Fast algorithms of discrete Fourier and cosine transforms in mixed formats

Differential formats allow for the representation of the signals in a digital form with small word length and at the same time they ensure a high fast-acting and simplicity of specialized processors. The use of differences and delta modulation (DM) for the economical representation and efficient processing of the stochastic signals needs the development of the existing methods and the working out of simple and fast processing methods in real time. Therefore discrete Fourier transform (DFT) and discrete cosine transform (DCT) methods were proposed in mixed formats which have the advantages as well DFT and DCT with PCM as DFT and DCT with differences or DM. In this work have been examined ways of choosing DM parameters with the aim of using them to these methods. At the same time, on the basis of computer simulation, the accuracy of these algorithms has been examined by means of signal voice processing. The use of DFT and DCT algorithms in such formats is expedient in neural systems.