Home

Introduction

SSP began after finding this document about numpy, scipy and matplotlib. Matplotlib in particular makes it easy to visualise what is going on inside signal processing algorithms. SSP is also an answer to tracter's being a little too tricky for experimentation with algorithms. Whilst tracter requires algorithms to be written frame-synchronously (real time ready), SSP assumes that you just want to read in the whole file, run an FFT over all frames, etc. It uses more memory than tracter, but it's quicker for evaluating new algorithms.

Tracter does influence the naming: Whilst I want to follow the python style guide, SSP uses UpperCamelCase for functions that would be component classes in tracter.

SSP is closely related to Bob being a Bob satellite package. However, SSP doesn't (yet) need to call out to C++ code, so it is also just a basic python module.