scispace - formally typeset
Search or ask a question

Showing papers on "Butterworth filter published in 1995"


Journal ArticleDOI
TL;DR: The authors provide a novel mapping of the proposed 1-D framework into 2-D that preserves the following: i) perfect reconstruction; ii) stability in the IIR case; iii) linear phase in the FIR case; iv) zeros at aliasing frequency; v) frequency characteristic of the filters.
Abstract: Proposes a novel framework for a new class of two-channel biorthogonal filter banks. The framework covers two useful subclasses: i) causal stable IIR filter banks. ii) linear phase FIR filter banks. There exists a very efficient structurally perfect reconstruction implementation for such a class. Filter banks of high frequency selectivity can be achieved by using the proposed framework with low complexity. The properties of such a class are discussed in detail. The design of the analysis/synthesis systems reduces to the design of a single transfer function. Very simple design methods are given both for FIR and IIR cases. Zeros of arbitrary multiplicity at aliasing frequency can be easily imposed, for the purpose of generating wavelets with regularity property. In the IIR case, two new classes of IIR maximally flat filters different from Butterworth filters are introduced. The filter coefficients are given in closed form. The wavelet bases corresponding to the biorthogonal systems are generated. the authors also provide a novel mapping of the proposed 1-D framework into 2-D. The mapping preserves the following: i) perfect reconstruction; ii) stability in the IIR case; iii) linear phase in the FIR case; iv) zeros at aliasing frequency; v) frequency characteristic of the filters. >

417 citations


Journal ArticleDOI
TL;DR: In this paper, an active power filter for single-phase power systems which are comprised of multiple nonlinear loads is presented. But the spectral performance of the active filter is not evaluated.
Abstract: This paper presents active power filters for single-phase power systems which are comprised of multiple nonlinear loads. The paper provides background on the operation of the filter, the details of the power circuit, the details of the control design, representative waveforms, and spectral performance for a filter which supports a 384 W AC controller and a 900 W uncontrolled bridge rectifier. Experimental data indicate that the active filter typically consumes 3% or less of the average load power, suggesting that a parallel filter is an efficient compensation approach. The spectral performance shows that the active filter brings the system into compliance with IEC-555 for decision frequencies in excess of 30 kHz. A discussion is presented outlining an alternative single-phase active filter which uses two controllable switches and is based on a half-bridge topology. >

212 citations


Proceedings ArticleDOI
08 Jun 1995
TL;DR: In this article, a 1.2 /spl mu/m CMOS process of a 5th order Butterworth polyphase filter with a bandwidth of 220 kHz centered around 250 kHz is reported.
Abstract: An analog integrated asymmetric polyphase filter is a key building block for the development of a high performance fully integrated low-IF receiver. The asymmetric polyphase filter makes it possible to suppress the mirror signal not at HF, but after quadrature demodulation at a low IF. The most important parameters for the polyphase filter are a high dynamic range and a good mirror signal suppression. This paper reports on the realisation in a 1.2 /spl mu/m CMOS process of a 5th order Butterworth polyphase filter with a bandwidth of 220 kHz centered around 250 kHz. Its measured mirror signal suppression is 64 dB. The active-RC implementation renders a 94.2 dB dynamic range at the input.

125 citations


Book
01 Sep 1995
TL;DR: In this article, the authors present a hands-on and academic approach to the design of EMI filters and the selection of components values using a mix of practical methods and theoretical analysis, including matrix methods using table data and the use of Fourier analysis, Laplace transforms and transfer function realization of LC structures.
Abstract: With today’s electrical and electronics systems requiring increased levels of performance and reliability, the design of robust EMI filters plays a critical role in EMC compliance. Using a mix of practical methods and theoretical analysis, EMI Filter Design, Third Edition presents both a hands-on and academic approach to the design of EMI filters and the selection of components values. The design approaches covered include matrix methods using table data and the use of Fourier analysis, Laplace transforms, and transfer function realization of LC structures. This edition has been fully revised and updated with additional topics and more streamlined content. New to the Third Edition Analysis techniques necessary for passive filter realization Matrix method and transfer function analysis approaches for LC filter structure design A more hands-on look at EMI filters and the overall design process Through this bestselling book’s proven design methodology and practical application of formal techniques, readers learn how to develop simple filter solutions. The authors examine the causes of common- and differential-mode noise and methods of elimination, the source and load impedances for various types of input power interfaces, and the load impedance aspect of EMI filter design. After covering EMI filter structures, topologies, and components, they provide insight into the sizing of components and protection from voltage transients, discuss issues that compromise filter performance, and present a goal for a filter design objective. The text also includes a matrix method for filter design, explains the transfer function method of LC structures and their equivalent polynomials, and gives a circuit design example and analysis techniques. The final chapter presents packaging solutions of EMI filters.

