System to reduce distortion due to coding with a sample-by-sample quantizer
View Patent ↗A distortion removal system for reducing the quantization distortion that results from quantization of a signal by sample-by-sample quantization techniques is disclosed. The distortion removal system uses the signal previously quantized with the sample-by-sample quantization techniques to determine an expected quantization distortion. The expected quantization distortion is determined for each of a plurality of frames of the signal. The expected quantization distortion is removed from each of the frames of the signal.
1. A method of reducing a quantization distortion created by quantization of a speech signal by a sample-by-sample quantizer, the speech signal including a plurality of frames, the method comprising:
detecting that one frame of the plurality of frames was previously quantized;
determining a quantization level for each of a plurality of samples of the one frame;
estimating an expected quantization distortion for each of the plurality of samples based on the quantization level of each of the plurality of samples;
summing up the expected quantized distortion of each of the plurality of samples to generate a summed quantization distortion; and
removing the summed quantization distortion from the one frame.
2. The method of claim 1 , wherein the determining includes using quantization indices to determine the quantization level for each of a plurality of samples.
3. The method of claim 2 , wherein the quantization indices are obtained by analyzing the one frame to determine a type of the sample-by-sample quantizer and requantizing the one frame.
4. The method of claim 1 where reducing the quantization distortion is part of a pre-processing of the speech signal prior to encoding.
5. The method of claim 1 where reducing the quantization distortion is part of a post-processing of the speech signal following decoding.
6. The method of claim 1 , wherein the removing is performed in the frequency domain.
7. The method of claim 1 , wherein the removing is performed in the time domain.
8. A method of reducing quantization distortion created by a sample-by-sample quantizer, the method comprising:
(a) determining quantization levels of a frame of a previously quantized signal;
(b) estimating an expected quantization distortion for each of the quantization levels;
(c) summing the expected quantization distortion for each of the quantization levels for the frame of the previously quantized signal; and
(d) removing the summed expected quantization distortion from the frame.
9. The method of claim 7 where (b) further comprises determining a distribution of the previously quantized signal within the quantization levels and storing the expected quantization distortion in a distortion table.
10. The method of claim 8 where (c) further comprises assuming a magnitude spectrum of the expected quantization distortion is flat and assuming a phase spectrum of the expected quantization distortion is the same as the previously quantized signal.
11. The method of claim 8 further comprising initially determining if a signal has been subject to quantization by the sample-by-sample quantizer.
12. The method of claim 10 further comprising executing the method after the initially determining if the signal has been subject to quantization by the sample-by-sample quantizer.
13. The method of claim 11 further comprising determining a type of sample-by-sample quantization.
14. The method of claim 7 where (a) further comprises quantizing a signal with a sample-by-sample quantizer prior o determining the quantization levels.
15. The method of claim 8 where (d) further comprises removing the expected quantization distortion in the frequency domain.
16. The method of claim 8 where (d) further comprises removing the expected quantization distortion in the time domain.
17. A method of estimating quantization distortion for a frame of a signal that has been quantized using sample-by-sample quantization, the method comprising:
(a) determining a distribution of the signal within a plurality of quantization levels;
(b) estimating an expected quantization distortion for each of the quantization levels based on the distribution;
(c) determining an expected quantization distortion for the frame as a function of the expected quantization distortion of each of the quantization levels;
(d) determining a phase spectrum of the expected quantization distortion of the frame.
18. The method of claim 17 further comprising (e) determining a spectral shape of the expected quantization distortion of the frame as a function of an error criterion used during the sample-by-sample quantization.
19. The method of claim 18 further comprising (f) approximating the spectral shape of the expected quantization distortion of the frame as flat.
20. The method of claim 17 further comprising (e) assuming that the phase spectrum of the expected quantization distortion of the frame is equal to the phase spectrum of the frame.
21. The method of claim 17 where (b) further comprises determining an upper boundary and a lower boundary for each of the quantization levels.
22. A method of reducing the quantization distortion created during quantization of a signal by a sample-by-sample quantizer, where a frame of the signal comprises a plurality of samples that are quantized to one of a plurality of quantization levels by the sample-by-sample quantizer, the method comprising:
(a) receiving the frame of the signal;
(b) identifying a quantization level of each of the samples;
(c) obtaining an expected quantization distortion of each of the samples;
(d) summing the expected quantization distortion of each of the samples; and
(e) removing the sum of the expected quantization distortion from the frame.
23. The method of claim 22 further comprising initially determining if the frame was previously quantized.
24. The method of claim 22 where (a) comprises receiving the signal with a base station.
25. The method of claim 22 where (a) comprises receiving the signal with a mobile communication device.
26. The method of claim 22 where (a) comprises receiving the signal with a public switched telephone network.
27. The method of claim 22 where (a) comprises receiving the signal from a communication medium.
28. The method of claim 22 where (a) comprises receiving the signal with a packet-based network.
29. The method of claim 22 where (c) further comprises determining a distribution of the samples within the quantization levels.
30. The method of claim 22 where (d) further comprises determining a magnitude spectrum of the expected quantization distortion, and determining a phase spectrum of the expected quantization distortion.
31. The method of claim 22 where the quantization levels are obtained from the signal without additional processing.
32. The method of claim 22 where (b) further comprises re-quantizing the signal.
33. The method of claim 22 where (c) further comprises retrieving the expected quantization distortion from a distortion table.
34. The method of claim 22 where (e) further comprises removing the expected quantization distortion in the frequency domain.
35. The method of claim 22 where (e) further comprises removing the expected quantization distortion in the time domain.
36. A distortion removal system for a frame of a signal that includes quantization distortion resulting from the frame being previously quantized to a plurality of quantization levels by a sample-by-sample quantizer, the distortion removal system comprising:
a distortion identification module operable to identify an expected quantization distortion for each of the quantization levels in the frame;
a summer module operable to sum the expected quantization distortion of each of the quantization levels in the frame; and
a distortion removal module operable to remove the summed expected quantization distortion.
37. The distortion removal system of claim 36 further comprising an initial processing module operable to determine and provide the quantization levels to the distortion identification module.
38. The distortion removal system of claim 37 where the initial processing module further comprises a sensing module and a quantization module.
39. The distortion removal system of claim 36 where the distortion identification module comprises a distortion determination module and a distortion table.
40. The distortion removal system of claim 36 where the distortion removal system is operable to pre-process the signal prior to encoding.
41. The distortion removal system of claim 36 where the distortion removal system is operable to post-process the signal following decoding.
42. The distortion removal system of claim 36 where the distortion removal module is operable to remove the expected quantization distortion in the frequency domain.
43. The distortion removal system of claim 36 where the distortion removal module is operable to remove the expected quantization distortion in the time domain.