Performance of software-defined and cognitive radios (SDRs and CRs) is limited by their analog and mixed-signal front-ends and back-ends mostly because conventional sampling and reconstruction (S&R) circuits are inconsistent with the sampling theory. Novel S&R circuits (NSCs and NRCs) based on the direct interpretation of the sampling theorem improve the dynamic range, attainable bandwidth, energy efficiency, reconfigurability, adaptivity, and scale of integration of SDRs and CRs.