95 citations


Journal ArticleDOI
TL;DR: The results show the gain in performance if an MMD filter is used instead of a Volterra filter, due to the fact that with a given computational power longer memory lengths can be achieved by the MMD model.

91 citations


Proceedings ArticleDOI
05 Mar 1995
TL;DR: In this article, a hybrid-active filter topology was proposed to minimize electric utility current harmonics at high power levels, which combines both passive and active filters to obtain the lowest power converter VA rating as compared to the power converter rating in the active filter and the series-active hybrid filter configurations.
Abstract: A novel hybrid-active filter topology to minimize electric utility current harmonics at high power levels is presented in this paper. The proposed topology combines both passive and active filters to obtain the lowest power converter VA rating as compared to the power converter rating in the active filter and the series-active hybrid filter configurations. This is demonstrated with experimental results from a laboratory model. Transient and steady state operation of the hybrid-active filter is also presented. The control of the proposed filter is extended to provide damping of resonances due to neighboring loads. Simulation results showing the effect of active damping are presented. >

70 citations


Journal ArticleDOI
TL;DR: In this article, a control method with a combined filter system which senses load current, source current, and line voltage to create reference signals for an active filter is described, and a small setup controlled by a DSP is built, and the validity of the proposed method is demonstrated by experimental results.
Abstract: This paper describes a control method with a combined filter system which senses load current, source current, and line voltage to create reference signals for an active filter. The transfer function of the active filter is identified and is used for the control system design. It is shown that the source current feedback is most effective in suppressing the harmonic-enlarging effects due to parallel resonance and the harmonic current generated by source harmonic voltages. A small setup controlled by a DSP was built, and the validity of the proposed method was demonstrated by experimental results. >

68 citations


Journal ArticleDOI
TL;DR: In this paper, a universal filter using two current-feedback amplifiers is presented, which can realize lowpass, bandpass, high-pass, notch and all-pass filters from the same configuration.
Abstract: A new universal filter using two current-feedback amplifiers is presented. This filter can realise lowpass, bandpass, highpass, notch and allpass filters from the same configuration. It has the following features: orthogonal control of the natural frequency and quality factor by grounded resistors, easy conversion into a voltage-controlled filter, minimum active components, without match components except for allpass applications, low active and passive sensitivities, and cascadability. Experimental results that confirm the theoretical analyses are obtained.

66 citations


Journal ArticleDOI
Lars Sundström1
TL;DR: In this article, the authors investigate how reconstruction filters and sampling rate for a digital signal component separator affect the performance of an LINC transmitter and find a nearly optimal cutoff frequency for a given filter approximation and the minimum sampling rate.
Abstract: The authors investigate how reconstruction filters and sampling rate for a digital signal component separator affect the performance of an LINC transmitter. A method is described that, for a specified adjacent channel interference level, finds a nearly optimal cutoff frequency for a given filter approximation and the minimum sampling rate.

66 citations


Proceedings ArticleDOI
11 Sep 1995
TL;DR: In this paper, a new transfer function approach in passive harmonic filter design for industrial and commercial power system applications is presented, which incorporates IEEE-519 distortion limits directly into the design and component specification process.
Abstract: This paper details a new transfer function approach in passive harmonic filter design for industrial and commercial power system applications. Filter placement along with six common filter configurations are presented. Harmonic impedance, voltage division and current division transfer functions are derived and used in a practical filter design procedure which incorporates IEEE-519 distortion limits directly into the design and component specification process. A simple four step filter design procedure is outlined and used in a variable speed motor drive pumping plant application.

