Minimum average configuration for a set of 'n' independent Bcg events

Since the analysis of the Bcg signal in the time domain is conveniently performed on a one-heartbeat block obtained by averaging a number of such blocks extracted from a continuous recording, the problem exists of which is the best relative phasing of the blocks before averaging. Supposing that each block has been expanded in a truncated Fourier series, an algorithm is found that allows one to determine the relative phase shift of the blocks; this is done by imposing the requirement that the integral variance over the period be a minimum. The influence of this procedure is tested on the width of confidence limits of a measure of signal dispersion of two different Bcg beds.