Look-ahead delta sigma modulator having an infinite impulse response filter with multiple look-ahead outputs

• US 20050156771A1
• Filed: 01/18/2005
• Published: 07/21/2005
• Est. Priority Date: 01/16/2004
• Status: Active Grant

• Current Assignee: Cirrus Logic Incorporated
• Original Assignee: Cirrus Logic Incorporated
• Inventors: Melanson, John L.
• Granted Patent: US 7,187,312 B2
• Time in Patent Office: Days
• US Class Current: 341/143
• CPC Class Codes: H03M 7/3011 of non-linear distortion, e.g. by temporarily adapting the operation upon detection of instability conditions avoiding instability by structural design H03M7/3035

Abstract

A delta sigma modulator look-ahead IIR filter response can be divided into a natural response and a pattern response using superposition techniques. The look-ahead delta sigma modulators of the signal processing systems described herein include an infinite impulse response filter (“IIR”) that produces multiple output look-ahead natural responses to input signals having a look-ahead depth of “M”. The multiple output lookahead IIR filter reduces the amount of processing and memory used by the delta sigma modulator to generate quantizer output values. The multiple output lookahead IIR filter uses extended delay stages and modified feedback coefficients to concurrently produce multiple look-ahead natural responses. In one embodiment, the multiple output lookahead IIR filter concurrently produces M look-ahead natural responses.

Claims

1. A method of processing a signal using a multiple output infinite impulse response (“

IIR”

) filter of a look-ahead delta-sigma modulator, the method comprising;

determining a set of state variables for respective stages of the IIR filter for a time t from a set of input signals, wherein at least a subset of the state variables of the respective stages of the IIR filter represent a current response and at least one future response elements of a natural response vector of the IIR filter, wherein t represents a time index value; and

determining a quantization output of the look-ahead delta sigma modulator for the time t using the natural response vector determined for the time t.

2. The method as in claim 1 wherein the look-ahead delta-sigma modulator has a look-ahead depth of M, the number of multiple natural response outputs of the IIR filter used for determining a quantization output of the look-ahead delta sigma modulator for the time t is less than or equal to M, and M is an integer greater than or equal to two.

3. The method as in claim 1 further comprising:

providing zero feedback from a quantizer of the look-ahead delta-sigma modulator to the IIR filter prior to determining the multiple natural response outputs of the IIR filter for each set of the input signal samples.

4. The method as in claim 1 wherein the set of multiple input signal samples comprises a current input signal and at least one future input signal sample.

5. The method as in claim 1 further comprising:

for each subsequent discrete time t+1;

providing the quantization output to the IIR filter;

updating state variables of the IIR filter using the provided quantization output and a current input signal sample;

repeating determining a next set of multiple natural response outputs of the IIR filter for a next set of input signal samples, wherein the next set of multiple outputs of the IIR filter represent a natural response vector of the IIR filter for each subsequent time t+1; and

determining a quantization output of the look-ahead delta sigma modulator for each subsequent time t+1 using the next set of multiple natural response outputs of the IIR filter.

6. The method as in claim 1 further comprising:

allocating only an amount of memory for state variables equal to the order of the ITR filter plus the depth of the look-ahead delta-sigma modulator.

7. The method as in claim 1 further comprising:

determining a set of forced pattern responses of the IIR filter of the look-ahead delta sigma modulator to a set of output candidate vectors; and

determining a quantization output further comprises;

quantizing each input signal sample by applying predetermined decision criteria to determine a best match between each forced pattern response and the natural IIR filter response vector and selecting output data from an output candidate vector associated with the forced pattern response used to determine the best match.

8. The method as in claim 7 wherein determining the set of forced pattern responses of the IIR filter of the look-ahead delta sigma modulator further comprises precomputing the set of forced pattern responses before quantizing each input signal data sample.

9. The method as in claim 1 further comprising:

determining the IIR filter natural response vector only once for quantization of each set of input signal samples.

10. The method as in claim 1 wherein the input signal data sample comprises audio input signal data.

11. The method as in claim 1 further comprising:

recording quantized input signal data on storage media.

12. The method as in claim 1 wherein determining a natural IIR filter response comprises setting feedback data to the IIR filter to at least approximately zero and filtering the set of input data signal samples.