62 citations


Journal ArticleDOI
TL;DR: In this paper, a method to choose an appropriate sampling rate for the predistorter is presented and different filter approximations and orders are examined for different amplifier configurations by means of simulation.
Abstract: The reconstruction filters are a major source of error in digital predistortion linearizers for RF power amplifiers. This paper presents a method that enables these filters to be designed for lowest filter complexity and minimum degradation of the linearizer performance. Furthermore, a method to choose an appropriate sampling rate for the predistorter is presented. Different filter approximations and orders are examined for different amplifier configurations by means of simulation. Of the filters tried, low-ripple Chebyshev and Butterworth filters perform best. It is demonstrated that it is possible to account for reasonable filter misalignment with only a small degradation in performance. It is also shown that the effect on the modulation is small with respect to intersymbol interference. >

Journal ArticleDOI
TL;DR: In this paper, the performance degradation for a multi-wavelength optical network with a cascade of 30 randomly misaligned (de)multiplexers is evaluated using computer simulation techniques.
Abstract: In this letter, we use computer simulation techniques to evaluate the performance degradation for a multiwavelength optical network with a cascade of 30 randomly misaligned (de)multiplexers. Two types of (de)multiplexers are used in the simulation and modeled as third-order and first-order Butterworth filters, both with a full width half maximum (FWHM) of 125 GHz. For WDM systems at 10 Gb/s per optical channel with 30 uniformly misaligned filters, laser misalignment tolerances (at 0.3 dB distortion-induced-eye-closure penalty) vary from /spl plusmn/30 GHz (/spl plusmn/0.24 nm) for systems using third-order Butterworth filters to /spl plusmn/20 GHz (/spl plusmn/16 mm) for systems using first-order filters. >

Journal ArticleDOI
TL;DR: In this article, the design of multiresolution analog filter banks, linear models of cochlear function, with power dissipation being a prime engineering constraint, is presented, where a reasonable goodness criterion is the information rate through the system, per watt of power dissipated.
Abstract: We detail the design of multiresolution analog filter banks, linear models of cochlear function, with power dissipation being a prime engineering constraint. We propose that a reasonable goodness criterion is the information rate through the system, per watt of power dissipated. Speech applications requiring filter banks with a wide frequency tuning range, from 20 Hz to 20 kHz, and low power consumption make the transconductance-C integrator in subthreshold CMOS the preferable integrator structure. As an example, the dynamic range of a lowpass filter is computed and subsequently used to design a filter bank that models faithfully cochlear micro-mechanics. The power consumption of the entire filter bank is computed from analytical expressions and is estimated as 355 nW, at 68 kb/s overall information rate at the output of the system.

Patent
Tatsuya Miya, Tsutomu Noguchi1
24 Oct 1995
TL;DR: In this article, a 3-element series parallel resonator, which performs series resonance and parallel resonance at a predetermined frequency by a combination of three elements including an inductor and a capacitor of concentrated constants, is used as a filter.
Abstract: A 3-element series parallel resonator, which performs series resonance and parallel resonance at a predetermined frequency by a combination of three elements including an inductor and a capacitor of concentrated constants, is used as a filter. This design can reduce the circuit constant and permits a filter to be formed on a semiconductor substrate. Accordingly, various micro-wave devices, such as an amplifier which requires a filter, a frequency converter, a multiplier and signal switcher, can be designed into an MMIC on a single chip.

Proceedings ArticleDOI
28 Apr 1995
TL;DR: It is shown that the MDFT filter bank structure provides a cancellation of all odd alias spectra and it is demonstrated that even perfect reconstruction can be obtained by an appropriate filter design.
Abstract: The Modified DFT (MDFT) polyphase filter bank has been introduced. It has been shown that this filter bank is able to cancel adjacent channel alias spectra resulting in almost perfect reconstruction. Compared to cosine modulated filter banks, the MDFT filter bank offers advantages such as linear phase analysis and synthesis filters, complex valued signal processing with a different mapping into the subbands and low propagation delays. In this paper, it is shown that the MDFT filter bank structure provides a cancellation of all odd alias spectra. Furthermore, it is demonstrated that even perfect reconstruction can be obtained by an appropriate filter design.

