Filterbank structure and method for filtering and separating an information signal into different bands, particularly for audio signal in hearing aids

• US 6,236,731 B1
• Filed: 04/16/1998
• Issued: 05/22/2001
• Est. Priority Date: 04/16/1997
• Status: Expired due to Term

• Current Assignee: K/S HIMPP
• Original Assignee: ON Semiconductor Corporation
• Inventors: Brennan, Robert, Schneider, Anthony Todd
• Primary Examiner(s): Kuntz, Curtis
• Assistant Examiner(s): Ni, Suhan
• Application Number:
  US09/060,823
• Time in Patent Office: 1,132 Days
• Field of Search: 381/23.1, 381/60, 381/74, 381/312, 381/313, 381/314, 381/316, 381/317, 381/320, 381/321, 381/323
• US Class Current: 381/316
• CPC Class Codes: H03H 17/0266 Filter banks
• CPC Class Codes: H03H 17/0294 Variable filters; Programmable filters
• CPC Class Codes: H04R 2225/43 Signal processing in hearing aids to enhance the speech intelligibility
• CPC Class Codes: H04R 2430/03 Synergistic effects of band splitting and sub-band processing
• CPC Class Codes: H04R 2460/03 Aspects of the reduction of energy consumption in hearing devices
• CPC Class Codes: H04R 25/356 Amplitude, e.g. amplitude shift or compression
• CPC Class Codes: H04R 25/407 Circuits for combining signals of a plurality of transducers
• CPC Class Codes: H04R 25/505 using digital signal processing

Abstract

A filterbank structure is provided which provides a flexible compromise between the conflicting goals of processing delay, filter sharpness, memory usage and band interaction. The filterbank has an adjustable number of bands and a stacking which provides for a selectable shift of band frequencies to one of two discrete sets of center frequencies. The width of the bands and hence the number of the bands is selected depending upon acceptable delay, memory usage, and processing speed required. The flexibility in terms of stacking of the bands provides twice the number of potential band edge placements, which is advantageous for hearing loss fitting, especially at low frequencies. The same filter coefficients can be used for analysis and synthesis, to reduce memory usage.

Claims

1. An oversampled filterbank for filtering an information signal, the filterbank having a filterbank structure comprising a filter means defining a filter bandwidth, said filter means filtering said information signal and separating said information signal into a plurality of frequency band signals, each representing one of a plurality of uniformly spaced frequency bands within said filter bandwidth, said frequency bands being stacked in one of an even and an odd manner and said frequency bands overlapping, such that the summation of the unmodified frequency band responses of the plurality of said frequency bands sums to a function within a predetermined passband ripple over said filter bandwidth, wherein the filter means includes a selection input enabling at least one of the following to be selected:

(i) the number of frequency band signals,

(ii) the bandwidth of said frequency bands,

(iii) selection of stacking of said frequency bands in one of an even and an odd manner,

(iv) the degree of overlap between said frequency bands;

(v) an oversampling factor by which said frequency band signals are sampled above the theoretical minimum of critical sampling.

2. A filterbank as claimed in claim 1, wherein the selection input enables at least one of the number of frequency bands and whether said frequency bands are stacked in an even or odd manner to be selected, said number of frequency bands being equal to N, and the filter means comprises:

(a) a first analysis filterbank means for separating said signal into the plurality of N separate frequency band signals;

(b) processing means for receiving and processing each of said separate frequency band signals to provide N separate processed frequency band signals; and

(c) a second synthesis filterbank means for receiving and recombining the N separate processed frequency band signals into a single output signal, wherein both of the first analysis filterbank means and the second synthesis filterbank means are connected to the selection input, the processing means being coupled between the first analysis filterbank means and the second synthesis filterbank means.

3. A filterbank as claimed in claim 2, which comprises a dedicated application specific integrated circuit (ASIC), said ASIC including the first analysis and the second synthesis filterbanks, and a programmable digital signal processor for controlling the number of frequency bands and the bandwidth of each frequency band, said digital signal processor being provided with the selection input.

4. A filterbank as claimed in claim 3, wherein said processing means includes a multiplier means for multiplying each of the frequency band signals by an adjustable gain to provide the N separate processed frequency band signals.

5. A filterbank as claimed in claim 4, wherein the multiplier means comprises one or more dedicated multiplier resources incorporated on the application specific integrated circuit.

6. A filterbank as claimed in claim 4, wherein the multiplier means comprises a multiplier resource provided on the programmable digital signal processor.

