We address the issues raised in the preceding comment by Mandelshtam and Carrington [Phys. Rev. E 65, 028701 (2002)], concerning the eigenvalue determination by the spectral filter methods. We argue that the Fourier transformation is the building block of all currently known time-domain spectral filter algorithms, and, therefore, the time-energy uncertainty principle affects them all in a similar manner. We also explain the situation when the correlation function method may be less suitable, in comparison to the filter diagonalization method, for the determination of eigenvalues, even though both share the same fundamental principles.