Panotti: A Convolutional Neural Network classifier for multichannel audio waveforms

This code has been publicly available and in continuous development since 2017 but for citation purposes I am creating this Zenodo release June 5 2018, in order to obtain a DOI.  I prefer 2017 as the year for the release; 2018 has not seen any notable change to the underlying structure. 

The method is essentially a simplified version of what I learned from Keunwoo Choi's code.  'Research' credit certainly goes to him, although perhaps my approach is more accessible to new students in machine learning audio applications.  My classifier is not SOTA, it is merely 'simple' enough to be reasonably understood by a teaching professor and his students.  It is a teaching example somewhere on the continuum toward research.   It is, however, quite successful at identifying a variety of sounds, provided that one has enough data (and augments it).