Dissertation
01 Jan 1995
TL;DR: The essential scale-space causality property for local extrema of a signal under this operation is proved and it is shown that structuring functions from the "elliptic poweroids" lead to favourable dimensionality and semi-group properties.
Abstract: This thesis develops and demonstrates an original approach to scale-space theory. A new scale-space theory based on a unified multiscale morphological dilation-erosion smoothing operator is presented. The essential scale-space causality property for local extrema of a signal under this operation is proved. This result holds for signals on $\IR\sp2$ and higher dimensions and for negative as well as positive scales. When applied to grayscale images we show that structuring functions from the "elliptic poweroids" lead to favourable dimensionality and semi-group properties. Paraboloids, in particular, allow efficient computation of the scale-space, and such an algorithm is presented. The generalised frequency response of this signal smoother, which is similar to that of a Butterworth filter (with an amplitude dependent corner frequency), is obtained. The filter is statistically characterised by obtaining second-order statistical properties of the output signal with independent and identically distributed uniform noise input. Similar scale-space results are obtained for the multiscale morphological closing-opening operator, and we show that the resulting scale-space fingerprints are identical to those of the dilation-erosion. To demonstrate the utility of the new theory, we present an approach for the recognition of multiple 3-D objects in range data via the local matching of surfaces. In this approach the reduced morphological scale-space fingerprint is used as the primitive for matching. The resulting recognition process is invariant to translation, rotation, limited scaling, and partial occlusion. The results of the proposed object recognition method showing the recognition of a scene containing nine faces at various positions, angles and scales is presented. In a second demonstration we show the recognition of eight mountains in a digital elevation map.

Patent
24 May 1995
TL;DR: In this article, a two-frequency band pass filter without the need for a control signal, with excellent insertion loss by converting input output impedance of 1st and 2nd band pass filters for the connection is obtained.
Abstract: PURPOSE: To obtain the two-frequency band pass filter without the need for a control signal, with excellent entire insertion loss by converting input output impedance of 1st and 2nd band pass filters for the connection. CONSTITUTION: When a pass band of a 2nd band pass filter is in existence at the outside of the pass band of a 1st band pass filter 11, a frequency band of the filter 11 corresponding to the pass band of the filter 12 is converted into a high impedance by impedance matching circuits 13 in pairs. Similarly, when a pass band of the 1st band pass filter 11 is in existence at the outside of the pass band of the 2nd band pass filter 12, a frequency band of the filter 12 corresponding to the pass band of the filter 12 is converted into a high impedance by impedance matching circuits 14 in pairs. Then the matching circuits 13, 14 are connected by common input output terminals 15, 16 to extract the voltage. Through the constitution above, the two-frequency band pass filter not giving effect on each pass band of the filters 11, 12 is obtained. COPYRIGHT: (C)1996,JPO

Patent
28 Nov 1995
TL;DR: In this article, the first FIR filter stage of a synchronized multirate digital filter is selected from a database of filter coefficient/parameter sets which are loaded at start time and used in place.
Abstract: A digital lowpass filter design for filtering nuclear magnetic resonance (NMR) signals in a magnetic resonance imaging (MRI) system supports multiple frequencies and is highly configurable without sacrificing speed and performance. The first FIR filter stage of a synchronized multirate digital filter is selected from a database of filter coefficient/parameter sets which are loaded at start time and used in place. Second (and later) FIR filter stages are fixed in the example. A large number of different filter frequencies can be selected rapidly without changing the parameterized filter program reducing the amount of time to start up the MRI system and allowing "on-the-fly" filter changes in real-time between echoes and/or A/D samples. For example, filter preselection by the host can be performed in less than 100 μs, and, one of several lowpass filters can be selected by the sequencer "on-the-fly" with a latency of less than 10 μs.

Proceedings ArticleDOI
09 May 1995
TL;DR: The spectral correlation density of this output is derived and a criterion is derived to construct an optimal 2-band perfect reconstruction filter bank which minimizes the averaged mean squared reconstruction error when the high pass band signal is dropped.
Abstract: Generally, the output of a filter bank for a stationary input signal is not stationary but cyclostationary. In this paper, by using cyclostationary spectral analysis, the spectral correlation density of this output is derived. Using this result we derive a criterion to construct an optimal 2-band perfect reconstruction filter bank which minimizes the averaged mean squared reconstruction error when the high pass band signal is dropped. By adding constrains to the filter coefficients, the biorthogonal filter bank, the conjugate quadrature filter bank and the biorthogonal linear phase filter bank are respectively obtained. Some numerical results are also presented for optimal biorthogonal and PR linear phase filter banks which are compared in terms of some performance measures.

Proceedings Article
01 Sep 1995
TL;DR: In this article, a 3V 2nd order lowpass continuous-time filter is presented based on a highly linear pseudo-differential transconductor, whose output common-mode voltage is controlled at the filter level using lossy-integrators.
Abstract: In this paper a 3V 2nd order lowpass continuous-time filter is presented. The filter is based on a highly linear pseudo-differential transconductor. The input common-mode signal is cancelled at the transconductor level using a feed-forward path. The output common-mode voltage is controlled at the filter level using lossy-integrators. A prototype cell has been realized in 1.2?BiCMOS technology. The pole frequency can be tuned in the range 12-55MHz. A THD of ?40dB is achieved for signals up to lVpp at 10MHz. The dynamic range is approximately 60dB.

