Overload protection for look-ahead delta sigma modulators

• US 7,081,843 B2
• Filed: 03/25/2005
• Issued: 07/25/2006
• Est. Priority Date: 07/19/2004
• Status: Active Grant

• Current Assignee: Cirrus Logic Incorporated
• Original Assignee: Cirrus Logic Incorporated
• Inventors: Melanson, John L.
• Primary Examiner(s): Barnie, Rexford
• Assistant Examiner(s): NGUYEN, KHAI M
• Application Number:
  US11/090,372
• Publication Number:
  US 20060012500A1
• Time in Patent Office: 487 Days
• Field of Search: 341143-144, 341/155
• US Class Current: 341/143
• CPC Class Codes: H03M 3/366 in feed-forward mode, e.g. using look-ahead circuits
• CPC Class Codes: H03M 3/43 the quantiser being a single bit one
• CPC Class Codes: H03M 3/438 the modulator having a higher order loop filter in the feedforward path

Abstract

Look-ahead delta sigma modulators of the signal processing systems described herein can anticipate quantizer overload. By anticipating quantizer overload, the look-ahead delta sigma modulators can select an output value y(n) that may have a lower SNR but will prevent quantizer overload in the future. A quantizer overload protection process determines the amount of look-ahead depth to drive state variables of a loop filter of the look-ahead delta sigma modulator to values that would prevent future quantizer overload. By substituting a quantizer of the look-ahead delta sigma modulator with a gain and determining a closed loop impulse response of a look-ahead delta-sigma modulator, the discrete time to achieve an absolute value maximum closed loop response magnitude of the look-ahead delta-sigma modulator has been determined to be directly related to the look-ahead depth that will prevent future quantizer overload.

Claims

1. A method of anticipating quantizer overload of a look-ahead delta sigma modulator, the method comprising:

determining a likelihood of quantizer overload of the look-ahead delta sigma modulator from a closed loop response of the look-ahead delta-sigma modulator to a closed loop feedback signal, wherein the look-ahead delta-sigma modulator includes a gain stage substitute for a quantizer, wherein the look-ahead delta sigma modulator has a look-ahead depth of n1 and n1 is a positive integer greater than one.

2. The method of claim 1 further comprising:

substituting a quantizer of the look-ahead delta sigma modulator with the gain stage having a constant gain.

3. The method of claim 1 wherein a look-ahead depth of the look-ahead delta sigma modulator equals n1, the method further comprising:

determining a number, n2, of input samples to anticipate quantizer overload of the look-ahead delta sigma modulator, wherein n2 is an integer;

determining a set of state variables for the loop filter of the look-ahead delta sigma modulator for at least n1 elements of an input signal vector X(n) and at least n1 elements of each output candidate vector Y(n)_i, wherein “

i”

is an element of at least a subset of the set {0, 1, . . . , 2<sup>n1−

1</sup>} and the n1 elements of input signal vector X(n) represent n1 input signal samples;

substituting a quantizer of the look-ahead delta sigma modulator with a gain stage to provide a closed loop feedback path in the look-ahead delta sigma modulator;

determining closed loop output responses of the look-ahead delta sigma modulator for n2 input samples using at least one set of determined state variables;

determining which closed loop output response has a lowest maximum absolute value closed loop filter output response; and

if overload of the look-ahead delta-sigma modulator is anticipated, selecting an output of the look-ahead delta sigma modulator from the output candidate vector corresponding to the determined closed loop output response having the lowest maximum absolute value closed loop look-ahead delta sigma modulator response.

4. A signal processing system comprising a look-ahead delta-sigma modulator with overload protection, the look-ahead delta-sigma modulator comprising:

an overload protection module to determine a likelihood of quantizer overload of the look-ahead delta sigma modulator from a closed loop response of the look-ahead delta-sigma modulator to a closed loop feedback signal, wherein the look-ahead delta-sigma modulator includes a gain stage substitute for a quantizer, wherein the look-ahead delta sigma modulator has a look-ahead depth of n1 and n1 is a positive integer greater than one.

