Toolboxes for modeling auditory perception have a surprisingly long history, starting with the Auditory Toolbox, first written by Malcom Slaney for Mathematica, in 1993, and then ported to Matlab in 1998. Here we present the Python Auditory Modeling Toolbox (PAMBOX), an open-source Python package for auditory modeling. The goal of the toolbox is to provide a collection of components that can be easily combined and extended to solve auditory modeling problems.
PAMBOX contains code for modeling cochlear filtering, envelope extraction, as well as modulation processing. The toolbox also includes speech intelligibility models. These models are commonly used to predict how well speech is understood in a given situation, such as in the presence of noise or reverberation. The intelligibility models use a simple and consistent "predict" API, inspired by scikit-learn's "fit and predict" API. This simplifies comparisons across models. PAMBOX also includes a framework for performing intelligibility experiments compatible with IPython.parallel.
Models that are not original to PAMBOX are validated against their original implementations, where available. PAMBOX is based on NumPy, SciPy, and Pandas. It is distributed under the Modified BSD License.