Performance comparison of square root raised-cosine and Lerner filters for the MDFT-TMUX filter bank

This paper deals with prototype filters, namely, square root raised - cosine (SRC) and Lerner filters in the modified discrete Fourier transform (MDFT) transmultiplexer (TMUX) filter bank. The error obtained in the reconstruction of the original signal is compared for various filter orders. These filters can be employed in multicarrier modulation system-overlapped discrete multitone (overlapped DMT) or discrete wavelet multitone (DWMT) system. However, with the SRC filter, a large number of computations are required. By employing Lerner filter, we can reduce this by atleast 3-folds. We have carried out simulation studies for an 8-channel MDFT-TMUX filter bank and the results obtained are presented.