5. The signal processing system of claim 4 wherein the look-ahead delta-sigma modulator further comprises:

an input to receive an input signal vector X(n), wherein the input signal vector X(n) has n1 elements;

a loop filter coupled to the input;

a quantizer coupled to the loop filter;

a gain module coupled to the loop filter; and

a memory to store a set of state variables for the loop filter of the look-ahead delta sigma modulator for at least n1 elements ofan input signal vector X(n) and at least n1 elements of each output candidate vector Y(n)_i, wherein “

i”

is an element of at least a subset of the set {0, 1, . . . , 2<sup>n1−

1</sup>}, n1 is a positive integer greater than one, and the n1 elements of input signal vector X(n) represent n1 input signal samples;

wherein the overload protection module is configured to;

substitute the quantizer with the gain stage to provide a closed loop feedback path in the look-ahead delta sigma modulator;

determine closed loop output responses of the look-ahead delta sigma modulator for n2 input samples using at least one set of determined state variables, wherein n2 represents a number of input samples to anticipate overload of the quantizer;

determine which closed loop output response has a lowest maximum absolute value closed loop output response; and

if overload of the look-ahead delta-sigma modulator is anticipated, select an output of the look-ahead delta sigma modulator from the output candidate vector corresponding to the determined closed loop output response having the lowest maximum absolute value closed loop output response.

6. A method of quantizing an input signal using a look-ahead delta-sigma modulator having quantizer overload protection, wherein a look-ahead depth of the look-ahead delta sigma modulator equals n1, the method comprising:

determining a set of state variables for a loop filter of the look-ahead delta sigma modulator for at least n1 elements ofan input signal vector X(n) and at least n1 elements of each output candidate vector Y(n)_i, wherein “

i”

is an element of at least a subset ofthe set {0, 1, . . . , 2<sup>n1−

1</sup>}, n1 is a positive integer greater than one, and the n1 elements of input signal vector X(n) represent n1 input signal samples;

substituting a quantizer of the look-ahead delta sigma modulator with a gain stage to provide a closed loop feedback path in the look-ahead delta sigma modulator;

determining closed loop output responses of the look-ahead delta sigma modulator for n2 input samples using at least one set of determined state variables, wherein n2 represents a number of input samples to anticipate quantizer overload of the look-ahead delta sigma modulator;

determining which closed loop output response has a lowest maximum absolute value closed loop output response; and

if overload of the look-ahead delta-sigma modulator is anticipated, selecting an output of the look-ahead delta sigma modulator from the output candidate vector corresponding to the determined closed loop output response having the lowest maximum absolute value closed loop output response.

7. The method of claim 6 further comprising:

determining the number n2 of input samples to anticipate quantizer overload of the look-ahead delta sigma modulator.

8. The method of claim 7 wherein determining the number n2 of input samples to anticipate quantizer overload of the look-ahead delta sigma modulator comprises determining a minimum discrete time to achieve an absolute value maximum closed loop response magnitude of the look-ahead delta-sigma modulator.

9. The method of claim 6 wherein selecting an output of the look-ahead delta sigma modulator from the output candidate vector corresponding to the determined closed loop output response having the lowest maximum absolute value closed loop output response further comprises selecting a leading bit from the output candidate vector corresponding to the determined closed loop output response having a lowest maximum absolute loop filter output value as an output of the look-ahead delta sigma modulator.

10. The method of claim 6 further comprising:

updating actual state variables using the selected output of the look-ahead delta-sigma modulator as feedback for the loop filter of the look-ahead delta-sigma modulator.

11. The method of claim 6 further comprising:

determining if overload of the look-ahead delta-sigma modulator is anticipated, wherein determining if overload of the look-ahead delta-sigma modulator is anticipated comprises;

determining if the lowest maximum absolute closed loop filter output value exceeds a predetermined value.

12. The method of claim 11 wherein the predetermined value equals 0.75 times a step size of the quantizer of the look-ahead delta-sigma modulator.

13. The method of claim 6 wherein determining a set of state variables for a loop filter of the look-ahead delta sigma modulator further comprises:

determining a set of state variables for the loop filter of the look-ahead delta sigma modulator for at least n1 elements of the input signal vector X(n) and at least n1 elements of each output candidate vector Y(n)₁.

