this software in the form of patch code to HTK and
its documentation without restriction, including without
limitation the rights to use, copy, modify, merge, publish,
distribute, sublicense, and/or sell copies of this work, and to
permit persons to whom this work is furnished to do so, subject
to the following conditions:
1. Once you apply the HTS patch to HTK, you must obey the
license of HTK.
2. The code must retain the above copyright notice, this list
of conditions and the following disclaimer.
3. Any modifications must be clearly marked as such.
NAGOYA INSTITUTE OF TECHNOLOGY, TOKYO INSTITUTE OF TECHNOLOGY,
HTS WORKING GROUP, AND THE CONTRIBUTORS TO THIS WORK DISCLAIM
ALL WARRANTIES WITH REGARD TO THIS SOFTWARE, INCLUDING ALL
IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS, IN NO EVENT
SHALL NAGOYA INSTITUTE OF TECHNOLOGY, TOKYO INSTITUTE OF
TECHNOLOGY, HTS WORKING GROUP, NOR THE CONTRIBUTORS BE LIABLE
FOR ANY SPECIAL, INDIRECT OR CONSEQUENTIAL DAMAGES OR ANY
DAMAGES WHATSOEVER RESULTING FROM LOSS OF USE, DATA OR PROFITS,
WHETHER IN AN ACTION OF CONTRACT, NEGLIGENCE OR OTHER TORTUOUS
ACTION, ARISING OUT OF OR IN CONNECTION WITH THE USE OR
PERFORMANCE OF THIS SOFTWARE.
> Several tools in HTS version 2.0 are independent of HTK (though most of them use the HTK library). The copyright of these tools is
The HMM-Based Speech Synthesis System (HTS)
HTS Working Group
Department of Computer Science
Nagoya Institute of Technology
and
Interdisciplinary Graduate School of Science and Engineering
Tokyo Institute of Technology
Copyright (c) 2001-2006
All Rights Reserved.
Permission is hereby granted, free of charge, to use and
distribute this software and its documentation without
restriction, including without limitation the rights to use,
copy, modify, merge, publish, distribute, sublicense, and/or
sell copies of this work, and to permit persons to whom this
work is furnished to do so, subject to the following conditions:
1. The code must retain the above copyright notice, this list
of conditions and the following disclaimer.
2. Any modifications must be clearly marked as such.
2. Redistributions in binary form must reproduce the above
copyright notice, this list of conditions and the following
disclaimer in the documentation and/or other materials
provided with the distribution. Otherwise, one must
contact the HTS working group.
3. Commercial products derived from this software must retain
the above copyright notice, and the following disclaimer.
Otherwise, one must contact the HTS working group.
NAGOYA INSTITUTE OF TECHNOLOGY, TOKYO INSTITUTE OF TECHNOLOGY,
HTS WORKING GROUP, AND THE CONTRIBUTORS TO THIS WORK DISCLAIM
ALL WARRANTIES WITH REGARD TO THIS SOFTWARE, INCLUDING ALL
IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS, IN NO EVENT
SHALL NAGOYA INSTITUTE OF TECHNOLOGY, TOKYO INSTITUTE OF
TECHNOLOGY, HTS WORKING GROUP, NOR THE CONTRIBUTORS BE LIABLE
FOR ANY SPECIAL, INDIRECT OR CONSEQUENTIAL DAMAGES OR ANY
DAMAGES WHATSOEVER RESULTING FROM LOSS OF USE, DATA OR PROFITS,
WHETHER IN AN ACTION OF CONTRACT, NEGLIGENCE OR OTHER TORTUOUS
ACTION, ARISING OUT OF OR IN CONNECTION WITH THE USE OR
PERFORMANCE OF THIS SOFTWARE.
** Installation [#s9493534]
> See the file "INSTALL" in the root directory. Note that HTS
needs HTK.
> Please expand HTS-2.0_for_HTK-3.4.tar.bz2 and see the extracted file "INSTALL". Note that HTS requires HTK.
![[PukiWiki]](skin/logo.png "[PukiWiki]"){kind=logo}
