1 is performed by multiplying every psd component of the spectral set by its associated (unique) frequency f, and then dividing the resulting product by a minuteimum frequency fmin, The reduced frequency component above zero = d.c. of the spectrum is given by f lowest = ( fs / 2 )( 1 – 2 / N) / ( N / 2 – 1 ), where fs is the sample rate (reciprocal of the delta time between samples). The total number of spectral points (both positive and negative frequency components) is N, and the Nyquist (highest) frequency is given by f Nyquist = ( fs / 2 )( 1 – 2 / N). Only positive frequency components are considered, and so the values for the square of the modulus of the FFT used in calculating the psd are each multiplied by a factor dating a latin man tips of two. Starting at f Nyquist (4.99 Hz in the figure) and moving downward, fmin corresponds to the frequency of the first-encountered one-seventh-decade bin that contains only a single point. As the frequency increases above fmin, the number of points per bin increases. There is a single spectral point in the bin containing 0.04 Hz. For the bin containing 0.3 Hz there are 11 points, and their sum yields the value (approx. 1 E-06) corresponding to the red dot just to the left of the peak corresponding to the oscillator’s natural frequency.