Skip to content

Latest commit

 

History

History
97 lines (83 loc) · 2.8 KB

README.md

File metadata and controls

97 lines (83 loc) · 2.8 KB
Raw

SSP - Speech Signal Processing module

Overview

SSP is a package for doing signal processing in python; the functionality is biassed towards speech signals. Top level programs include a feature extracter for speech recognition, and a vocoder for both coding and speech synthesis. The vocoder is based on linear prediction, but with several experimental excitation models. A continuous pitch extraction algorithm is also provided, built around standard components and a Kalman filter.

There is a "sister" package, libssp, that includes translations of some algorithms in C++. Libssp is built around libube that makes this translation easier.

SSP is released under a BSD licence. See the file COPYING for details.

Installation

To install SSP from git, just clone it into a working directory:

cd ~/src  # Just for example
git clone [email protected]:idiap/ssp.git
cd ssp

Or, if you work in a different directory:

cd where-I-work
ln -s ~/src/ssp/pitch.py
ln -s ~/src/ssp/codec.py

i.e., you shouldn't need to set PYTHONPATH

Then you can say

pitch.py test.wav  # Graphical view of what's going on
codec.py -h
codec.py -r 22050 -p EM1.wav EM1.lf0

If you know what buildout is:

python bootstrap.py
./bin/buildout

Alternatively, SSP is available on PyPI at https://pypi.python.org/pypi/ssp That's available by typing

pip install ssp  # Root level installation

or

pip install ssp --user  # User level in ~/.local

Note that the PyPI package is very likely to be out of date!

Otherwise, see the wiki at https://github.com/idiap/ssp/wiki

Publications

The pitch tracker in SSP is documented in the paper:

@Article{Garner2013,
  author =       "Garner, Philip N. and Cernak, Milos and Motlicek,
                  Petr",
  title =        "A Simple Continuous Pitch Estimation Algorithm",
  journal =      "IEEE Signal Processing Letters",
  year =         2013,
  month =        "January",
  volume =       20,
  number =       1,
  pages =        "102--105",
  doi =          "10.1109/LSP.2012.2231675"
}

and there is a downloadable pdf.

The codec is documented in a technical report:

@TechReport{GarnerTech2015,
  author =      "Garner, Philip N. and Cernak, Milos and Potard, Blaise",
  title =       "A simple continuous excitation model for parametric vocoding",
  institution = "Idiap Research Institute",
  year =        2015,
  type =        "Idiap-RR",
  number =	    "03-2015",
  month =       "January"
}

and again there is a downloadable pdf.

-- Phil Garner June 2013