The paper proposes numerical methods to calculate the data transfer rate over a communication channel with real signals in the presence of a mixture of Gaussian and non-Gaussian correlated noise. This paper considers the transmission of discrete and continuous messages. The estimate of the data transmission rate given by the proposed methods is much more accurate than the estimation given by the Shannon limit. The proposed methods can be used for the information-theoretical analysis of the wide range of systems: systems of digital and analogous communication, radiolocation and radio navigation.