Method for noise shaping and a noise shaping filter

Patent metadata

Field	Value
Publication number	US-10523937-B2
Application number	US-201313946457-A
Country	US
Kind code	B2
Filing date	Jul 19, 2013
Priority date	Jul 19, 2013
Publication date	Dec 31, 2019
Grant date	Dec 31, 2019

Abstract

A method for noise shaping includes: reducing a bit-depth of an input signal to obtain a quantized input signal; feeding a quantization error corresponding to the bit-depth reduction of the input signal into a feedback loop to the input signal, the feedback loop comprising a first quantization stage, a second quantization stage and a correction stage, both the first and second quantization stages operating at the bit-depth of the input signal and the correction stage operating at a bit-depth of the quantization error; and generating a noise-shaped output signal at lower clock rate than the input signal based on the feedback loop.

First claim

What is claimed is:

1. A method for noise shaping, the method comprising:
reducing a bit-depth of an input signal to obtain a quantized input signal;
feeding a quantization error corresponding to the bit-depth reduction of the input signal into a feedback loop to the input signal, the feedback loop comprising a first quantization stage, a second quantization stage and a correction stage, both the first and second quantization stages operating at the bit-depth of the input signal and the correction stage operating at a bit-depth of the quantization error; and
generating a noise-shaped output signal at lower clock rate than the input signal based on the feedback loop.

2. The method of claim 1 , wherein a bit-depth reduction of the first quantization stage corresponds to a bit-depth reduction of the second quantization stage.

3. The method of claim 1 , comprising:
reducing a bit-depth of an input to the first quantization stage to a bit-depth of the quantization error when passing the input through the first quantization stage.

4. The method of claim 1 , wherein an input to the first quantization stage corresponds to an input to the correction stage.

5. The method of claim 1 , wherein an input to the second quantization stage is different from an input to the correction stage.

6. The method of claim 1 , comprising:
using an additive property for the combination of the first and second quantization stages and the correction stage.

7. The method of claim 6 , wherein the additive property is according to the following:
Q ( a 1 +Q ( a 2))= Q ( a 1)+ Q ( a 2)+delta( a 1),

where Q(a 1 ) denotes the first quantization stage, Q(a 2 ) denotes the second quantization stage, delta(a 1 ) denotes the correction stage, a 1 denotes an input to both the first quantization stage and the correction stage and a 2 denotes an input to the second quantization stage.

8. The method of claim 1 , wherein the correction stage outputs a one if an input to the correction stage is smaller than zero and a fractional part of that input corresponds to two to the half bit-depth of the quantization error and otherwise outputs a zero.

9. The method of claim 8 , comprising:
determining the fractional part of the input of the correction stage by applying a modulo operation to the input of the correction stage, the modulo operation with respect to two to the bit-depth of the quantization error.

10. The method of claim 1 , comprising:
implementing the combination of the first quantization stage and the correction stage based on a lookup table.

11. The method of claim 10 , wherein a size of the lookup table corresponds to two to a bit-depth of the quantization error.

12. The method of claim 10 , wherein a bit-depth of an output of the lookup table is by one bit greater than a bit-depth of an input to the lookup table.

13. The method of claim 1 , wherein the feedback loop comprises a filtering of the quantization error with filter coefficients whose real and imaginary parts are integers.

14. A noise shaping filter, comprising:
a quantizer circuit configured to reduce a bit-depth of an input signal to obtain a quantized input signal;
a feedback loop circuit configured to combine a quantization error of the quantizer with the input signal to obtain a noise-shaped output signal,
wherein the feedback loop circuit comprises a combination of a first and a second quantization stage operating at the bit-depth of the input signal and a correction stage operating at a bit-depth of the quantization error.

15. The noise shaping filter of claim 14 , wherein the feedback loop circuit is configured to provide the noise-shaped output signal in a lower clock rate than the input signal.

16. The noise shaping filter of claim 14 , comprising a lookup table configured to implement the combination of the first quantization stage and the correction stage.

17. The noise shaping filter of claim 16 , configured to combine a first and a second clock sample of the input signal providing one clock sample of the noise-shaped output signal.

18. The noise shaping filter of claim 17 , wherein a bit-depth of the combination of the first and second clock samples of the input signal is by one bit larger than a bit-depth of the first or the second clock sample of the input signal.

19. A noise shaper, comprising:
a first and a second noise shaping filter circuit, each of them comprising:
a quantizer circuit configured to reduce a bit-depth of an input signal to obtain a quantized input signal;
a feedback loop circuit configured to combine a quantization error of the quantizer circuit with the input signal to obtain a noise-shaped output signal,
wherein the feedback loop circuit comprises a combination of a first and a second quantization stage operating at the bit-depth of the input signal and a correction stage operating at a bit-depth of the quantization error.

20. The noise shaper of claim 19 , wherein the feedback loop circuits of the first and second noise shaping filters are configured to provide the noise-shaped output signal in a lower clock rate than the input signal.

21. The noise shaper of claim 19 , configured to provide two clock samples of the noise-shaped output signal in every two clocks.

22. The noise shaper of claim 19 , configured to receive two clock samples of the input signal in every clock.

23. The noise shaper of claim 19 ,
wherein the first noise shaping filter circuit is configured to combine a first and a second clock sample of the input signal providing a first clock sample of the noise-shaped output signal; and
wherein the second noise shaping filter circuit is configured to combine the second and a third clock sample of the input signal providing a second clock sample of the noise-shaped output signal.

24. The noise shaper of claim 23 , comprising a shift register configured to store the first, second and third clock samples of the input signal.

25. The noise shaper of claim 23 , comprising a bit adder configured to transform a range of the input signal to positive numbers or to negative numbers before being processed by the first and second noise shaping filters.

26. A mobile device, comprising:
a baseband processor configured to provide a baseband signal; and
a transmitter configured to modulate the baseband signal on a radio frequency carrier and to transmit the modulated baseband signal through an air link to a base station,
wherein the transmitter comprises a noise shaper configured to noise-shape the baseband signal, the noise shaper comprising:
a first and a second noise shaping filter, each of them comprising:
a quantizer configured to reduce a bit-depth of the baseband signal to obtain a quantized baseband signal;
a feedback loop configured to combine a quantization error of the quantizer with the baseband signal to obtain a noise-shaped baseband signal,
wherein the feedback loop comprises a combination of a first and a second quantization stage operating at the bit-depth of the baseband signal and a correction stage operating at a bit-depth of the quantization error.

27. The mobile device of claim 26 , wherein the transmitter further comprises:
a digital-to-analog converter configured to convert the noise-shaped baseband signal to an analog baseband signal; and
a modulator configured to modulate the analog baseband signal to the radio frequency carrier.

Assignees

• Kerner Michael
• Dinur Nati
• Gurevitz Assaf
• Intel Corp

Inventors

• Kerner Michael
• Dinur Nati
• Gurevitz Assaf

Classifications

• H04N19/126

  Details of normalisation or weighting functions, e.g. normalisation matrices or variable uniform quantisers · CPC title

• H04N19/117 Primary

  Filters, e.g. for pre-processing or post-processing (sub-band filter banks H04N19/635) · CPC title

• H04N19/154

  Measured or subjectively estimated visual quality after decoding, e.g. measurement of distortion (use of rate-distortion criteria H04N19/147) · CPC title