Interpolation filter and method for digitally interpolating a digital signal

• US 20020184278A1
• Filed: 05/01/2002
• Published: 12/05/2002
• Est. Priority Date: 07/05/2000
• Status: Active Grant

• Current Assignee: Intel Corporation
• Original Assignee: Infineon Technologies AG
• Inventors: Menkhoff, Andreas
• Granted Patent: US 7,225,213 B2
• Time in Patent Office: Days
• US Class Current: 708/313
• CPC Class Codes: H03H 17/0294 Variable filters; Programmable filters
• CPC Class Codes: H03H 17/0664 where the output-delivery frequency is lower than the input sampling frequency, i.e. decimation

Abstract

The invention relates to an interpolation filter and to a method for filtering a digital input signal. The interpolation filter has an amplitude characteristic with a low-pass-shaped damping curve in the useful signal frequency range of the digital input signal. The group delay time of the interpolation filter is essentially constant in the useful signal frequency range and can be adjusted within a clock period of the equidistant digital signal.

Claims

1. An interpolation filter for filtering a digital input signal whose amplitude response exhibits a lowpass-type attenuation curve in the useful signal frequency band Δ

f_nutz of the digital input range.

3. The interpolation filter as claimed in claim 1 or 2, wherein connected downstream of the interpolation filter (5) is a highpass filter (9) for compensating the lowpass-type amplitude response.

4. The interpolation filter as claimed in one of the preceding claims, wherein the group delay τ

of the interpolation filter (5) runs in an essentially constant fashion in the useful signal frequency band Δ

f_nutz of the digital input signal.

5. The interpolation filter as claimed in one of the preceding claims, wherein the digital input signal is an equidistant digital signal with a predetermined clock pulse period T_in.

6. The interpolation filter as claimed in one of the preceding claims, wherein the group delay τ

of the interpolation filter (5) can be set inside a clock pulse period T_in of the digital input signal.

7. The interpolation filter as claimed in one of the preceding claims, wherein the ratio of the clock pulse periods of the digital input signal T_ln and the digital output signal T_aus filtered by the interpolation filter (5) can be set.

8. The interpolation filter as claimed in claim 3, wherein the interpolation filter (5) and the downstream highpass filter (9) together exhibit a sinc filter characteristic.

9. The interpolation filter as claimed in one of the preceding claims, characterized in that a further interpolation filter can be connected upstream of the interpolation filter (5) for the purpose of constricting the useful signal frequency band Δ

f_nutz.

10. The interpolation filter as claimed in claim 9, wherein the interpolation filter that can be connected upstream is a polyphase filter.

11. The interpolation filter as claimed in one of the preceding claims, wherein it has:

a filter coefficient generator (15) for generating filter coefficients as a function of a base function BF;

a multiplier (13) for multiplying the digital input signal by the generated filter coefficients, and

an accumulator (23) for accumulating the digital input signal weighted by the multiplication.

12. The interpolation filter as claimed in one of the preceding claims, defined by a storage device (20) for storing the base function.

13. The interpolation filter as claimed in one of the preceding claims 1 to 11, defined by a base function generator for generating the base function as a function of fundamental functions.

14. The interpolation filter as claimed in claim 13, defined by a storage device for storing the fundamental functions.

15. The interpolation filter as claimed in one of the preceding claims, wherein a controllable switching device (28) is provided for reading out the weighted digital input signal as a digital output signal.

16. The interpolation filter as claimed in one of the preceding claims, wherein the accumulator (23) consists of an adder (24) and a register (26) whose output is fed back to an input of the adder (24).

18. The method as claimed in claim 17, in which the filter coefficients of the interpolation filter (5) are determined as a function of a base function BF.

25. The method as claimed in one of the preceding claims, in which a multiplicity of sets of filter coefficients of the interpolation filter (5) are generated as a function of the base function BF which in each case exhibit in the useful signal frequency band Δ

f_nutz an essentially identical amplitude response and different group delays τ

, there subsequently being selected for the purpose of determining the filter coefficients of the interpolation filter (5) that set of filter coefficients whose group delay τ

corresponds to the set desired group delay τ

_soll.

2. An interpolation filter for filtering a digital input signal whose amplitude response exhibits a lowpass-type attenuation curve both inside and outside the useful signal frequency band Δ

f_nutz.

17. A method for digital interpolation of a digital input signal, having the following steps:

(a) receiving a digital input signal with a predetermined clock frequency f_in;

(b) determining filter coefficients of a settable interpolation filter whose amplitude response exhibits a lowpass-type attenuation curve in the useful signal frequency band of the digital input signal;

(c) filtering the digital input signal by means of the set interpolation filter.

19. The method as claimed in one of the preceding claims 17 or 18, in which the base function BF is stored in a memory (20).

20. The method as claimed in one of the preceding claims 17 or 18, in which the base function BF is generated from prescribed fundamental functions GF.

21. The method as claimed in claim 20, in which a first fundamental function is a time-limited power sine function.

22. The method as claimed in claim 21, in which the first fundamental function is as follows:

h₁(t)=sin[t·

π

/n]^m·

σ

(t)−

sin[t·

π

/n]^m·

σ

(t−

n)m,n>

=1 m,nε

R,

σ

(t−

n) being the unit-step function at the instant n.

23. The method as claimed in claim 20, in which the second fundamental function GF is a first-order sample-and-hold function.

24. The method as claimed in claim 21, in which the second fundamental function is as follows:

h₂(t)=σ

(t)−

σ

(t−

n),

σ

(t−

n) being the unit-step function at the instant n.