His model of the signal was that its power fell are the square of the frequency. So, ln(1+e^[-x*x]) always came to K*-x^2 , where K represents the clock and noise spectrum. So when sending the log, x was recovered by taking the square rootL:(-K*X^2)^(1/2), and that led to sampling at twice the signal rate to recover the signal and led naturally to a binary counting system. We need to talk cubic signals.
But first, why do signals fall off as f^2 in the engineering world? Mainly because disruptions come from echoes vertical to the line of sight. At high frequencies, the return echo from the surface of a wire, for example, falls on the next cycle, while at low freque3ncies it falls on the current cycle.
No comments:
Post a Comment