One 2-hour beam-formed, mixed, downsampled data file. After
beamforming, the time series was multiplied by a 260 Hz sine wave of
amplitude 1 and
then downsampled by grouping (averaging) 128 points. I now realize that
in frequency-shifting I should multiply by a sine wave of amplitude 2 to
conserve the amplitude of the shifted signal. I will do this in the
future.