01 Jan 1995
TL;DR: In this paper, a generalization of the transversal filter, called Laguerre filter, is presented, which is obtained by replacing each delay by a first order all-pass section, and by applying a first-order low-pass filter (with the same pole used in the all pass) to the filter's input signal.
Abstract: In this tutorial paper we present a generalization of the transversal filter, called Laguerre filter, and study some of its more remarkable properties. This filter is obtained by replacing each delay of the transversal filter by a first order all-pass section, and by applying a first order low-pass filter (with the same pole used in the all-pass sections) to the filter’s input signal. Both the transversal and the lattice forms of the Laguerre filter are discussed. We also deduce the stationarity conditions of the mean-square error of a Laguerre filter (transversal or lattice) with respect to its pole position.

Journal ArticleDOI
TL;DR: Butterworth filters of second or higher order periodically inserted along a soliton transmission line, although they are flat at bandpass center, give to first order a nonzero restoring force for frequency and amplitude displacements.
Abstract: Butterworth filters of second or higher order periodically inserted along a soliton transmission line, although they are flat at bandpass center, give to first order a nonzero restoring force for frequency and amplitude displacements. The amplified spontaneous emission generated at the filter center frequency is significantly less than that generated with fixed etalons for the same damping strength.

Journal ArticleDOI
TL;DR: In this article, the effect of reconstruction filters on direct upconversion is analyzed in a multichannel environment and it is shown that the reconstruction filters cause gain and phase imbalances to be frequency dependent.
Abstract: The effect of reconstruction filters on direct upconversion is analyzed in a multichannel environment. It is shown that the reconstruction filters cause gain and phase imbalances to be frequency dependent, a significant issue when considering the stringent adjacent channel interference (ACI) requirements for radio communication systems. The paper introduces an adaptive compensation technique as a means of overcoming these frequency dependent imbalances. Additionally the error vector between the mismatched filters is shown to be directly related to the amount of sideband rejection. This analysis is then used to develop a methodology for analysing the performance of the novel compensation technique with respect to the mismatched filters. Analysis of the classical all-pole filter algorithms reveal the sensitivity of direct upconversion in a multichannel environment to filter order, type, percentage mismatch and differing ACI level. >

Patent
07 Jun 1995
TL;DR: In this paper, a multiple-stage tracking filter with a self-calibrating RC oscillator, a resistor and a capacitor is presented, where the capacitor is controlled by a switch which is connected to a timing signal.
Abstract: A multiple stage tracking filter includes a self-calibrating RC oscillator, a resistor connected to the self-calibrating RC oscillator and a capacitor connected to the self-calibrating RC oscillator. The filter further includes a switched capacitor filter element connected to the self-calibrating RC oscillator. The switched capacitor filter elements include a switch which is controlled by a timing signal from the self-calibrating RC oscillator. A method of filtering a signal includes the steps of operating a self-calibrating RC oscillator to generate a timing signal, tuning a plurality of cascaded filter elements with the generated timing signal and passing a signal through the plurality of tuned cascaded filter elements.

PatentDOI
TL;DR: In this article, a method of constructing an acoustic filter incorporating a technique in which the filter is modelled with one or more distributed two port elements is presented, where each distributed element is defined by a characteristic impedance and length and includes a waveguide filter which does not require damping.
Abstract: A method of constructing an acoustic filter incorporates a technique in which the filter is modelled with one or more distributed two port elements. The or each distributed element is defined by a characteristic impedance and length and includes a waveguide filter which does not require damping. The or each filter section is characterized in that it has at least two resonances which are used to shape a specified response for the filter. A substantially reactive acoustic filter constructed according to the method and a loudspeaker system incorporating such an acoustic filter are also disclosed.

