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


and read the help messages.

2) To build and install all tools, type:


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 \
Start running demonstration as follows:

% cd HTS-demo_CMU-ARCTIC-SLT
% make

After composing training data, HMMs are estimated and speech waveforms are synthesized.