DISCRETE NUMERICAL FILTERS.
Abstract
Low-pass filters using a finite convolution integral representation are discussed. This truncation of the underlying Fourier integrals causes undesirable oscillations in the frequency response of the filter that are suppressed by apodizing functions which smooth the frequency response and do not seriously degrade the sharpness of the cutoff. In this study, the low-pass filter is extended to the bandpass and differentiating filters, and further extension is made to the case of filters for discrete, equally spaced data. An analysis of the error in going from the continuous to the discrete case is presented, and it is shown that this error can easily be kept to less than 1 in 10,000. Finally, some examples of filter applications are given. (Author)
Document Details
- Document Type
- Technical Report
- Publication Date
- Jan 01, 1966
- Accession Number
- AD0480134
Entities
People
- R. J. Stirton
Organizations
- Naval Air Weapons Station China Lake