Using the basic concept of Emri and Tschoegl, the algorithm for calculating relaxation time spectra has been improved such that excellent results are provided in the difficult case of polymers with narrow molar mass distributions. These spectra can be compared with those calculated by nonlinear regularization (Weese 1992), which we regard as a very exact method, and show equally good results with even less mathematical effort. Examples of dense relaxation time spectra (up to eight points per decade) are given for nearly monodisperse polystyrene melts and for mixtures of these up to four components. The relaxation time spectra describe the dynamic mechanical experimental data in each case with an overall error of less than 3%. The filtering method used to avoid physically senseless oscillations has been proven to resolve the characteristic peaks contributed by monodisperse polymers accurately.