The idea of improving the methods of processing the received radio signals used in radio communication and radiolocation systems is considered. It is proposed to consider the radio signal as a noise-resistant code with repetition of its periods (symbols), each of which carries the same information about the value of its amplitude and initial phase. The possibility of realization of coherent reception of radio signals with amplitude or biphasic discrete modulation under the influence of powerful noise on them by their most "reliable" (suitable for calculations) fragments, determined by the values of samples, received at moments when the amplitude of a received signal shall be equal to zero (samples correspond to the value of the noise component – point A) or be maximum – point E (the phase of signals shifted with respect to the phase of noise, by π/2 or 3π/4). The considered approach on processing of a mix of a signal and noise allows to define some varieties of "reliable" fragments (at difference of phases of noise and a signal on π/2 or 3π/4, at a match of zero values of signals and noise, at difference of phases of noise and a signal on the value smaller, than ±π/24) on values of samples in points A and E, and also to exclude from consideration fragments, decision on which often is erroneous. The possibility of obtaining an additional estimation ∆ of the quality of the decision on a fragment of a discrete signal (BPSK), with already calculated its average value SСР on the set of its fragments, which allows to select the best or several best fragments for the decision on the received signal is shown. Fragments with a difference of noise and signal phases on the value smaller than ±π/24 occur most often from all considered variants and allow to distinguish with high accuracy the initial phases of opposite signals (BPSK). To solve the problem, we used all possible temporal representations of fragments of signal, noise and a mixture of signal and noise, mathematical models and formulas of the theory of circuits and signals.