The effects of computational complexity on the characteristics of a physical signal that is
reconstructed from its representation of sampled data are analyzed. It is found that a more
complex algorithm does not only require longer time to implement, but also yields an
erroneous reconstruction. The reconstruction suffers from contrast degradation, phase shifts,
and attenuation of details relative to the true signal. These unwanted effects are caused by
the existence of spurious frequencies in the computed spectrum due to rounding‐off errors …