VocAddict / psola

Python package implementing the TD-PSOLA algorithm for speech processing

psola

Implements the TD-PSOLA algorithm.

Requirements

• Numpy
• Matplotlib
• Scipy

Basic Usage

Install from PyPI

pip install psola

or from the source directory

python setup.py install

and use the command line script to interface with the package

psola_exec ...

Project Structure

psola
│
└───psola (source code)
│   │   experiment_config.py (define parameters for a run)
│   │   errors.py (project specific exceptions)
│   │   
│   └───exec (holds executables)
│   │   │   psola_exec (main script to run PSOLA implementation)
│   │   
│   └───pitch (modules for estimating and marking pitch)
│   │   │   estimation.py (estimates pitch for a speech time series)
│   │   │   mark.py (marks pitch periods)
│   │   │   pitch_config.py (holds data and experiemental params)
│   │
│   └───plot (plotting routines)
│   │
│   └───utilities (functions not explicitly part of PSOLA)
│   │   │   low_pass_filter.py (implements a standard LPF)
│   │   │   center_clipping.py (center clipping for spectral whitening)
│   │
│   └───voice (detecting voiced and unvoiced regions)
│   │   │   detection.py (segments voiced and unvoiced regions)
│
└───tests (unit tests)
│   
└───docs (documentation)