7. A filterbank as claimed in claim 1, wherein the selection input enables whether said frequency bands are stacked in an even or odd manner to be selected.

8. A filterbank as claimed in claim 1, wherein the selection input enables whether said frequency bands abut, overlap, or are spaced apart from one another to be selected.

9. A filterbank as claimed in claim 1, wherein the selection input enables the decimation factor to be selected.

10. A filterbank as claimed in claim 1, wherein the selection input enables the degree of overlap in the frequency bands to be selected, including selection of abutting and spaced apart frequency bands.

11. A filterbank as claimed in claim 1, which includes a shared memory interface, for interfacing with a programmable digital signal processor.

12. A filterbank as claimed in claim 1, which includes low frequency processing means for additional processing of low-frequency bands to provide additional resolution.

13. A filterbank as claimed in claim 2, which includes a prefiltering means connected to the first analysis filterbank means, for modifying the gain of at least one selected portion of the frequency spectrum of said information signal.

14. A filterbank as claimed in claim 2 or 13, which includes a postfiltering means connected to the second filterbank means, for postfiltering the single output signal.

15. A filterbank as claimed in claim 3, wherein the first analysis filterbank means, the processing means and the second synthesis filterbank means utilize digital signal processing, the first analysis filterbank means being adapted to receive an input digital sample stream and the second synthesis filterbank means providing an output digital data stream as the output signal wherein the filterbank includes an analog-to-digital conversion means connected to said first analysis filterbank for receiving an original analog signal and for converting said analog signal into an input digital sample stream at an initial input sampling rate which forms said information signal for the analysis filterbank, and a digital-to-analog conversion means connected to said second synthesis filterbank for converting the output digital data stream to form an analog version of said single output signal.

16. A filterbank as claimed in claim 15, wherein the first analysis filterbank means comprises:

(a) a blocking means for receiving the input digital sample stream and blocking a first number, R where R≦

N, of the digital samples so as to provide a blocked input digital sample stream, the ratio of N/R corresponding to an oversampling factor;

(b) an analysis window means for applying an analysis window function to the input digital sample stream to provide a windowed blocked digital sample stream, said analysis window function being defined by a set of analysis window coefficients;

(c) a time folding means for overlapping and adding blocks of said windowed blocked digital sample stream, each of said blocks comprising N digital samples, to provide a summed block of N digital samples; and

(d) a discrete transform means for receiving said summed block of N digital samples and transforming the signal into a discrete frequency domain signal having N components, the N components corresponding to the N frequency band signals.

17. A filterbank as claimed in claim 16, wherein the second synthesis filterbank means comprises:

(a) an inverse discrete transform means for receiving said N processed frequency band signals and for effecting an inverse transform to form a block of N digital samples;

(b) a replication and concatenation means for replicating and concatenating said processed block of N digital samples to provide a periodically extended block of N digital samples;

(c) a synthesis window means for applying a synthesis window function to said extended block of N digital samples to provide a windowed periodically extended block of N digital samples, said synthesis window function being defined by a set of synthesis window coefficients; and

(d) a summation buffer means for receiving said windowed periodically extended block of N digital samples and adding said windowed periodically extended samples to the shifted contents of the buffer each time a new windowed periodically extended sample is received, so as to provide a processed information signal.

18. A filterbank as claimed in claim 17, wherein said synthesis window function is based on a decimated version of said analysis window function.

19. A filterbank as claimed in claim 16, 17, or 18 in which the programmable digital signal processor is operable to vary said analysis window coefficients and said synthesis window coefficients.

20. A filterbank as claimed in claim 16 or 17, in which the programmable digital signal processor is operable to vary said oversampling factor, and in which the oversampling factor is at least equal to 2.

21. A filterbank as claimed in claim 3 in which the programmable digital signal processor is operable to provide either even stacking or odd stacking of the frequency bands within said filter bandwidth.

22. A filterbank as claimed in claim 1, 2, or 3, in which said filterbank is incorporated in a digital hearing aid.

23. A filterbank as claimed in claim 1, wherein the selection input enables the number of frequency band signals to be selected.

24. A filterbank as claimed in claim 23, wherein the selection input further enables the bandwidth of said frequency bands to be selected.

