Performance Enhancements

[riggsd]

Look into ensuring that all signal-processing code is vectorized as native numpy C-code, or look into other alternatives like:

http://berkeleyanalytics.com/bottleneck/

https://github.com/pydata/numexpr

https://cython.readthedocs.io/en/latest/src/tutorial/numpy.html

[riggsd]

Replacing our already performance-tuned mean calculation in calculate_amplitudes() with bottleneck.nanmean() gives a 25% speed increase.