CMU Sphinx, also called Sphinx in short, is the general term to describe a group of speech recognition systems developed at Carnegie Mellon University. These include a series of speech recognizers (Sphinx 2 – 4) and an acoustic model trainer (SphinxTrain). Here are brief steps for compilation and usage of CMU Sphinx in Debian.
- Download sphinxbase and pocketsphinx from github:
123$ apt-get install git$ git clone https://github.com/cmusphinx/sphinxbase.git$ git clone https://github.com/cmusphinx/pocketsphinx.git - Install prerequisite packages as:
123$ su# apt-get install autoconf libtool automake bison python-dev swig libpulse-dev libasound2-dev# exit - Add write access to /usr/local for current user as:
12$ su# chown -R 'i' /usr/local - Export following paths:
12export LD_LIBRARY_PATH=/usr/local/libexport PKG_CONFIG_PATH=/usr/local/lib/pkgconfig - Change directory to sphinx base and compile it:
1234$ cd sphinxbase$ ./autogen.sh$ make$ make install - In case of any update:
1234$ make clean all$ ./configure$ make$ make install - Then change directory to pocketsphinx and compile it:
1234$ cd pocketsphinx$ ./autogen.sh$ make$ make install - In case of any update:
1234$ make clean all$ ./configure$ make$ make install - To test run below command and wait for ready. Say a word in microphone and it will guess and type that word:
1$ pocketsphinx_continuous -inmic yes