Hidden Markov Model Based Text To Speech
HTK 3.4 source code:
http://htk.eng.cam.ac.uk/download.shtml∞
HDecode source code:
http://htk.eng.cam.ac.uk/extensions/index.shtml∞
After downloading HTK-3.4.tar.gz and HDecode-3.4.tar.gz, expand them
% tar -zxvf HTK-3.4.tar.gz
% tar -zxvf HDecode-3.4.tar.gz
Both of these will be extracted into the same htk folder
And then apply HTS-2.0_for_HTK-3.4.patch
% cd htk
% patch -p1 -d . < HTS-2.0_for_HTK-3.4.patch (path to the patch from the htk folder)
HTK
./configure
make
make install
HDecode
5. make hdecode
6. make install-hdecode
Installing SPTK (speech signal processing toolkit)
SPTK source:
http://downloads.sourceforge.net/sp-tk/SPTK-3.1.tar.gz∞
0) Enter 'src' directory.
1) Type
% ./configure --help
and read the help messages.
2) To build and install all tools, type:
% ./configure --prefix=/usr/local/SPTK
% make
% make install
Other requirements
- speech tools
- festival
- tcl8.4
- tcl8.4-dev
- sox, normally available in linux.
- tcl-tk supporting snack
- perl, normally available in linux.
Installing HTS-demo_CMU-ARCTIC-SLT (Demo for extracting speech parameters and training HMMs)
Source: HTS-demo_CMU-ARCTIC-SLT.tar.bz2 for HTS-2.0.1
http://hts.sp.nitech.ac.jp/archives/2.0.1/HTS-demo_CMU-ARCTIC-SLT.tar.bz2∞
Setup HTS-demo_CMU-ARCTIC-SLT by running configure script:
% cd HTS-demo_CMU-ARCTIC-SLT
% ./configure --with-tcl-search-path=/usr/local/
ActiveTcl/bin \
--with-fest-search-path=/usr/local/festival/examples \
--with-sptk-search-path=/usr/local/SPTK/bin \
--with-hts-search-path=/usr/local/HTS-2.0_for_HTK-3.4/bin
Start running demonstration as follows:
% cd HTS-demo_CMU-ARCTIC-SLT
% make
After composing training data, HMMs are estimated and speech waveforms are synthesized.