Spectral analysis of audio signal using complex discrete cosine transform

The main method of spectral analysis of the sound broadcasting signal (SBS) remains the fast Fourier transform. One of its disadvantages is insufficient resolution, which does not allow to estimate closely spaced spectral components of the SBS, as well as a sufficiently rich spectrum of side lobes deter-mined by the window function used. The need to use window functions also leads to the need to work with overlaps of the analyzed sequence. The resolution of the discrete cosine transform (DCT) is twice as high, but its use was constrained by a wide range of side lobes. The advantage of DCT is also the absence of the need to use window functions, and, accordingly, overlaps. The authors propose an algorithm for a complex DCT with side lobe compensation and in-creased accuracy of determining the parameters of the spectral components due to multiple transposition of the SB spectrum. The algorithm allows to bring the resolution and accuracy of spectral analysis closer to the parameters of the human peripheral auditory analyzer. The developed algorithm of complex cosine discrete transform provides formation of an estimate similar to DFT but with twice higher resolution and lower level of side lobes due to use of active window functions. The obtained results showed the efficiency of the developed algorithm and the prospects of its use in analysis of the spectrum of the audio signal, its coding and processing.