Systems and methods for a partial sum digital fir filter

• US 6,718,355 B2
• Filed: 02/05/2001
• Issued: 04/06/2004
• Est. Priority Date: 02/05/2001
• Status: Expired due to Term

• Current Assignee: Synaptics Incorporated
• Original Assignee: Conexant Systems Incorporated (Synaptics Incorporated)
• Inventors: Wang, Wilson, Felts, Benjamin E. III
• Primary Examiner(s): Malzahn, David H.
• Application Number:
  US09/777,622
• Publication Number:
  US 20020138531A1
• Time in Patent Office: 1,156 Days
• Field of Search: 708/319
• US Class Current: 708/319
• CPC Class Codes: H03H 17/0227 Measures concerning the coefficients
• CPC Class Codes: H03H 17/06 Non-recursive filters

Abstract

A digital FIR filter is provided that inputs a series of data samples x[0] . . . x[n] and generates a partial sum output PS[i], where i≦n. The partial sum output is a weighted version of the a difference between a partial sum of the previous i−1 data samples, PS[i−1], and the current data sample x[n] added to the current data sample x[n]. The filter includes a plurality of weighting stages. Each weighting stage includes a first adder for subtracting the current data sample x[n] from the previous partial sum PS[i−1], a multiplier that multiplies the difference by a weighting coefficient, and a second adder that sums the weighted difference with the current data sample. The filter also includes a plurality of delay elements, each of which inputs a partial sum and imposes a unit delay on the partial sum before supplying it to a weighting stage.

Claims

1. A digital filter, comprising:

a plurality of stages, wherein each of the plurality of stages comprises;

a first adder configured to subtract an input from a first partial sum producing a sum,

a multiplier configured to multiply the sum by a gain coefficient producing a product, and

a second adder configured to sum the product with the input to produce a second partial sum.

2. The digital filter of claim 1, wherein each stage comprises a delay element for storing the first partial sum for a predetermined period.

3. The digital filter of claim 1, wherein the gain coefficient is normalized.

4. The digital filter of claim 3, wherein the gain coefficient is normalized at least in part by dividing the gain for each of the plurality of stages by the total gain up to each of the plurality of stages.

5. The digital filter of claim 1, wherein the input is a data sample.

6. The digital filter of claim 1, wherein the input is an input line.

7. The digital filter of claim 1, wherein the digital filter is configured for n inputs, and wherein the digital filter comprises n−

1 stages.

8. The digital filter of claim 1, wherein the digital filter is configured for n inputs, and wherein there are n/2 stages.

9. The digital filter of claim 1, wherein the gain coefficients are adjustable.

10. A digital filter, comprising:

a plurality of stages, wherein each of the plurality of stages comprises;

a first adder configured to subtract a first partial sum from an input producing a sum,

a multiplier configured to multiply the sum by a gain coefficient producing a product, and

a second adder configured to sum the product with the first partial sum to produce a second partial sum.

11. The digital filter of claim 10, wherein each stage comprises a delay element for storing the first partial sum for a predetermined period.

12. The digital filter of claim 10, wherein the gain coefficient is normalized.

13. The digital filter of claim 12, wherein the gain coefficient is normalized at least in part by dividing the gain for each of the plurality of stages by the total gain up to each of the plurality of stages.

14. An encoding system, comprising:

a digital filter comprising a plurality of stages, each of the plurality of stages configured to receive a first partial sum and an input; and

an adaptive filter configured to receive the difference between the first partial sum and the input for each of the plurality of stages, and to produce an adjustable gain coefficient for use by each of the plurality of stages based on the difference between the first partial sums and the inputs.

15. The encoding system of claim 14, wherein the adjustable gain coefficients are produced from predetermined sets of gain coefficients.

16. The encoding system of claim 15, wherein the predetermined sets of gain coefficients comprises one set for more aggressive filtering and one set for less aggressive filtering.

17. A digital filter, comprising:

a plurality of stages, wherein each of the plurality of stages comprises;

a means for subtracting an input from a first partial sum producing a sum,

a means for multiplying the sum by a gain coefficient producing a product, and

a means for summing the product with the input to produce a second partial sum.

18. The digital filter of claim 17, comprising a means for delaying the first partial sum for a predetermined period.

19. The digital filter of claim 18, comprising a means for normalizing the gain coefficient.

20. A method for digital filtering, comprising:

subtracting an input from a first partial sum producing a sum,

multiplying the sum by a gain coefficient producing a product, and

summing the product with the input to produce a second partial sum.

21. The method of claim 20, comprising delaying the first partial sum for a predetermined period.

22. The method of claim 21, comprising normalizing the gain coefficient.