25. An oversampled filterbank, for filtering an information signal, the filterbank having a filterbank structure comprising:

(a) a first analysis filterbank means for separating said signal into a plurality of N separate frequency bands, wherein the first analysis filterbank means includes transform means for transforming the audio signal into the frequency domain, with the N separate frequency band signals being present in the frequency domain;

(b) processing means for receiving and processing each of said separate frequency band signals to provide N separate processed frequency band signals, wherein said processing means includes a multiplier means for multiplying each of the frequency band signals by an adjustable gain to provide the N separate processed frequency band signals;

(c) a second synthesis filterbank means for receiving and recombining the N separate processed frequency band signals into a single output signal, the processing means being coupled between the first analysis filterbank means and the second synthesis filterbank means, wherein the second synthesis filterbank means includes inverse transform means for transforming the N separate frequency band signals into the single output signal in the time domain;

(d) a selection input connected to both of the first analysis filterbank means and the second synthesis filterbank means, to enable the number of bands and the bandwidth of each frequency band to be selected, wherein the selection input further enables at least one of the following to be selected;

(i) the number of frequency band signals;

(ii) the bandwidth of said frequency bands;

(iii) selection of stacking of said frequency bands in one of an even and an odd manner;

(iv) the degree of overlap between said frequency bands;

(v) an oversampling factor by which said frequency band is sampled above the theoretical minimum of critical sampling wherein the filterbank comprises an application specific integrated circuit, said application specific integrated circuit including the first analysis and the second synthesis filterbanks, and a programable digital signal processor for controlling the number of frequency bands and the bandwidth of each frequency band, said digital signal processor being provided with the selection input.

26. A filterbank as claimed in claim 25, which is adapted to receive a single real monaural information signal, wherein said transform means generates non-negative frequency band signals and negative frequency band signals, said negative frequency band signals being derivable from the non-negative frequency band signals, and said processing means processes only said non-negative frequency band signals.

27. A filterbank as claimed in claim 25, wherein the filterbank is adapted to filter an audio signal comprising first and second real monaural information signals which are combined into a complex stereo signal and wherein said transform means generates N combined frequency band signals, and wherein said processing means includes:

(a) channel separation means for separating the N combined frequency band signals into the N frequency band signals corresponding to said first information signal and the N frequency band signals corresponding to said second information signal, each of said N frequency band signals comprising non-negative and negative frequency band signals;

(b) first independent channel processing means connected to the channel separation means for receiving and processing each of said separate frequency band signals of said first information signal to provide a first set of N separate processed frequency band signals;

(c) second independent channel processing means connected to channel separation means for receiving and processing each of said separate frequency band signals of said second information signal to provide a second set of N separate processed frequency band signals; and

(d) channel combination means connected to the first and second independent channel processing means for combining said first set of N processed separate frequency band signals and said second set of N processed separate frequency band signals.

28. A filterbank as claimed in claim 27, wherein said first and second independent channel processing means each process only the non-negative frequency band signals of the corresponding information signal, the negative frequency band signals being derivable from the non-negative frequency band signals.

29. A method of processing an information signal to selectively modify different frequency bands, the method comprising the steps of:

(1) defining a filter frequency bandwidth to be analyzed;

(2) dividing the filter frequency bandwidth into a plurality of uniformly spaced bands and defining characteristics of the filter bands, and selecting stacking of said frequency bands into one of an even and an odd manner and the degree of overlap between said frequency bands;

(3) filtering the information signal to separate the signal into a plurality of frequency band signals, each representing one of said uniform filter bands, including oversampling the frequency band signals by a fraction above a theoretical initial sampling, thereby determining a decimation factor;

(4) processing the frequency band signals; and

(5) recombining the signals of the individual bands to form an output signal;

wherein the method includes as an additional step (6), at least one of;

(i) selecting the number of frequency band signals,

(ii) selecting the bandwidth of said frequency bands,

(iii) selecting stacking of said frequency bands in one of an even and an odd manner,

(iv) selecting the degree of overlap of said frequency bands, and

(v) selecting the oversampling a decimation factor by which said frequency band signals are sampled above critical sampling.

30. A method as claimed in claim 29, which includes selecting the degree of overlap of said frequency bands and providing for the selection of abutting and spaced apart frequency bands.

31. A method as claimed in claim 29, wherein the input in step (6) enables whether said frequency bands abut, overlap, or are spaced apart from one another to be selected.

32. A method as claimed in claim 29, wherein the input in step (6) enables the decimation factor to be selected.

33. A method as claimed in claim 29, wherein the input in step (6) enables whether said frequency bands are stacked in an even or odd manner to be selected.

34. A method as claimed in claim 29, wherein step (4) comprises setting a gain for each frequency band and multiplying each frequency band signal by the respective set gain.

35. A method as claimed in claim 29 which comprises applying the method to an audio information signal in a digital hearing aid.