14. The method of claim 6 further comprising:

determining a best match between an output candidate vector and the input signal vector X(n); and

if overload of the look-ahead delta-sigma modulator is not anticipated, selecting an output of the look-ahead delta sigma modulator from the output candidate vector that best matches the input signal vector X(n).

15. The method of claim 6 wherein determining closed loop output responses of the look-ahead delta sigma modulator for n2 input samples using at least one set of determined state variables comprises determining closed loop output responses of the look-ahead delta sigma modulator for n2 input samples using a set of determined state variables from the output candidate vector Y(n) that best matches the input signal vector X(n).

16. The method of claim 6 wherein the at least one set of determined state variables comprises state variables determined from one or more input samples in a subset of a group of T input samples, wherein T is an integer greater than or equal to two.

17. The method of claim 16 wherein the one or more input samples in the subset of a group of T input samples subset represents an average of at least a subset of the T input samples.

18. The method of claim 6 wherein the input signal vector X(n) represents input data derived from audio input signal data.

19. The method of claim 6 further comprising:

recording quantized quantizer input signal data on storage media.

20. A signal processing system to quantize an input signal using a look-ahead delta-sigma modulator with quantizer overload protection module, wherein a look-ahead depth of the look-ahead delta sigma modulator equals n1, the signal processing system comprising:

an input to receive an input signal vector X(n), wherein the input signal vector X(n) has n1 elements and n1 is a positive integer greater than one and equals a look-ahead depth of the look-ahead delta-sigma modulator;

a loop filter coupled to the input;

a quantizer coupled to the loop filter;

a gain stage coupled to the loop filter; and

a memory to store a set of state variables for the loop filter of the look-ahead delta sigma modulator for at least n1 elements of an input signal vector X(n) and at least n1 elements of each output candidate vector Y(n)_i, wherein “

i”

is an element of at least a subset of the set {0, 1, . . . , 2<sup>n1−

1</sup>}, n1 is a positive integer greater than one, and the n1 elements of input signal vector X(n) represent n1 input signal samples;

wherein the overload protection module is configured to;

substitute the quantizer with the gain stage to provide a closed loop feedback path in the look-ahead delta sigma modulator;

determine closed loop output responses of the look-ahead delta sigma modulator for n2 input samples using at least one set of determined state variables, wherein n2 represents a number of input samples to anticipate overload of the quantizer;

determine which closed loop output response has a lowest maximum absolute value closed loop output response; and

if overload of the look-ahead delta-sigma modulator is anticipated, select an output of the look-ahead delta sigma modulator from the output candidate vector corresponding to the determined closed loop output response having the lowest maximum absolute value closed loop output response.

21. The signal processing system of claim 20 wherein the number n2 of input samples to anticipate quantizer overload of the look-ahead delta sigma is at least equal to a discrete time to achieve an absolute value maximum closed loop response magnitude of the look-ahead delta-sigma modulator.

22. The signal processing system of claim 20 wherein the overload protection module is further configured to select a leading bit from the output candidate vector corresponding to the determined closed loop output response having a lowest maximum absolute loop filter output value as an output of the look-ahead delta sigma modulator.

23. The signal processing system of claim 20 wherein the overload protection module is further configured to:

update actual state variables using the selected output of the look-ahead delta-sigma modulator as feedback for the loop filter of the look-ahead delta-sigma modulator.

24. The signal processing system of claim 20 wherein the overload protection module is further configured to:

determine if overload of the look-ahead delta-sigma modulator is anticipated, wherein to determine if overload of the look-ahead delta-sigma modulator is anticipated comprises;

determining if the lowest maximum absolute closed loop filter output value exceeds a predetermined value.

25. The signal processing system of claim 24 wherein the predetermined value equals 0.75 times a step size of the quantizer of the look-ahead delta-sigma modulator.

26. The signal processing system of claim 20 wherein the overload protection module is further configured to:

determining a set of state variables for the loop filter of the look-ahead delta sigma modulator for at least n1 elements of the input signal vector X(n) and at least n1 elements of each output candidate vector Y(n)₁.

