Low-pass filtering with filters defined in the Discrete Trigonometric Transform domains

In this paper a novel approach to discrete signal filtering based on the Discrete Trigonometric Transforms (DTTs) is proposed. The concept is derived from the generalized convolution, modified by the rectangular windowing applied to the convolution matrix. The experimental results obtained for low-pass filters defined in domains of all 16 DTTs, as compared to the relevantly adjusted Finite Impulse Response (FIR) filter characteristic in the Fourier domain, indicate that the proposed approach may be treated as offering resembling alternatives to the classical Fourier-based FIR filtering. This observation opens new direction of interesting DTT applications.