13. A method of processing a signal using a multiple output infinite impulse response (“

IIR”

) filter of a look-ahead delta-sigma modulator, the method comprising;

updating a set of state variables of the IIR filter;

conducting a memory-less operation to determine elements of a natural response vector using the updated set of state variables, wherein the elements of the natural response vector include a current response output and at least one future response output of the IIR filter to a set of input signal samples; and

determining a quantization output of the look-ahead delta sigma modulator for time t using the natural response outputs.

14. The method of claim 13 wherein conducting a memory-less operation to determine elements of a natural response vector further comprises:

providing at least approximately zero feedback from a quantizer of the look-ahead delta-sigma modulator to determine the elements of the natural response.

15. The method of claim 13 further comprises:

storing the updated state variables once in a memory; and

wherein conducting a memory-less operation to determine elements of a natural response vector comprises retrieving the stored updated variables once from the memory to determine elements of the natural response vector using the updated set of state variables.

16. A signal processing system comprising:

a look-ahead delta sigma modulator having a depth of M, the look-ahead delta

sigma modulator comprising;

an infinite impulse response (IIR) loop filter to filter a set of input signal samples, the IIR loop filter having K natural response outputs for each of K elements of a natural response vector SNAT_t for each time t, wherein at least one output is for a future output and 1<

K≦

M, wherein K and M are integers; and

a quantizer to determine a quantization output for each time t using the natural response vector SNAT_t.

17. The signal processing system as in claim 16 wherein the set of input signal samples comprises a current input signal sample and M-1 future input signal samples.

18. The signal processing system as in claim 16 further comprising:

a forced pattern response generator coupled to the IIR loop filter to determine a set of forced pattern responses of the loop filter to a set of output candidate vectors.

19. The signal processing system as in claim 16 wherein the pattern generator determines the set of forced pattern responses of the loop filter to a set of vectors obtained by setting signal input data to the loop filter to at least approximately zero and filtering the set of vectors.

20. The signal processing system as in claim 16 wherein the quantizer further comprises a function generator to quantize each input signal data sample by applying predetermined decision criteria to determine a best match between each forced pattern response and the natural response vector SNAT_t and selecting output data from the output candidate vector associated with the forced pattern response used to determine the best match.

21. The signal processing system as in claim 16 wherein the IIR loop filter further comprises:

a first set filter stages;

a second set of filter delay stages sequentially coupled to a last of the first set of filter stages, wherein elements of the natural response vector SNAT_t are state variables of the second set of filter delay stages.

22. The signal processing system as in claim 16 wherein the IIR loop filter further comprises:

a first set filter stages;

a second set of filter delay stages sequentially coupled to a last of the first set of filter stages, wherein elements of the natural response vector SNAT_t are derived from state variables of the second set of filter delay stages.

23. The signal processing system as in claim 16 wherein the natural response generator determines the natural response vector SNAT_t by setting feedback data to the loop filter to at least approximately zero and filtering current input signal sample and M-1 future input signal samples.

24. The signal processing system as in claim 16 wherein the natural response generator determines the natural response vector SNAT_t only once for quantization of each filtered input signal data sample.

25. The signal processing system as in claim 16 wherein the input signal data sample comprises audio input signal data.

26. The signal processing system as in claim 16 further comprising:

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

27. The signal processing system as in claim 16 wherein the look-ahead delta-sigma modulator is implemented using software.

28. An apparatus for processing a signal using a multiple output infinite impulse response (“

IIR”

) filter of a look-ahead delta-sigma modulator, the apparatus comprising;

means for determining a set of state variables for respective stages of the IIR filter for a time t from a set of input signals, wherein at least a subset of the state variables of the respective stages of the IIR filter represent a current response and at least one future response elements of a natural response vector of the IIR filter, wherein t represents a time index value; and

means for determining a quantization output of the look-ahead delta sigma modulator for the time t using the natural response vector determined for the time t.

29. A method of determining feedback coefficients for a multiple output, L-order, infinite impulse response (“

IIR”

) filter of a look-ahead delta-sigma modulator of depth M, wherein the IIR filter includes a plurality of state variables and M is an integer greater than one, the method comprising;

determining feedback coefficients for the IIR filter from the state variables using (i) an at least approximately zero input, (ii) feedback coefficients setting a characteristic of the IIR filter without K-1 natural response delay stages, wherein 1<

K≦

M, and (iii) an impulse function forced feedback response followed by L-1 at least approximately zeros, wherein the feedback coefficients for the IIR filter comprise state variables of each of the initial stages and the K-1 natural response delay stages after completion of the impulse function and L-1 zeros, wherein K and L are integers.