Digital lowpass filtering using the discrete Hilbert transform

It has been shown in the recent literature that the use of discrete Hilbert transform in digital filtering reduces the operation of filtering to that of a matrix multiplication, In this paper, we present an alternative matrix formulation for a digital low pass filter, which requires less storage and reduces the number of multiplications by a factor of two. From a detailed analysis of the errors due to quantization and roundoff, it is shown that the new formulation has superior mean squared signal to error ratio.