Improving the Nyquist Rate for Complex Stationary Process Sampling

Separately using the real and imaginary parts of a complex sampled sequence can improve the recovery of the complex continuous time stationary processes. We look for an exact reconstruction when the Nyquist classical condition is weakened. More precisely, when the sampling interval is unity, we give a necessary and sufficient condition for error-free retrieval of a stationary process with spectral density support on [–2π, 2π], i.e twice the classical support.