27. The signal processing system of claim 20 further comprising:

signal processing and recording equipment to process output data from the quantizer and record the processed output data on storage media.

28. The signal processing system of claim 20 wherein the input signal vector X(n) represents input data derived from audio input signal data.

29. The signal processing system of claim 20 wherein the quantizer is configured to determine a best match between an output candidate vector and the input signal vector X(n) and, if overload of the look-ahead delta-sigma modulator is not anticipated, the quantizer is further configured to select an output of the look-ahead delta sigma modulator from the output candidate vector that best matches the input signal vector X(n).

30. The signal processing system of claim 20 wherein the at least one set of determined state variables comprises state variables determined from one or more input samples in a subset of a group of T input samples, wherein T is an integer greater than or equal to two.

31. The signal processing system of claim 30 wherein the one or more input samples in the subset of a group of T input samples subset represents an average of at least a subset of the T input samples.

32. An apparatus to quantize an input signal using a look-ahead delta-sigma modulator with quantizer overload protection, wherein a look-ahead depth of the look-ahead delta sigma modulator equals n1, the apparatus comprising:

means for determining a set of state variables for a loop filter of the look-ahead delta sigma modulator for at least n1 elements of an input signal vector X(n) and at least n1 elements of each output candidate vector Y(n)_i, wherein “

i”

is an element of at least a subset of the set {0, 1, . . . , 2<sup>n1−

1</sup>}, n1 is a positive integer greater than one, and the n1 elements of input signal vector X(n) represent n1 input signal samples;

means for substituting a quantizer of the look-ahead delta sigma modulator with a gain stage to provide a closed loop feedback path in the look-ahead delta sigma modulator;

means for determining closed loop output responses of the look-ahead delta-sigma modulator for n2 input samples using at least one set of determined state variables, wherein n2 represents a number of input samples to anticipate quantizer overload of the look-ahead delta sigma modulator;

means for determining which closed loop output response has a lowest maximum absolute value closed loop output response; and

means for selecting an output of the look-ahead delta sigma modulator from the output candidate vector corresponding to the determined closed loop output response having the lowest maximum absolute value closed loop output response if overload of the look-ahead delta-sigma modulator is anticipated.

33. A method comprising:

substituting a gain stage for a quantizer in a look-ahead delta sigma modulator, wherein the look-ahead delta sigma modulator has a look-ahead depth of n1 and n1 is a positive integer greater than one;

providing a first input signal to the gain stage;

determining a feedback signal from the gain stage;

providing the feedback signal to a loop filter of the look-ahead delta sigma modulator;

providing a second input signal to the loop filter;

determining an output of the loop filter from a combination of the input signal and the feedback signal;

substituting the quantizer for the gain stage in the look-ahead delta sigma modulator; and

selecting a quantization output using the loop filter.

34. The method of claim 33 wherein substituting a gain state further comprises substituting a gain stage having a constant gain.

35. The method of claim 33 wherein providing a second input signal further comprises providing an audio input signal to the loop filter of the look-ahead delta sigma modulator.

36. The method of claim 33 further comprising:

recording the quantization output on storage media.

37. A signal processing system comprising:

a look-ahead delta sigma modulator, wherein the look-ahead delta sigma modulator has a look-ahead depth of n1 and n1 is a positive integer greater than one, the look-ahead delta sigma modulator comprising;

a loop filter to filter a combination of an input signal and a feedback signal;

a quantizer coupled to an output of the loop filter to select a quantization output using an output of the loop filter;

a gain stage coupled to the output of the loop filter to generate the feedback signal; and

a module to substitute the gain stage for the quantizer to allow the gain stage to receive an output signal from the loop filter and generate the feedback signal for the loop filter.

38. The signal processing system of claim 37 wherein the gain stage has a constant gain.

39. The signal processing system of claim 37 wherein the input signal is an audio input signal.

40. The signal processing system of claim 37 further comprising:

signal processing and recording equipment to process output data from the quantizer and record the processed output data on storage media.