Patent
17 Nov 1995
TL;DR: In this article, adjusted filter control parameters are provided to the filter, which filters the unfiltered signal (134) to produce the filtered signal (136), and the filter has a filtering function that agrees with the desired filtering control parameters (124).
Abstract: A method (100) and data processing system (120) for filtering unfiltered signals (134) to produce filtered signals (136) in accordance with desired filter control parameters (124) Adjusted filter control parameters (130) are generated based upon user provided filter control parameters (124) and operational characteristics (128) of a filter (132) to be implemented The adjusted filter control parameters (130) are provided to the filter (132) which filters the unfiltered signal (134) to produce the filtered signal (136) Using the adjusted filter control parameters (130), the filter (132) has a filtering function that agrees with the desired filter control parameters (124) to effectively band decouple the filter from bands of surrounding filters A graphic equalizer (250) implements a multi-band digital filter function that substantially reduces the effects of interband coupling, producing a more accurate match to the original filter response specification

Journal ArticleDOI
TL;DR: In this paper, the error-rate performance of an ON-OFF keyed lightwave receiver having an optical amplifier followed by a Fabry-Perot filter is compared with the well known result for a matched optical filter.
Abstract: The error-rate performance of an ON-OFF keyed lightwave receiver having an optical amplifier followed by a Fabry-Perot filter is analysed. Receivers with sampling and with integrate-and-dump threshold comparison are compared with the well known result for a matched optical filter. The Fabry-Perot filter decreases the sensitivity at 10/sup -9/error rate from 76 to 89 photon/bit (peak) with optimum optical filter bandwidth and postdetection integration time. >

Book ChapterDOI
01 Jan 1995
TL;DR: The realization of a 5th order lowpass Butterworth filter centered around a low IF of 250 kHz is presented and is designed for a low-IF receiver for high quality applications.
Abstract: When it comes to integratability, the zero-IF receiver is an alternative for the heterodyne or IF receiver. In recent years it has been introduced in several applications, but its performance can not be compared to that of the IF receiver yet. In this paper the principle of the low-IF receiver is introduced. The low-IF receiver has a topology which is closely related to the zero-IF receiver. Like the zero-IF receiver, the implementation of a low-IF receiver can be done with a high degree of integration. Its performance can however be better. The low-IF receiver is not sensitive to DC-offsets or LO to RF crosstalk. The principle of the low-IF receiver is based on the replacement of the LF lowpass filter of a zero-IF receiver with an analog integrated active polyphase filter. An active polyphase filter can perform a lowpass filter operation shifted on to a center frequency. In this paper the realization of a 5th order lowpass Butterworth filter centered around a low IF of 250 kHz is presented. This polyphase filter is designed for a low-IF receiver for high quality applications. The filter is implemented with the active-RC technique in a standard 1.2 μm CMOS process.

Patent
16 Nov 1995
TL;DR: In this paper, a two-stage filter is constructed by a single resonator so that a small plane filter is obtained, which has the satisfactory damping characteristic with small loss in a high frequency area and also has the stable characteristic against its external environment.
Abstract: PROBLEM TO BE SOLVED: To provide a small filter which has the satisfactory damping characteristic with small loss in a high frequency area and also has the stable characteristic against its external environment. SOLUTION: This filter has a specific structure where the input/output coupled circuits 102-103 are connected to a ring resonator 101 with a distance of 1/4 ring length secured between them and accordingly the discontinuity is caused in terms of an electromagnetic field at a position 107 that is symmetrical to the coupling points 105-106 of the circuits 102-103. Then a negative resistance circuit using a distributed coupling line 104 connected to a ring resonator line and a transistor is added to connect together two modes orthogonal to each other in the resonator 101. Thus, a two-stage filter is constructed by a single resonator so that a small plane filter is obtained. Furthermore, the satisfactory damping characteristic is attained for the plane filter since a damping pole is produced at a place near a passing band.

Patent
27 Dec 1995
TL;DR: In this article, a multi-stage comb filter is proposed for achieving substantial attenuation of aliasing or imaging bands of a signal to be filtered, where the comb filter can perform decimation or interpolation, depending upon its application.
Abstract: A digital filter is provided for achieving substantial attenuation of aliasing or imaging bands of a signal to be filtered. The digital filter employs a comb filter technique, wherein the comb filter can perform decimation or interpolation, depending upon its application. The comb filter is a multi-stage element, having more than one stage, and having an overall word length, W L , optimally reduced. The total number of terms within the cumulative set of stages is also optimally reduced. The comb decimation or interpolation filter architecture is therefore of minimum size if employed in hardware, or utilizes minimal operations if employed in software. A filter element within the comb decimation or interpolation filter includes a z-transform C K (Z) term. The filter element can be reduced to a simple z-transform 1+z -1 term if the stage of interest includes a decimate-by-two or interpolate-by-two rate change switch.