FIR filter amplitude is inconsistent

Created by: alurban

In signal.filter_design.fir_from_transfer(), the overall amplitude scaling of the final filter is off by a factor of Nsize * Ntaps / 2, where Nsize is the length of the input array and Ntaps the requested number of taps. The fix should be very simple, and will impact the amplitude of whitening filters designed in TimeSeries.whiten().

@duncanmmacleod, this should be a 0.12.1 milestone. Sorry for the confusion! I’ll post a PR as soon as I can.