36. A method as claimed in claim 35, including enhancing low-frequency bands by additional processing.

37. A method as claimed in claim 36, which includes subdividing at least one low-frequency band into smaller bands for individual processing.

38. A method as claimed in claim 35, which comprises selecting the number of bands and the width of the bands, in dependence upon acceptable delays and desired resolution.

39. A method as claimed in claim 29, wherein the input in step (6) enables the number of frequency band signals to be selected.

40. A method as claimed in claim 39, wherein the input in step (6) further enables the bandwidth of said frequency bands to be selected.

41. A method as claimed in claim 29, the method further comprising:

(a) in step (3) separating said information signal into N separate frequency band signals;

(b) in step (4) processing each of said separate frequency band signals to provide N separate processed frequency band signals;

(c) in step (5), recombining the N separate processed frequency band signals to form the output signal; and

(d) selecting either the number of frequency bands or whether said frequency bands are stacked in an even or odd manner.

42. A method as claimed in claim 41, which includes transforming the information signal into the frequency domain, providing N separate frequency band signals in the frequency domain, and effecting an inverse transform of the N separate processed frequency band signals into the output signal in the time domain.

43. A method as claimed in claim 42, which comprises generating non-negative frequency band signals and negative frequency band signals, the negative frequency band signals being derivable from the non-negative frequency band signals, and, in step (4), processing only said non-negative frequency band signals.

44. A method as claimed in claim 43, which comprises:

(a) receiving an information signal comprising first and second real monaural information signals combined into a complex stereo signal and, when transforming the information signal into the frequency domain, generating N combined frequency band signals;

(b) separating the N combined frequency band signals into the N frequency band signals corresponding to said first information signal and the frequency band signals corresponding to said second information signal, each of said N frequency band signals comprising non-negative and negative frequency band signals;

(c) processing each of said separate frequency band signals of said first information signal to provide a first set of N separate processed frequency band signals;

(d) processing each of said separate frequency band signals of said second information signal to provide a second set of N separate processed frequency band signals; and

(e) combining said first set of N processed separate frequency band signals and said second set of N processed separate frequency band signals.

45. A method as claimed in claim 44, which comprises, for each of said first and second information signals processing only the non-negative frequency band signals of the corresponding information signal.

46. A method as claimed in claim 42, wherein the method is effected by digital signal processing, the method including:

passing an original analog signal through an analog-to-digital conversion means to convert the analog signal into an input digital sample stream at an initial input sampling rate which forms said information signal, and wherein the output signal after the inverse transform comprises an output digital data stream; and

effecting a digital to analog conversion of the output digital data stream to form an analog version of the output signal.

47. A method as claimed in claim 46, wherein the step of transforming the information signal into the frequency domain comprises:

(a) blocking the input digital sample stream into blocks of R samples, where R≦

N, to provide a blocked input digital sample stream, the ratio of N/R corresponding to an oversampling factor;

(b) applying an analysis window function to the input digital sample stream to provide a windowed blocked digital sample stream, said analysis window function being defined by a set of analysis window coefficients;

(c) overlapping and adding blocks of said windowed blocked digital sample stream, each of said blocks comprising N digital samples, to provide a summed block of N digital samples; and

(d) transforming the signal into a discrete frequency domain signal having N components, the N components corresponding to the N frequency band signals.

48. A method as claimed in claim 47, wherein the step of effecting an inverse transform of the N separate processed frequency band signals comprises:

(a) effecting an inverse transform to form a block of N digital samples;

(b) replicating and concatenating said processed block of N digital samples to provide a periodically extended block of N digital samples;

(c) applying a synthesis window function to said extended block of N digital samples to provide a windowed periodically extended block of N digital samples, said synthesis window function being defined by a set of synthesis window coefficients;

(d) adding said windowed periodically extended block of N samples to a summation buffer; and

(e) each time a new windowed periodically extended sample is received, shifting the contents of the summation buffer by R samples from one end of the summation buffer towards the other end of the buffer, providing zeros to fill the R empty samples at the one end of the buffer, and passing the R samples displaced out of the other end of the buffer to an output to provide a processed information signal.

49. A method as claimed in claim 48 wherein the analysis window coefficients and the synthesis window coefficients can be varied.

50. A method as claimed in claim 48, which includes forming the synthesis window function by decimating the analysis window function.

51. A method as claimed in claims 42, 46 or 48, which includes selecting stacking of the frequency bands in one of even stacking and odd stacking, within the frequency bandwidth.

52. A method as claimed in claim 47 or 48 wherein the oversampling factor can be varied and is at least equal to 2.